3 \cfg{text-title-align}{left}
4 \cfg{text-chapter-align}{left}
5 \cfg{text-chapter-numeric}{true}
6 \cfg{text-chapter-suffix}{. }
7 \cfg{text-chapter-underline}{-}
8 \cfg{text-section-align}{0}{left}
9 \cfg{text-section-numeric}{0}{true}
10 \cfg{text-section-suffix}{0}{. }
11 \cfg{text-section-underline}{0}{-}
12 \cfg{text-section-align}{1}{left}
13 \cfg{text-section-numeric}{1}{true}
14 \cfg{text-section-suffix}{1}{. }
15 \cfg{text-section-underline}{1}{-}
16 \cfg{text-versionid}{0}
18 \cfg{html-contents-filename}{index.html}
19 \cfg{html-template-filename}{%k.html}
20 \cfg{html-index-filename}{docindex.html}
21 \cfg{html-leaf-level}{1}
22 \cfg{html-contents-depth-0}{1}
23 \cfg{html-contents-depth-1}{3}
24 \cfg{html-leaf-contains-contents}{true}
26 \define{dash} \u2013{-}
28 \title Developer documentation for Simon Tatham's puzzle collection
30 This is a guide to the internal structure of Simon Tatham's Portable
31 Puzzle Collection (henceforth referred to simply as \q{Puzzles}),
32 for use by anyone attempting to implement a new puzzle or port to a
35 This guide is believed correct as of r6190. Hopefully it will be
36 updated along with the code in future, but if not, I've at least
37 left this version number in here so you can figure out what's
38 changed by tracking commit comments from there onwards.
40 \C{intro} Introduction
42 The Puzzles code base is divided into four parts: a set of
43 interchangeable front ends, a set of interchangeable back ends, a
44 universal \q{middle end} which acts as a buffer between the two, and
45 a bunch of miscellaneous utility functions. In the following
46 sections I give some general discussion of each of these parts.
48 \H{intro-frontend} Front end
50 The front end is the non-portable part of the code: it's the bit
51 that you replace completely when you port to a different platform.
52 So it's responsible for all system calls, all GUI interaction, and
53 anything else platform-specific.
55 The current front ends in the main code base are for Windows, GTK
56 and MacOS X; I also know of a third-party front end for PalmOS.
58 The front end contains \cw{main()} or the local platform's
59 equivalent. Top-level control over the application's execution flow
60 belongs to the front end (it isn't, for example, a set of functions
61 called by a universal \cw{main()} somewhere else).
63 The front end has complete freedom to design the GUI for any given
64 port of Puzzles. There is no centralised mechanism for maintaining
65 the menu layout, for example. This has a cost in consistency (when I
66 \e{do} want the same menu layout on more than one platform, I have
67 to edit two pieces of code in parallel every time I make a change),
68 but the advantage is that local GUI conventions can be conformed to
69 and local constraints adapted to. For example, MacOS X has strict
70 human interface guidelines which specify a different menu layout
71 from the one I've used on Windows and GTK; there's nothing stopping
72 the OS X front end from providing a menu layout consistent with
75 Although the front end is mostly caller rather than the callee in
76 its interactions with other parts of the code, it is required to
77 implement a small API for other modules to call, mostly of drawing
78 functions for games to use when drawing their graphics. The drawing
79 API is documented in \k{drawing}; the other miscellaneous front end
80 API functions are documented in \k{frontend-api}.
82 \H{intro-backend} Back end
84 A \q{back end}, in this collection, is synonymous with a \q{puzzle}.
85 Each back end implements a different game.
87 At the top level, a back end is simply a data structure, containing
88 a few constants (flag words, preferred pixel size) and a large
89 number of function pointers. Back ends are almost invariably callee
90 rather than caller, which means there's a limitation on what a back
91 end can do on its own initiative.
93 The persistent state in a back end is divided into a number of data
94 structures, which are used for different purposes and therefore
95 likely to be switched around, changed without notice, and otherwise
96 updated by the rest of the code. It is important when designing a
97 back end to put the right pieces of data into the right structures,
98 or standard midend-provided features (such as Undo) may fail to
101 The functions and variables provided in the back end data structure
102 are documented in \k{backend}.
104 \H{intro-midend} Middle end
106 Puzzles has a single and universal \q{middle end}. This code is
107 common to all platforms and all games; it sits in between the front
108 end and the back end and provides standard functionality everywhere.
110 People adding new back ends or new front ends should generally not
111 need to edit the middle end. On rare occasions there might be a
112 change that can be made to the middle end to permit a new game to do
113 something not currently anticipated by the middle end's present
114 design; however, this is terribly easy to get wrong and should
115 probably not be undertaken without consulting the primary maintainer
116 (me). Patch submissions containing unannounced mid-end changes will
117 be treated on their merits like any other patch; this is just a
118 friendly warning that mid-end changes will need quite a lot of
119 merits to make them acceptable.
121 Functionality provided by the mid-end includes:
123 \b Maintaining a list of game state structures and moving back and
124 forth along that list to provide Undo and Redo.
126 \b Handling timers (for move animations, flashes on completion, and
127 in some cases actually timing the game).
129 \b Handling the container format of game IDs: receiving them,
130 picking them apart into parameters, description and/or random seed,
131 and so on. The game back end need only handle the individual parts
132 of a game ID (encoded parameters and encoded game description);
133 everything else is handled centrally by the mid-end.
135 \b Handling standard keystrokes and menu commands, such as \q{New
136 Game}, \q{Restart Game} and \q{Quit}.
138 \b Pre-processing mouse events so that the game back ends can rely
139 on them arriving in a sensible order (no missing button-release
140 events, no sudden changes of which button is currently pressed,
143 \b Handling the dialog boxes which ask the user for a game ID.
145 \b Handling serialisation of entire games (for loading and saving a
146 half-finished game to a disk file, or for handling application
147 shutdown and restart on platforms such as PalmOS where state is
148 expected to be saved).
150 Thus, there's a lot of work done once by the mid-end so that
151 individual back ends don't have to worry about it. All the back end
152 has to do is cooperate in ensuring the mid-end can do its work
155 The API of functions provided by the mid-end to be called by the
156 front end is documented in \k{midend}.
158 \H{intro-utils} Miscellaneous utilities
160 In addition to these three major structural components, the Puzzles
161 code also contains a variety of utility modules usable by all of the
162 above components. There is a set of functions to provide
163 platform-independent random number generation; functions to make
164 memory allocation easier; functions which implement a balanced tree
165 structure to be used as necessary in complex algorithms; and a few
166 other miscellaneous functions. All of these are documented in
169 \H{intro-structure} Structure of this guide
171 There are a number of function call interfaces within Puzzles, and
172 this guide will discuss each one in a chapter of its own. After
173 that, (\k{writing}) discusses how to design new games, with some
174 general design thoughts and tips.
176 \C{backend} Interface to the back end
178 This chapter gives a detailed discussion of the interface that each
179 back end must implement.
181 At the top level, each back end source file exports a single global
182 symbol, which is a \c{const struct game} containing a large number
183 of function pointers and a small amount of constant data. This
184 structure is called by different names depending on what kind of
185 platform the puzzle set is being compiled on:
187 \b On platforms such as Windows and GTK, which build a separate
188 binary for each puzzle, the game structure in every back end has the
189 same name, \cq{thegame}; the front end refers directly to this name,
190 so that compiling the same front end module against a different back
191 end module builds a different puzzle.
193 \b On platforms such as MacOS X and PalmOS, which build all the
194 puzzles into a single monolithic binary, the game structure in each
195 back end must have a different name, and there's a helper module
196 \c{list.c} (constructed automatically by the same Perl script that
197 builds the \cw{Makefile}s) which contains a complete list of those
200 On the latter type of platform, source files may assume that the
201 preprocessor symbol \c{COMBINED} has been defined. Thus, the usual
202 code to declare the game structure looks something like this:
205 \c #define thegame net /* or whatever this game is called */
206 \e iii iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
209 \c const struct game thegame = {
210 \c /* lots of structure initialisation in here */
211 \e iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
214 Game back ends must also internally define a number of data
215 structures, for storing their various persistent state. This chapter
216 will first discuss the nature and use of those structures, and then
217 go on to give details of every element of the game structure.
219 \H{backend-structs} Data structures
221 Each game is required to define four separate data structures. This
222 section discusses each one and suggests what sorts of things need to
225 \S{backend-game-params} \c{game_params}
227 The \c{game_params} structure contains anything which affects the
228 automatic generation of new puzzles. So if puzzle generation is
229 parametrised in any way, those parameters need to be stored in
232 Most puzzles currently in this collection are played on a grid of
233 squares, meaning that the most obvious parameter is the grid size.
234 Many puzzles have additional parameters; for example, Mines allows
235 you to control the number of mines in the grid independently of its
236 size, Net can be wrapping or non-wrapping, Solo has difficulty
237 levels and symmetry settings, and so on.
239 A simple rule for deciding whether a data item needs to go in
240 \c{game_params} is: would the user expect to be able to control this
241 data item from either the preset-game-types menu or the \q{Custom}
242 game type configuration? If so, it's part of \c{game_params}.
244 \c{game_params} structures are permitted to contain pointers to
245 subsidiary data if they need to. The back end is required to provide
246 functions to create and destroy \c{game_params}, and those functions
247 can allocate and free additional memory if necessary. (It has not
248 yet been necessary to do this in any puzzle so far, but the
249 capability is there just in case.)
251 \c{game_params} is also the only structure which the game's
252 \cw{compute_size()} function may refer to; this means that any
253 aspect of the game which affects the size of the window it needs to
254 be drawn in must be stored in \c{game_params}. In particular, this
255 imposes the fundamental limitation that random game generation may
256 not have a random effect on the window size: game generation
257 algorithms are constrained to work by starting from the grid size
258 rather than generating it as an emergent phenomenon. (Although this
259 is a restriction in theory, it has not yet seemed to be a problem.)
261 \S{backend-game-state} \c{game_state}
263 While the user is actually playing a puzzle, the \c{game_state}
264 structure stores all the data corresponding to the current state of
267 The mid-end keeps \c{game_state}s in a list, and adds to the list
268 every time the player makes a move; the Undo and Redo functions step
269 back and forth through that list.
271 Therefore, a good means of deciding whether a data item needs to go
272 in \c{game_state} is: would a player expect that data item to be
273 restored on undo? If so, put it in \c{game_state}, and this will
274 automatically happen without you having to lift a finger. If not
275 \dash for example, the deaths counter in Mines is precisely
276 something that does \e{not} want to be reset to its previous state
277 on an undo \dash then you might have found a data item that needs to
278 go in \c{game_ui} instead.
280 During play, \c{game_state}s are often passed around without an
281 accompanying \c{game_params} structure. Therefore, any information
282 in \c{game_params} which is important during play (such as the grid
283 size) must be duplicated within the \c{game_state}. One simple
284 method of doing this is to have the \c{game_state} structure
285 \e{contain} a \c{game_params} structure as one of its members,
286 although this isn't obligatory if you prefer to do it another way.
288 \S{backend-game-drawstate} \c{game_drawstate}
290 \c{game_drawstate} carries persistent state relating to the current
291 graphical contents of the puzzle window. The same \c{game_drawstate}
292 is passed to every call to the game redraw function, so that it can
293 remember what it has already drawn and what needs redrawing.
295 A typical use for a \c{game_drawstate} is to have an array mirroring
296 the array of grid squares in the \c{game_state}; then every time the
297 redraw function was passed a \c{game_state}, it would loop over all
298 the squares, and physically redraw any whose description in the
299 \c{game_state} (i.e. what the square needs to look like when the
300 redraw is completed) did not match its description in the
301 \c{game_drawstate} (i.e. what the square currently looks like).
303 \c{game_drawstate} is occasionally completely torn down and
304 reconstructed by the mid-end, if the user somehow forces a full
305 redraw. Therefore, no data should be stored in \c{game_drawstate}
306 which is \e{not} related to the state of the puzzle window, because
307 it might be unexpectedly destroyed.
309 The back end provides functions to create and destroy
310 \c{game_drawstate}, which means it can contain pointers to
311 subsidiary allocated data if it needs to. A common thing to want to
312 allocate in a \c{game_drawstate} is a \c{blitter}; see
313 \k{drawing-blitter} for more on this subject.
315 \S{backend-game-ui} \c{game_ui}
317 \c{game_ui} contains whatever doesn't fit into the above three
320 A new \c{game_ui} is created when the user begins playing a new
321 instance of a puzzle (i.e. during \q{New Game} or after entering a
322 game ID etc). It persists until the user finishes playing that game
323 and begins another one (or closes the window); in particular,
324 \q{Restart Game} does \e{not} destroy the \c{game_ui}.
326 \c{game_ui} is useful for implementing user-interface state which is
327 not part of \c{game_state}. Common examples are keyboard control
328 (you wouldn't want to have to separately Undo through every cursor
329 motion) and mouse dragging. See \k{writing-keyboard-cursor} and
330 \k{writing-howto-dragging}, respectively, for more details.
332 Another use for \c{game_ui} is to store highly persistent data such
333 as the Mines death counter. This is conceptually rather different:
334 where the Net cursor position was \e{not important enough} to
335 preserve for the player to restore by Undo, the Mines death counter
336 is \e{too important} to permit the player to revert by Undo!
338 A final use for \c{game_ui} is to pass information to the redraw
339 function about recent changes to the game state. This is used in
340 Mines, for example, to indicate whether a requested \q{flash} should
341 be a white flash for victory or a red flash for defeat; see
342 \k{writing-flash-types}.
344 \H{backend-simple} Simple data in the back end
346 In this section I begin to discuss each individual element in the
347 back end structure. To begin with, here are some simple
348 self-contained data elements.
350 \S{backend-name} \c{name}
354 This is a simple ASCII string giving the name of the puzzle. This
355 name will be used in window titles, in game selection menus on
356 monolithic platforms, and anywhere else that the front end needs to
357 know the name of a game.
359 \S{backend-winhelp} \c{winhelp_topic}
361 \c const char *winhelp_topic;
363 This member is used on Windows only, to provide online help.
364 Although the Windows front end provides a separate binary for each
365 puzzle, it has a single monolithic help file; so when a user selects
366 \q{Help} from the menu, the program needs to open the help file and
367 jump to the chapter describing that particular puzzle.
369 Therefore, each chapter in \c{puzzles.but} is labelled with a
370 \e{help topic} name, similar to this:
372 \c \cfg{winhelp-topic}{games.net}
374 And then the corresponding game back end encodes the topic string
375 (here \cq{games.net}) in the \c{winhelp_topic} element of the game
378 \H{backend-params} Handling game parameter sets
380 In this section I present the various functions which handle the
381 \c{game_params} structure.
383 \S{backend-default-params} \cw{default_params()}
385 \c game_params *(*default_params)(void);
387 This function allocates a new \c{game_params} structure, fills it
388 with the default values, and returns a pointer to it.
390 \S{backend-fetch-preset} \cw{fetch_preset()}
392 \c int (*fetch_preset)(int i, char **name, game_params **params);
394 This function is used to populate the \q{Type} menu, which provides
395 a list of conveniently accessible preset parameters for most games.
397 The function is called with \c{i} equal to the index of the preset
398 required (numbering from zero). It returns \cw{FALSE} if that preset
399 does not exist (if \c{i} is less than zero or greater than the
400 largest preset index). Otherwise, it sets \c{*params} to point at a
401 newly allocated \c{game_params} structure containing the preset
402 information, sets \c{*name} to point at a newly allocated C string
403 containing the preset title (to go on the \q{Type} menu), and
406 If the game does not wish to support any presets at all, this
407 function is permitted to return \cw{FALSE} always.
409 \S{backend-encode-params} \cw{encode_params()}
411 \c char *(*encode_params)(game_params *params, int full);
413 The job of this function is to take a \c{game_params}, and encode it
414 in a string form for use in game IDs. The return value must be a
415 newly allocated C string, and \e{must} not contain a colon or a hash
416 (since those characters are used to mark the end of the parameter
417 section in a game ID).
419 Ideally, it should also not contain any other potentially
420 controversial punctuation; bear in mind when designing a string
421 parameter format that it will probably be used on both Windows and
422 Unix command lines under a variety of exciting shell quoting and
423 metacharacter rules. Sticking entirely to alphanumerics is the
424 safest thing; if you really need punctuation, you can probably get
425 away with commas, periods or underscores without causing anybody any
426 major inconvenience. If you venture far beyond that, you're likely
427 to irritate \e{somebody}.
429 (At the time of writing this, all existing games have purely
430 alphanumeric string parameter formats. Usually these involve a
431 letter denoting a parameter, followed optionally by a number giving
432 the value of that parameter, with a few mandatory parts at the
433 beginning such as numeric width and height separated by \cq{x}.)
435 If the \c{full} parameter is \cw{TRUE}, this function should encode
436 absolutely everything in the \c{game_params}, such that a subsequent
437 call to \cw{decode_params()} (\k{backend-decode-params}) will yield
438 an identical structure. If \c{full} is \cw{FALSE}, however, you
439 should leave out anything which is not necessary to describe a
440 \e{specific puzzle instance}, i.e. anything which only takes effect
441 when a new puzzle is \e{generated}. For example, the Solo
442 \c{game_params} includes a difficulty rating used when constructing
443 new puzzles; but a Solo game ID need not explicitly include the
444 difficulty, since to describe a puzzle once generated it's
445 sufficient to give the grid dimensions and the location and contents
446 of the clue squares. (Indeed, one might very easily type in a puzzle
447 out of a newspaper without \e{knowing} what its difficulty level is
448 in Solo's terminology.) Therefore, Solo's \cw{encode_params()} only
449 encodes the difficulty level if \c{full} is set.
451 \S{backend-decode-params} \cw{decode_params()}
453 \c void (*decode_params)(game_params *params, char const *string);
455 This function is the inverse of \cw{encode_params()}
456 (\k{backend-encode-params}). It parses the supplied string and fills
457 in the supplied \c{game_params} structure. Note that the structure
458 will \e{already} have been allocated: this function is not expected
459 to create a \e{new} \c{game_params}, but to modify an existing one.
461 This function can receive a string which only encodes a subset of
462 the parameters. The most obvious way in which this can happen is if
463 the string was constructed by \cw{encode_params()} with its \c{full}
464 parameter set to \cw{FALSE}; however, it could also happen if the
465 user typed in a parameter set manually and missed something out. Be
466 prepared to deal with a wide range of possibilities.
468 When dealing with a parameter which is not specified in the input
469 string, what to do requires a judgment call on the part of the
470 programmer. Sometimes it makes sense to adjust other parameters to
471 bring them into line with the new ones. In Mines, for example, you
472 would probably not want to keep the same mine count if the user
473 dropped the grid size and didn't specify one, since you might easily
474 end up with more mines than would actually fit in the grid! On the
475 other hand, sometimes it makes sense to leave the parameter alone: a
476 Solo player might reasonably expect to be able to configure size and
477 difficulty independently of one another.
479 This function currently has no direct means of returning an error if
480 the string cannot be parsed at all. However, the returned
481 \c{game_params} is almost always subsequently passed to
482 \cw{validate_params()} (\k{backend-validate-params}), so if you
483 really want to signal parse errors, you could always have a \c{char
484 *} in your parameters structure which stored an error message, and
485 have \cw{validate_params()} return it if it is non-\cw{NULL}.
487 \S{backend-free-params} \cw{free_params()}
489 \c void (*free_params)(game_params *params);
491 This function frees a \c{game_params} structure, and any subsidiary
492 allocations contained within it.
494 \S{backend-dup-params} \cw{dup_params()}
496 \c game_params *(*dup_params)(game_params *params);
498 This function allocates a new \c{game_params} structure and
499 initialises it with an exact copy of the information in the one
500 provided as input. It returns a pointer to the new duplicate.
502 \S{backend-can-configure} \c{can_configure}
504 \c int can_configure;
506 This boolean data element is set to \cw{TRUE} if the back end
507 supports custom parameter configuration via a dialog box. If it is
508 \cw{TRUE}, then the functions \cw{configure()} and
509 \cw{custom_params()} are expected to work. See \k{backend-configure}
510 and \k{backend-custom-params} for more details.
512 \S{backend-configure} \cw{configure()}
514 \c config_item *(*configure)(game_params *params);
516 This function is called when the user requests a dialog box for
517 custom parameter configuration. It returns a newly allocated array
518 of \cw{config_item} structures, describing the GUI elements required
519 in the dialog box. The array should have one more element than the
520 number of controls, since it is terminated with a \cw{C_END} marker
521 (see below). Each array element describes the control together with
522 its initial value; the front end will modify the value fields and
523 return the updated array to \cw{custom_params()} (see
524 \k{backend-custom-params}).
526 The \cw{config_item} structure contains the following elements:
533 \c{name} is an ASCII string giving the textual label for a GUI
534 control. It is \e{not} expected to be dynamically allocated.
536 \c{type} contains one of a small number of \c{enum} values defining
537 what type of control is being described. The meaning of the \c{sval}
538 and \c{ival} fields depends on the value in \c{type}. The valid
543 \dd Describes a text input box. (This is also used for numeric
544 input. The back end does not bother informing the front end that the
545 box is numeric rather than textual; some front ends do have the
546 capacity to take this into account, but I decided it wasn't worth
547 the extra complexity in the interface.) For this type, \c{ival} is
548 unused, and \c{sval} contains a dynamically allocated string
549 representing the contents of the input box.
553 \dd Describes a simple checkbox. For this type, \c{sval} is unused,
554 and \c{ival} is \cw{TRUE} or \cw{FALSE}.
558 \dd Describes a drop-down list presenting one of a small number of
559 fixed choices. For this type, \c{sval} contains a list of strings
560 describing the choices; the very first character of \c{sval} is used
561 as a delimiter when processing the rest (so that the strings
562 \cq{:zero:one:two}, \cq{!zero!one!two} and \cq{xzeroxonextwo} all
563 define a three-element list containing \cq{zero}, \cq{one} and
564 \cq{two}). \c{ival} contains the index of the currently selected
565 element, numbering from zero (so that in the above example, 0 would
566 mean \cq{zero} and 2 would mean \cq{two}).
570 Note that for this control type, \c{sval} is \e{not} dynamically
571 allocated, whereas it was for \c{C_STRING}.
577 \dd Marks the end of the array of \c{config_item}s. All other fields
580 The array returned from this function is expected to have filled in
581 the initial values of all the controls according to the input
582 \c{game_params} structure.
584 If the game's \c{can_configure} flag is set to \cw{FALSE}, this
585 function is never called and need not do anything at all.
587 \S{backend-custom-params} \cw{custom_params()}
589 \c game_params *(*custom_params)(config_item *cfg);
591 This function is the counterpart to \cw{configure()}
592 (\k{backend-configure}). It receives as input an array of
593 \c{config_item}s which was originally created by \cw{configure()},
594 but in which the control values have since been changed in
595 accordance with user input. Its function is to read the new values
596 out of the controls and return a newly allocated \c{game_params}
597 structure representing the user's chosen parameter set.
599 (The front end will have modified the controls' \e{values}, but
600 there will still always be the same set of controls, in the same
601 order, as provided by \cw{configure()}. It is not necessary to check
602 the \c{name} and \c{type} fields, although you could use
603 \cw{assert()} if you were feeling energetic.)
605 This function is not expected to (and indeed \e{must not}) free the
606 input \c{config_item} array. (If the parameters fail to validate,
607 the dialog box will stay open.)
609 If the game's \c{can_configure} flag is set to \cw{FALSE}, this
610 function is never called and need not do anything at all.
612 \S{backend-validate-params} \cw{validate_params()}
614 \c char *(*validate_params)(game_params *params, int full);
616 This function takes a \c{game_params} structure as input, and checks
617 that the parameters described in it fall within sensible limits. (At
618 the very least, grid dimensions should almost certainly be strictly
619 positive, for example.)
621 Return value is \cw{NULL} if no problems were found, or
622 alternatively a (non-dynamically-allocated) ASCII string describing
623 the error in human-readable form.
625 If the \c{full} parameter is set, full validation should be
626 performed: any set of parameters which would not permit generation
627 of a sensible puzzle should be faulted. If \c{full} is \e{not} set,
628 the implication is that these parameters are not going to be used
629 for \e{generating} a puzzle; so parameters which can't even sensibly
630 \e{describe} a valid puzzle should still be faulted, but parameters
631 which only affect puzzle generation should not be.
633 (The \c{full} option makes a difference when parameter combinations
634 are non-orthogonal. For example, Net has a boolean option
635 controlling whether it enforces a unique solution; it turns out that
636 it's impossible to generate a uniquely soluble puzzle with wrapping
637 walls and width 2, so \cw{validate_params()} will complain if you
638 ask for one. However, if the user had just been playing a unique
639 wrapping puzzle of a more sensible width, and then pastes in a game
640 ID acquired from somebody else which happens to describe a
641 \e{non}-unique wrapping width-2 puzzle, then \cw{validate_params()}
642 will be passed a \c{game_params} containing the width and wrapping
643 settings from the new game ID and the uniqueness setting from the
644 old one. This would be faulted, if it weren't for the fact that
645 \c{full} is not set during this call, so Net ignores the
646 inconsistency. The resulting \c{game_params} is never subsequently
647 used to generate a puzzle; this is a promise made by the mid-end
648 when it asks for a non-full validation.)
650 \H{backend-descs} Handling game descriptions
652 In this section I present the functions that deal with a textual
653 description of a puzzle, i.e. the part that comes after the colon in
654 a descriptive-format game ID.
656 \S{backend-new-desc} \cw{new_desc()}
658 \c char *(*new_desc)(game_params *params, random_state *rs,
659 \c char **aux, int interactive);
661 This function is where all the really hard work gets done. This is
662 the function whose job is to randomly generate a new puzzle,
663 ensuring solubility and uniqueness as appropriate.
665 As input it is given a \c{game_params} structure and a random state
666 (see \k{utils-random} for the random number API). It must invent a
667 puzzle instance, encode it in string form, and return a dynamically
668 allocated C string containing that encoding.
670 Additionally, it may return a second dynamically allocated string in
671 \c{*aux}. (If it doesn't want to, then it can leave that parameter
672 completely alone; it isn't required to set it to \cw{NULL}, although
673 doing so is harmless.) That string, if present, will be passed to
674 \cw{solve()} (\k{backend-solve}) later on; so if the puzzle is
675 generated in such a way that a solution is known, then information
676 about that solution can be saved in \c{*aux} for \cw{solve()} to
679 The \c{interactive} parameter should be ignored by almost all
680 puzzles. Its purpose is to distinguish between generating a puzzle
681 within a GUI context for immediate play, and generating a puzzle in
682 a command-line context for saving to be played later. The only
683 puzzle that currently uses this distinction (and, I fervently hope,
684 the only one which will \e{ever} need to use it) is Mines, which
685 chooses a random first-click location when generating puzzles
686 non-interactively, but which waits for the user to place the first
687 click when interactive. If you think you have come up with another
688 puzzle which needs to make use of this parameter, please think for
689 at least ten minutes about whether there is \e{any} alternative!
691 Note that game description strings are not required to contain an
692 encoding of parameters such as grid size; a game description is
693 never separated from the \c{game_params} it was generated with, so
694 any information contained in that structure need not be encoded
695 again in the game description.
697 \S{backend-validate-desc} \cw{validate_desc()}
699 \c char *(*validate_desc)(game_params *params, char *desc);
701 This function is given a game description, and its job is to
702 validate that it describes a puzzle which makes sense.
704 To some extent it's up to the user exactly how far they take the
705 phrase \q{makes sense}; there are no particularly strict rules about
706 how hard the user is permitted to shoot themself in the foot when
707 typing in a bogus game description by hand. (For example, Rectangles
708 will not verify that the sum of all the numbers in the grid equals
709 the grid's area. So a user could enter a puzzle which was provably
710 not soluble, and the program wouldn't complain; there just wouldn't
711 happen to be any sequence of moves which solved it.)
713 The one non-negotiable criterion is that any game description which
714 makes it through \cw{validate_desc()} \e{must not} subsequently
715 cause a crash or an assertion failure when fed to \cw{new_game()}
716 and thence to the rest of the back end.
718 The return value is \cw{NULL} on success, or a
719 non-dynamically-allocated C string containing an error message.
721 \S{backend-new-game} \cw{new_game()}
723 \c game_state *(*new_game)(midend *me, game_params *params,
726 This function takes a game description as input, together with its
727 accompanying \c{game_params}, and constructs a \c{game_state}
728 describing the initial state of the puzzle. It returns a newly
729 allocated \c{game_state} structure.
731 Almost all puzzles should ignore the \c{me} parameter. It is
732 required by Mines, which needs it for later passing to
733 \cw{midend_supersede_game_desc()} (see \k{backend-supersede}) once
734 the user has placed the first click. I fervently hope that no other
735 puzzle will be awkward enough to require it, so everybody else
736 should ignore it. As with the \c{interactive} parameter in
737 \cw{new_desc()} (\k{backend-new-desc}), if you think you have a
738 reason to need this parameter, please try very hard to think of an
739 alternative approach!
741 \H{backend-states} Handling game states
743 This section describes the functions which create and destroy
744 \c{game_state} structures.
746 (Well, except \cw{new_game()}, which is in \k{backend-new-game}
747 instead of under here; but it deals with game descriptions \e{and}
748 game states and it had to go in one section or the other.)
750 \S{backend-dup-game} \cw{dup_game()}
752 \c game_state *(*dup_game)(game_state *state);
754 This function allocates a new \c{game_state} structure and
755 initialises it with an exact copy of the information in the one
756 provided as input. It returns a pointer to the new duplicate.
758 \S{backend-free-game} \cw{free_game()}
760 \c void (*free_game)(game_state *state);
762 This function frees a \c{game_state} structure, and any subsidiary
763 allocations contained within it.
765 \H{backend-ui} Handling \c{game_ui}
767 \S{backend-new-ui} \cw{new_ui()}
769 \c game_ui *(*new_ui)(game_state *state);
771 This function allocates and returns a new \c{game_ui} structure for
772 playing a particular puzzle. It is passed a pointer to the initial
773 \c{game_state}, in case it needs to refer to that when setting up
774 the initial values for the new game.
776 \S{backend-free-ui} \cw{free_ui()}
778 \c void (*free_ui)(game_ui *ui);
780 This function frees a \c{game_ui} structure, and any subsidiary
781 allocations contained within it.
783 \S{backend-encode-ui} \cw{encode_ui()}
785 \c char *(*encode_ui)(game_ui *ui);
787 This function encodes any \e{important} data in a \c{game_ui}
788 structure in string form. It is only called when saving a
789 half-finished game to a file.
791 It should be used sparingly. Almost all data in a \c{game_ui} is not
792 important enough to save. The location of the keyboard-controlled
793 cursor, for example, can be reset to a default position on reloading
794 the game without impacting the user experience. If the user should
795 somehow manage to save a game while a mouse drag was in progress,
796 then discarding that mouse drag would be an outright \e{feature}.
798 A typical thing that \e{would} be worth encoding in this function is
799 the Mines death counter: it's in the \c{game_ui} rather than the
800 \c{game_state} because it's too important to allow the user to
801 revert it by using Undo, and therefore it's also too important to
802 allow the user to revert it by saving and reloading. (Of course, the
803 user could edit the save file by hand... But if the user is \e{that}
804 determined to cheat, they could just as easily modify the game's
807 \S{backend-decode-ui} \cw{decode_ui()}
809 \c void (*decode_ui)(game_ui *ui, char *encoding);
811 This function parses a string previously output by \cw{encode_ui()},
812 and writes the decoded data back into the provided \c{game_ui}
815 \S{backend-changed-state} \cw{changed_state()}
817 \c void (*changed_state)(game_ui *ui, game_state *oldstate,
818 \c game_state *newstate);
820 This function is called by the mid-end whenever the current game
821 state changes, for any reason. Those reasons include:
823 \b a fresh move being made by \cw{interpret_move()} and
826 \b a solve operation being performed by \cw{solve()} and
829 \b the user moving back and forth along the undo list by means of
830 the Undo and Redo operations
832 \b the user selecting Restart to go back to the initial game state.
834 The job of \cw{changed_state()} is to update the \c{game_ui} for
835 consistency with the new game state, if any update is necessary. For
836 example, Same Game stores data about the currently selected tile
837 group in its \c{game_ui}, and this data is intrinsically related to
838 the game state it was derived from. So it's very likely to become
839 invalid when the game state changes; thus, Same Game's
840 \cw{changed_state()} function clears the current selection whenever
843 When \cw{anim_length()} or \cw{flash_length()} are called, you can
844 be sure that there has been a previous call to \cw{changed_state()}.
845 So \cw{changed_state()} can set up data in the \c{game_ui} which will
846 be read by \cw{anim_length()} and \cw{flash_length()}, and those
847 functions will not have to worry about being called without the data
848 having been initialised.
850 \H{backend-moves} Making moves
852 This section describes the functions which actually make moves in
853 the game: that is, the functions which process user input and end up
854 producing new \c{game_state}s.
856 \S{backend-interpret-move} \cw{interpret_move()}
858 \c char *(*interpret_move)(game_state *state, game_ui *ui,
859 \c game_drawstate *ds,
860 \c int x, int y, int button);
862 This function receives user input and processes it. Its input
863 parameters are the current \c{game_state}, the current \c{game_ui}
864 and the current \c{game_drawstate}, plus details of the input event.
865 \c{button} is either an ASCII value or a special code (listed below)
866 indicating an arrow or function key or a mouse event; when
867 \c{button} is a mouse event, \c{x} and \c{y} contain the pixel
868 coordinates of the mouse pointer relative to the top left of the
869 puzzle's drawing area.
871 \cw{interpret_move()} may return in three different ways:
873 \b Returning \cw{NULL} indicates that no action whatsoever occurred
874 in response to the input event; the puzzle was not interested in it
877 \b Returning the empty string (\cw{""}) indicates that the input
878 event has resulted in a change being made to the \c{game_ui} which
879 will require a redraw of the game window, but that no actual
880 \e{move} was made (i.e. no new \c{game_state} needs to be created).
882 \b Returning anything else indicates that a move was made and that a
883 new \c{game_state} must be created. However, instead of actually
884 constructing a new \c{game_state} itself, this function is required
885 to return a string description of the details of the move. This
886 string will be passed to \cw{execute_move()}
887 (\k{backend-execute-move}) to actually create the new
888 \c{game_state}. (Encoding moves as strings in this way means that
889 the mid-end can keep the strings as well as the game states, and the
890 strings can be written to disk when saving the game and fed to
891 \cw{execute_move()} again on reloading.)
893 The return value from \cw{interpret_move()} is expected to be
894 dynamically allocated if and only if it is not either \cw{NULL}
895 \e{or} the empty string.
897 After this function is called, the back end is permitted to rely on
898 some subsequent operations happening in sequence:
900 \b \cw{execute_move()} will be called to convert this move
901 description into a new \c{game_state}
903 \b \cw{changed_state()} will be called with the new \c{game_state}.
905 This means that if \cw{interpret_move()} needs to do updates to the
906 \c{game_ui} which are easier to perform by referring to the new
907 \c{game_state}, it can safely leave them to be done in
908 \cw{changed_state()} and not worry about them failing to happen.
910 (Note, however, that \cw{execute_move()} may \e{also} be called in
911 other circumstances. It is only \cw{interpret_move()} which can rely
912 on a subsequent call to \cw{changed_state()}.)
914 The special key codes supported by this function are:
916 \dt \cw{LEFT_BUTTON}, \cw{MIDDLE_BUTTON}, \cw{RIGHT_BUTTON}
918 \dd Indicate that one of the mouse buttons was pressed down.
920 \dt \cw{LEFT_DRAG}, \cw{MIDDLE_DRAG}, \cw{RIGHT_DRAG}
922 \dd Indicate that the mouse was moved while one of the mouse buttons
923 was still down. The mid-end guarantees that when one of these events
924 is received, it will always have been preceded by a button-down
925 event (and possibly other drag events) for the same mouse button,
926 and no event involving another mouse button will have appeared in
929 \dt \cw{LEFT_RELEASE}, \cw{MIDDLE_RELEASE}, \cw{RIGHT_RELEASE}
931 \dd Indicate that a mouse button was released. The mid-end
932 guarantees that when one of these events is received, it will always
933 have been preceded by a button-down event (and possibly some drag
934 events) for the same mouse button, and no event involving another
935 mouse button will have appeared in between.
937 \dt \cw{CURSOR_UP}, \cw{CURSOR_DOWN}, \cw{CURSOR_LEFT},
940 \dd Indicate that an arrow key was pressed.
942 \dt \cw{CURSOR_SELECT}
944 \dd On platforms which have a prominent \q{select} button alongside
945 their cursor keys, indicates that that button was pressed.
947 In addition, there are some modifiers which can be bitwise-ORed into
948 the \c{button} parameter:
950 \dt \cw{MOD_CTRL}, \cw{MOD_SHFT}
952 \dd These indicate that the Control or Shift key was pressed
953 alongside the key. They only apply to the cursor keys, not to mouse
954 buttons or anything else.
956 \dt \cw{MOD_NUM_KEYPAD}
958 \dd This applies to some ASCII values, and indicates that the key
959 code was input via the numeric keypad rather than the main keyboard.
960 Some puzzles may wish to treat this differently (for example, a
961 puzzle might want to use the numeric keypad as an eight-way
962 directional pad), whereas others might not (a game involving numeric
963 input probably just wants to treat the numeric keypad as numbers).
967 \dd This mask is the bitwise OR of all the available modifiers; you
968 can bitwise-AND with \cw{~MOD_MASK} to strip all the modifiers off
971 \S{backend-execute-move} \cw{execute_move()}
973 \c game_state *(*execute_move)(game_state *state, char *move);
975 This function takes an input \c{game_state} and a move string as
976 output from \cw{interpret_move()}. It returns a newly allocated
977 \c{game_state} which contains the result of applying the specified
978 move to the input game state.
980 This function may return \cw{NULL} if it cannot parse the move
981 string (and this is definitely preferable to crashing or failing an
982 assertion, since one way this can happen is if loading a corrupt
983 save file). However, it must not return \cw{NULL} for any move
984 string that really was output from \cw{interpret_move()}: this is
985 punishable by assertion failure in the mid-end.
987 \S{backend-can-solve} \c{can_solve}
991 This boolean field is set to \cw{TRUE} if the game's \cw{solve()}
992 function does something. If it's set to \cw{FALSE}, the game will
993 not even offer the \q{Solve} menu option.
995 \S{backend-solve} \cw{solve()}
997 \c char *(*solve)(game_state *orig, game_state *curr,
998 \c char *aux, char **error);
1000 This function is called when the user selects the \q{Solve} option
1003 It is passed two input game states: \c{orig} is the game state from
1004 the very start of the puzzle, and \c{curr} is the current one.
1005 (Different games find one or other or both of these convenient.) It
1006 is also passed the \c{aux} string saved by \cw{new_desc()}
1007 (\k{backend-new-desc}), in case that encodes important information
1008 needed to provide the solution.
1010 If this function is unable to produce a solution (perhaps, for
1011 example, the game has no in-built solver so it can only solve
1012 puzzles it invented internally and has an \c{aux} string for) then
1013 it may return \cw{NULL}. If it does this, it must also set
1014 \c{*error} to an error message to be presented to the user (such as
1015 \q{Solution not known for this puzzle}); that error message is not
1016 expected to be dynamically allocated.
1018 If this function \e{does} produce a solution, it returns a move
1019 string suitable for feeding to \cw{execute_move()}
1020 (\k{backend-execute-move}).
1022 \H{backend-drawing} Drawing the game graphics
1024 This section discusses the back end functions that deal with
1027 \S{backend-new-drawstate} \cw{new_drawstate()}
1029 \c game_drawstate *(*new_drawstate)(drawing *dr, game_state *state);
1031 This function allocates and returns a new \c{game_drawstate}
1032 structure for drawing a particular puzzle. It is passed a pointer to
1033 a \c{game_state}, in case it needs to refer to that when setting up
1036 This function may not rely on the puzzle having been newly started;
1037 a new draw state can be constructed at any time if the front end
1038 requests a forced redraw. For games like Pattern, in which initial
1039 game states are much simpler than general ones, this might be
1040 important to keep in mind.
1042 The parameter \c{dr} is a drawing object (see \k{drawing}) which the
1043 function might need to use to allocate blitters. (However, this
1044 isn't recommended; it's usually more sensible to wait to allocate a
1045 blitter until \cw{set_size()} is called, because that way you can
1046 tailor it to the scale at which the puzzle is being drawn.)
1048 \S{backend-free-drawstate} \cw{free_drawstate()}
1050 \c void (*free_drawstate)(drawing *dr, game_drawstate *ds);
1052 This function frees a \c{game_drawstate} structure, and any
1053 subsidiary allocations contained within it.
1055 The parameter \c{dr} is a drawing object (see \k{drawing}), which
1056 might be required if you are freeing a blitter.
1058 \S{backend-preferred-tilesize} \c{preferred_tilesize}
1060 \c int preferred_tilesize;
1062 Each game is required to define a single integer parameter which
1063 expresses, in some sense, the scale at which it is drawn. This is
1064 described in the APIs as \cq{tilesize}, since most puzzles are on a
1065 square (or possibly triangular or hexagonal) grid and hence a
1066 sensible interpretation of this parameter is to define it as the
1067 size of one grid tile in pixels; however, there's no actual
1068 requirement that the \q{tile size} be proportional to the game
1069 window size. Window size is required to increase monotonically with
1070 \q{tile size}, however.
1072 The data element \c{preferred_tilesize} indicates the tile size
1073 which should be used in the absence of a good reason to do otherwise
1074 (such as the screen being too small, or the user explicitly
1075 requesting a resize if that ever gets implemented).
1077 \S{backend-compute-size} \cw{compute_size()}
1079 \c void (*compute_size)(game_params *params, int tilesize,
1082 This function is passed a \c{game_params} structure and a tile size.
1083 It returns, in \c{*x} and \c{*y}, the size in pixels of the drawing
1084 area that would be required to render a puzzle with those parameters
1087 \S{backend-set-size} \cw{set_size()}
1089 \c void (*set_size)(drawing *dr, game_drawstate *ds,
1090 \c game_params *params, int tilesize);
1092 This function is responsible for setting up a \c{game_drawstate} to
1093 draw at a given tile size. Typically this will simply involve
1094 copying the supplied \c{tilesize} parameter into a \c{tilesize}
1095 field inside the draw state; for some more complex games it might
1096 also involve setting up other dimension fields, or possibly
1097 allocating a blitter (see \k{drawing-blitter}).
1099 The parameter \c{dr} is a drawing object (see \k{drawing}), which is
1100 required if a blitter needs to be allocated.
1102 Back ends may assume (and may enforce by assertion) that this
1103 function will be called at most once for any \c{game_drawstate}. If
1104 a puzzle needs to be redrawn at a different size, the mid-end will
1105 create a fresh drawstate.
1107 \S{backend-colours} \cw{colours()}
1109 \c float *(*colours)(frontend *fe, int *ncolours);
1111 This function is responsible for telling the front end what colours
1112 the puzzle will need to draw itself.
1114 It returns the number of colours required in \c{*ncolours}, and the
1115 return value from the function itself is a dynamically allocated
1116 array of three times that many \c{float}s, containing the red, green
1117 and blue components of each colour respectively as numbers in the
1120 The second parameter passed to this function is a front end handle.
1121 The only things it is permitted to do with this handle are to call
1122 the front-end function called \cw{frontend_default_colour()} (see
1123 \k{frontend-default-colour}) or the utility function called
1124 \cw{game_mkhighlight()} (see \k{utils-game-mkhighlight}). (The
1125 latter is a wrapper on the former, so front end implementors only
1126 need to provide \cw{frontend_default_colour()}.) This allows
1127 \cw{colours()} to take local configuration into account when
1128 deciding on its own colour allocations. Most games use the front
1129 end's default colour as their background, apart from a few which
1130 depend on drawing relief highlights so they adjust the background
1131 colour if it's too light for highlights to show up against it.
1133 Note that the colours returned from this function are for
1134 \e{drawing}, not for printing. Printing has an entirely different
1135 colour allocation policy.
1137 \S{backend-anim-length} \cw{anim_length()}
1139 \c float (*anim_length)(game_state *oldstate, game_state *newstate,
1140 \c int dir, game_ui *ui);
1142 This function is called when a move is made, undone or redone. It is
1143 given the old and the new \c{game_state}, and its job is to decide
1144 whether the transition between the two needs to be animated or can
1147 \c{oldstate} is the state that was current until this call;
1148 \c{newstate} is the state that will be current after it. \c{dir}
1149 specifies the chronological order of those states: if it is
1150 positive, then the transition is the result of a move or a redo (and
1151 so \c{newstate} is the later of the two moves), whereas if it is
1152 negative then the transition is the result of an undo (so that
1153 \c{newstate} is the \e{earlier} move).
1155 If this function decides the transition should be animated, it
1156 returns the desired length of the animation in seconds. If not, it
1159 State changes as a result of a Restart operation are never animated;
1160 the mid-end will handle them internally and never consult this
1161 function at all. State changes as a result of Solve operations are
1162 also not animated by default, although you can change this for a
1163 particular game by setting a flag in \c{flags} (\k{backend-flags}).
1165 The function is also passed a pointer to the local \c{game_ui}. It
1166 may refer to information in here to help with its decision (see
1167 \k{writing-conditional-anim} for an example of this), and/or it may
1168 \e{write} information about the nature of the animation which will
1169 be read later by \cw{redraw()}.
1171 When this function is called, it may rely on \cw{changed_state()}
1172 having been called previously, so if \cw{anim_length()} needs to
1173 refer to information in the \c{game_ui}, then \cw{changed_state()}
1174 is a reliable place to have set that information up.
1176 Move animations do not inhibit further input events. If the user
1177 continues playing before a move animation is complete, the animation
1178 will be abandoned and the display will jump straight to the final
1181 \S{backend-flash-length} \cw{flash_length()}
1183 \c float (*flash_length)(game_state *oldstate, game_state *newstate,
1184 \c int dir, game_ui *ui);
1186 This function is called when a move is completed. (\q{Completed}
1187 means that not only has the move been made, but any animation which
1188 accompanied it has finished.) It decides whether the transition from
1189 \c{oldstate} to \c{newstate} merits a \q{flash}.
1191 A flash is much like a move animation, but it is \e{not} interrupted
1192 by further user interface activity; it runs to completion in
1193 parallel with whatever else might be going on on the display. The
1194 only thing which will rush a flash to completion is another flash.
1196 The purpose of flashes is to indicate that the game has been
1197 completed. They were introduced as a separate concept from move
1198 animations because of Net: the habit of most Net players (and
1199 certainly me) is to rotate a tile into place and immediately lock
1200 it, then move on to another tile. When you make your last move, at
1201 the instant the final tile is rotated into place the screen starts
1202 to flash to indicate victory \dash but if you then press the lock
1203 button out of habit, then the move animation is cancelled, and the
1204 victory flash does not complete. (And if you \e{don't} press the
1205 lock button, the completed grid will look untidy because there will
1206 be one unlocked square.) Therefore, I introduced a specific concept
1207 of a \q{flash} which is separate from a move animation and can
1208 proceed in parallel with move animations and any other display
1209 activity, so that the victory flash in Net is not cancelled by that
1212 The input parameters to \cw{flash_length()} are exactly the same as
1213 the ones to \cw{anim_length()}.
1215 Just like \cw{anim_length()}, when this function is called, it may
1216 rely on \cw{changed_state()} having been called previously, so if it
1217 needs to refer to information in the \c{game_ui} then
1218 \cw{changed_state()} is a reliable place to have set that
1221 (Some games use flashes to indicate defeat as well as victory;
1222 Mines, for example, flashes in a different colour when you tread on
1223 a mine from the colour it uses when you complete the game. In order
1224 to achieve this, its \cw{flash_length()} function has to store a
1225 flag in the \c{game_ui} to indicate which flash type is required.)
1227 \S{backend-redraw} \cw{redraw()}
1229 \c void (*redraw)(drawing *dr, game_drawstate *ds,
1230 \c game_state *oldstate, game_state *newstate, int dir,
1231 \c game_ui *ui, float anim_time, float flash_time);
1233 This function is responsible for actually drawing the contents of
1234 the game window, and for redrawing every time the game state or the
1235 \c{game_ui} changes.
1237 The parameter \c{dr} is a drawing object which may be passed to the
1238 drawing API functions (see \k{drawing} for documentation of the
1239 drawing API). This function may not save \c{dr} and use it
1240 elsewhere; it must only use it for calling back to the drawing API
1241 functions within its own lifetime.
1243 \c{ds} is the local \c{game_drawstate}, of course, and \c{ui} is the
1246 \c{newstate} is the semantically-current game state, and is always
1247 non-\cw{NULL}. If \c{oldstate} is also non-\cw{NULL}, it means that
1248 a move has recently been made and the game is still in the process
1249 of displaying an animation linking the old and new states; in this
1250 situation, \c{anim_time} will give the length of time (in seconds)
1251 that the animation has already been running. If \c{oldstate} is
1252 \cw{NULL}, then \c{anim_time} is unused (and will hopefully be set
1253 to zero to avoid confusion).
1255 \c{flash_time}, if it is is non-zero, denotes that the game is in
1256 the middle of a flash, and gives the time since the start of the
1257 flash. See \k{backend-flash-length} for general discussion of
1260 The very first time this function is called for a new
1261 \c{game_drawstate}, it is expected to redraw the \e{entire} drawing
1262 area. Since this often involves drawing visual furniture which is
1263 never subsequently altered, it is often simplest to arrange this by
1264 having a special \q{first time} flag in the draw state, and
1265 resetting it after the first redraw.
1267 When this function (or any subfunction) calls the drawing API, it is
1268 expected to pass colour indices which were previously defined by the
1269 \cw{colours()} function.
1271 \H{backend-printing} Printing functions
1273 This section discusses the back end functions that deal with
1274 printing puzzles out on paper.
1276 \S{backend-can-print} \c{can_print}
1280 This flag is set to \cw{TRUE} if the puzzle is capable of printing
1281 itself on paper. (This makes sense for some puzzles, such as Solo,
1282 which can be filled in with a pencil. Other puzzles, such as
1283 Twiddle, inherently involve moving things around and so would not
1284 make sense to print.)
1286 If this flag is \cw{FALSE}, then the functions \cw{print_size()}
1287 and \cw{print()} will never be called.
1289 \S{backend-can-print-in-colour} \c{can_print_in_colour}
1291 \c int can_print_in_colour;
1293 This flag is set to \cw{TRUE} if the puzzle is capable of printing
1294 itself differently when colour is available. For example, Map can
1295 actually print coloured regions in different \e{colours} rather than
1296 resorting to cross-hatching.
1298 If the \c{can_print} flag is \cw{FALSE}, then this flag will be
1301 \S{backend-print-size} \cw{print_size()}
1303 \c void (*print_size)(game_params *params, float *x, float *y);
1305 This function is passed a \c{game_params} structure and a tile size.
1306 It returns, in \c{*x} and \c{*y}, the preferred size in
1307 \e{millimetres} of that puzzle if it were to be printed out on paper.
1309 If the \c{can_print} flag is \cw{FALSE}, this function will never be
1312 \S{backend-print} \cw{print()}
1314 \c void (*print)(drawing *dr, game_state *state, int tilesize);
1316 This function is called when a puzzle is to be printed out on paper.
1317 It should use the drawing API functions (see \k{drawing}) to print
1320 This function is separate from \cw{redraw()} because it is often
1323 \b The printing function may not depend on pixel accuracy, since
1324 printer resolution is variable. Draw as if your canvas had infinite
1327 \b The printing function sometimes needs to display things in a
1328 completely different style. Net, for example, is very different as
1329 an on-screen puzzle and as a printed one.
1331 \b The printing function is often much simpler since it has no need
1332 to deal with repeated partial redraws.
1334 However, there's no reason the printing and redraw functions can't
1335 share some code if they want to.
1337 When this function (or any subfunction) calls the drawing API, the
1338 colour indices it passes should be colours which have been allocated
1339 by the \cw{print_*_colour()} functions within this execution of
1340 \cw{print()}. This is very different from the fixed small number of
1341 colours used in \cw{redraw()}, because printers do not have a
1342 limitation on the total number of colours that may be used. Some
1343 puzzles' printing functions might wish to allocate only one \q{ink}
1344 colour and use it for all drawing; others might wish to allocate
1345 \e{more} colours than are used on screen.
1347 One possible colour policy worth mentioning specifically is that a
1348 puzzle's printing function might want to allocate the \e{same}
1349 colour indices as are used by the redraw function, so that code
1350 shared between drawing and printing does not have to keep switching
1351 its colour indices. In order to do this, the simplest thing is to
1352 make use of the fact that colour indices returned from
1353 \cw{print_*_colour()} are guaranteed to be in increasing order from
1354 zero. So if you have declared an \c{enum} defining three colours
1355 \cw{COL_BACKGROUND}, \cw{COL_THIS} and \cw{COL_THAT}, you might then
1359 \c c = print_mono_colour(dr, 1); assert(c == COL_BACKGROUND);
1360 \c c = print_mono_colour(dr, 0); assert(c == COL_THIS);
1361 \c c = print_mono_colour(dr, 0); assert(c == COL_THAT);
1363 If the \c{can_print} flag is \cw{FALSE}, this function will never be
1366 \H{backend-misc} Miscellaneous
1368 \S{backend-can-format-as-text} \c{can_format_as_text}
1370 \c int can_format_as_text;
1372 This boolean field is \cw{TRUE} if the game supports formatting a
1373 game state as ASCII text (typically ASCII art) for copying to the
1374 clipboard and pasting into other applications. If it is \cw{FALSE},
1375 front ends will not offer the \q{Copy} command at all.
1377 If this field is \cw{FALSE}, the function \cw{text_format()}
1378 (\k{backend-text-format}) is not expected to do anything at all.
1380 \S{backend-text-format} \cw{text_format()}
1382 \c char *(*text_format)(game_state *state);
1384 This function is passed a \c{game_state}, and returns a newly
1385 allocated C string containing an ASCII representation of that game
1386 state. It is used to implement the \q{Copy} operation in many front
1389 This function should only be called if the back end field
1390 \c{can_format_as_text} (\k{backend-can-format-as-text}) is
1393 The returned string may contain line endings (and will probably want
1394 to), using the normal C internal \cq{\\n} convention. For
1395 consistency between puzzles, all multi-line textual puzzle
1396 representations should \e{end} with a newline as well as containing
1397 them internally. (There are currently no puzzles which have a
1398 one-line ASCII representation, so there's no precedent yet for
1399 whether that should come with a newline or not.)
1401 \S{backend-wants-statusbar} \cw{wants_statusbar()}
1403 \c int wants_statusbar;
1405 This boolean field is set to \cw{TRUE} if the puzzle has a use for a
1406 textual status line (to display score, completion status, currently
1409 \S{backend-is-timed} \c{is_timed}
1413 This boolean field is \cw{TRUE} if the puzzle is time-critical. If
1414 so, the mid-end will maintain a game timer while the user plays.
1416 If this field is \cw{FALSE}, then \cw{timing_state()} will never be
1417 called and need not do anything.
1419 \S{backend-timing-state} \cw{timing_state()}
1421 \c int (*timing_state)(game_state *state, game_ui *ui);
1423 This function is passed the current \c{game_state} and the local
1424 \c{game_ui}; it returns \cw{TRUE} if the game timer should currently
1427 A typical use for the \c{game_ui} in this function is to note when
1428 the game was first completed (by setting a flag in
1429 \cw{changed_state()} \dash see \k{backend-changed-state}), and
1430 freeze the timer thereafter so that the user can undo back through
1431 their solution process without altering their time.
1433 \S{backend-flags} \c{flags}
1437 This field contains miscellaneous per-backend flags. It consists of
1438 the bitwise OR of some combination of the following:
1440 \dt \cw{BUTTON_BEATS(x,y)}
1442 \dd Given any \cw{x} and \cw{y} from the set \{\cw{LEFT_BUTTON},
1443 \cw{MIDDLE_BUTTON}, \cw{RIGHT_BUTTON}\}, this macro evaluates to a
1444 bit flag which indicates that when buttons \cw{x} and \cw{y} are
1445 both pressed simultaneously, the mid-end should consider \cw{x} to
1446 have priority. (In the absence of any such flags, the mid-end will
1447 always consider the most recently pressed button to have priority.)
1449 \dt \cw{SOLVE_ANIMATES}
1451 \dd This flag indicates that moves generated by \cw{solve()}
1452 (\k{backend-solve}) are candidates for animation just like any other
1453 move. For most games, solve moves should not be animated, so the
1454 mid-end doesn't even bother calling \cw{anim_length()}
1455 (\k{backend-anim-length}), thus saving some special-case code in
1456 each game. On the rare occasion that animated solve moves are
1457 actually required, you can set this flag.
1459 \dt \cw{REQUIRE_RBUTTON}
1461 \dd This flag indicates that the puzzle cannot be usefully played
1462 without the use of mouse buttons other than the left one. On some
1463 PDA platforms, this flag is used by the front end to enable
1464 right-button emulation through an appropriate gesture. Note that a
1465 puzzle is not required to set this just because it \e{uses} the
1466 right button, but only if its use of the right button is critical to
1467 playing the game. (Slant, for example, uses the right button to
1468 cycle through the three square states in the opposite order from the
1469 left button, and hence can manage fine without it.)
1471 \dt \cw{REQUIRE_NUMPAD}
1473 \dd This flag indicates that the puzzle cannot be usefully played
1474 without the use of number-key input. On some PDA platforms it causes
1475 an emulated number pad to appear on the screen. Similarly to
1476 \cw{REQUIRE_RBUTTON}, a puzzle need not specify this simply if its
1477 use of the number keys is not critical.
1479 \H{backend-initiative} Things a back end may do on its own initiative
1481 This section describes a couple of things that a back end may choose
1482 to do by calling functions elsewhere in the program, which would not
1483 otherwise be obvious.
1485 \S{backend-newrs} Create a random state
1487 If a back end needs random numbers at some point during normal play,
1488 it can create a fresh \c{random_state} by first calling
1489 \c{get_random_seed} (\k{frontend-get-random-seed}) and then passing
1490 the returned seed data to \cw{random_new()}.
1492 This is likely not to be what you want. If a puzzle needs randomness
1493 in the middle of play, it's likely to be more sensible to store some
1494 sort of random state within the \c{game_state}, so that the random
1495 numbers are tied to the particular game state and hence the player
1496 can't simply keep undoing their move until they get numbers they
1499 This facility is currently used only in Net, to implement the
1500 \q{jumble} command, which sets every unlocked tile to a new random
1501 orientation. This randomness \e{is} a reasonable use of the feature,
1502 because it's non-adversarial \dash there's no advantage to the user
1503 in getting different random numbers.
1505 \S{backend-supersede} Supersede its own game description
1507 In response to a move, a back end is (reluctantly) permitted to call
1508 \cw{midend_supersede_game_desc()}:
1510 \c void midend_supersede_game_desc(midend *me,
1511 \c char *desc, char *privdesc);
1513 When the user selects \q{New Game}, the mid-end calls
1514 \cw{new_desc()} (\k{backend-new-desc}) to get a new game
1515 description, and (as well as using that to generate an initial game
1516 state) stores it for the save file and for telling to the user. The
1517 function above overwrites that game description, and also splits it
1518 in two. \c{desc} becomes the new game description which is provided
1519 to the user on request, and is also the one used to construct a new
1520 initial game state if the user selects \q{Restart}. \c{privdesc} is
1521 a \q{private} game description, used to reconstruct the game's
1522 initial state when reloading.
1524 The distinction between the two, as well as the need for this
1525 function at all, comes from Mines. Mines begins with a blank grid
1526 and no idea of where the mines actually are; \cw{new_desc()} does
1527 almost no work in interactive mode, and simply returns a string
1528 encoding the \c{random_state}. When the user first clicks to open a
1529 tile, \e{then} Mines generates the mine positions, in such a way
1530 that the game is soluble from that starting point. Then it uses this
1531 function to supersede the random-state game description with a
1532 proper one. But it needs two: one containing the initial click
1533 location (because that's what you want to happen if you restart the
1534 game, and also what you want to send to a friend so that they play
1535 \e{the same game} as you), and one without the initial click
1536 location (because when you save and reload the game, you expect to
1537 see the same blank initial state as you had before saving).
1539 I should stress again that this function is a horrid hack. Nobody
1540 should use it if they're not Mines; if you think you need to use it,
1541 think again repeatedly in the hope of finding a better way to do
1542 whatever it was you needed to do.
1544 \C{drawing} The drawing API
1546 The back end function \cw{redraw()} (\k{backend-redraw}) is required
1547 to draw the puzzle's graphics on the window's drawing area, or on
1548 paper if the puzzle is printable. To do this portably, it is
1549 provided with a drawing API allowing it to talk directly to the
1550 front end. In this chapter I document that API, both for the benefit
1551 of back end authors trying to use it and for front end authors
1552 trying to implement it.
1554 The drawing API as seen by the back end is a collection of global
1555 functions, each of which takes a pointer to a \c{drawing} structure
1556 (a \q{drawing object}). These objects are supplied as parameters to
1557 the back end's \cw{redraw()} and \cw{print()} functions.
1559 In fact these global functions are not implemented directly by the
1560 front end; instead, they are implemented centrally in \c{drawing.c}
1561 and form a small piece of middleware. The drawing API as supplied by
1562 the front end is a structure containing a set of function pointers,
1563 plus a \cq{void *} handle which is passed to each of those
1564 functions. This enables a single front end to switch between
1565 multiple implementations of the drawing API if necessary. For
1566 example, the Windows API supplies a printing mechanism integrated
1567 into the same GDI which deals with drawing in windows, and therefore
1568 the same API implementation can handle both drawing and printing;
1569 but on Unix, the most common way for applications to print is by
1570 producing PostScript output directly, and although it would be
1571 \e{possible} to write a single (say) \cw{draw_rect()} function which
1572 checked a global flag to decide whether to do GTK drawing operations
1573 or output PostScript to a file, it's much nicer to have two separate
1574 functions and switch between them as appropriate.
1576 When drawing, the puzzle window is indexed by pixel coordinates,
1577 with the top left pixel defined as \cw{(0,0)} and the bottom right
1578 pixel \cw{(w-1,h-1)}, where \c{w} and \c{h} are the width and height
1579 values returned by the back end function \cw{compute_size()}
1580 (\k{backend-compute-size}).
1582 When printing, the puzzle's print area is indexed in exactly the
1583 same way (with an arbitrary tile size provided by the printing
1584 module \c{printing.c}), to facilitate sharing of code between the
1585 drawing and printing routines. However, when printing, puzzles may
1586 no longer assume that the coordinate unit has any relationship to a
1587 pixel; the printer's actual resolution might very well not even be
1588 known at print time, so the coordinate unit might be smaller or
1589 larger than a pixel. Puzzles' print functions should restrict
1590 themselves to drawing geometric shapes rather than fiddly pixel
1593 \e{Puzzles' redraw functions may assume that the surface they draw
1594 on is persistent}. It is the responsibility of every front end to
1595 preserve the puzzle's window contents in the face of GUI window
1596 expose issues and similar. It is not permissible to request that the
1597 back end redraw any part of a window that it has already drawn,
1598 unless something has actually changed as a result of making moves in
1601 Most front ends accomplish this by having the drawing routines draw
1602 on a stored bitmap rather than directly on the window, and copying
1603 the bitmap to the window every time a part of the window needs to be
1604 redrawn. Therefore, it is vitally important that whenever the back
1605 end does any drawing it informs the front end of which parts of the
1606 window it has accessed, and hence which parts need repainting. This
1607 is done by calling \cw{draw_update()} (\k{drawing-draw-update}).
1609 In the following sections I first discuss the drawing API as seen by
1610 the back end, and then the \e{almost} identical function-pointer
1611 form seen by the front end.
1613 \H{drawing-backend} Drawing API as seen by the back end
1615 This section documents the back-end drawing API, in the form of
1616 functions which take a \c{drawing} object as an argument.
1618 \S{drawing-draw-rect} \cw{draw_rect()}
1620 \c void draw_rect(drawing *dr, int x, int y, int w, int h,
1623 Draws a filled rectangle in the puzzle window.
1625 \c{x} and \c{y} give the coordinates of the top left pixel of the
1626 rectangle. \c{w} and \c{h} give its width and height. Thus, the
1627 horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1628 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1631 \c{colour} is an integer index into the colours array returned by
1632 the back end function \cw{colours()} (\k{backend-colours}).
1634 There is no separate pixel-plotting function. If you want to plot a
1635 single pixel, the approved method is to use \cw{draw_rect()} with
1636 width and height set to 1.
1638 Unlike many of the other drawing functions, this function is
1639 guaranteed to be pixel-perfect: the rectangle will be sharply
1640 defined and not anti-aliased or anything like that.
1642 This function may be used for both drawing and printing.
1644 \S{drawing-draw-rect-outline} \cw{draw_rect_outline()}
1646 \c void draw_rect_outline(drawing *dr, int x, int y, int w, int h,
1649 Draws an outline rectangle in the puzzle window.
1651 \c{x} and \c{y} give the coordinates of the top left pixel of the
1652 rectangle. \c{w} and \c{h} give its width and height. Thus, the
1653 horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1654 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1657 \c{colour} is an integer index into the colours array returned by
1658 the back end function \cw{colours()} (\k{backend-colours}).
1660 From a back end perspective, this function may be considered to be
1661 part of the drawing API. However, front ends are not required to
1662 implement it, since it is actually implemented centrally (in
1663 \cw{misc.c}) as a wrapper on \cw{draw_polygon()}.
1665 This function may be used for both drawing and printing.
1667 \S{drawing-draw-line} \cw{draw_line()}
1669 \c void draw_line(drawing *dr, int x1, int y1, int x2, int y2,
1672 Draws a straight line in the puzzle window.
1674 \c{x1} and \c{y1} give the coordinates of one end of the line.
1675 \c{x2} and \c{y2} give the coordinates of the other end. The line
1676 drawn includes both those points.
1678 \c{colour} is an integer index into the colours array returned by
1679 the back end function \cw{colours()} (\k{backend-colours}).
1681 Some platforms may perform anti-aliasing on this function.
1682 Therefore, do not assume that you can erase a line by drawing the
1683 same line over it in the background colour; anti-aliasing might
1684 lead to perceptible ghost artefacts around the vanished line.
1686 This function may be used for both drawing and printing.
1688 \S{drawing-draw-polygon} \cw{draw_polygon()}
1690 \c void draw_polygon(drawing *dr, int *coords, int npoints,
1691 \c int fillcolour, int outlinecolour);
1693 Draws an outlined or filled polygon in the puzzle window.
1695 \c{coords} is an array of \cw{(2*npoints)} integers, containing the
1696 \c{x} and \c{y} coordinates of \c{npoints} vertices.
1698 \c{fillcolour} and \c{outlinecolour} are integer indices into the
1699 colours array returned by the back end function \cw{colours()}
1700 (\k{backend-colours}). \c{fillcolour} may also be \cw{-1} to
1701 indicate that the polygon should be outlined only.
1703 The polygon defined by the specified list of vertices is first
1704 filled in \c{fillcolour}, if specified, and then outlined in
1707 \c{outlinecolour} may \e{not} be \cw{-1}; it must be a valid colour
1708 (and front ends are permitted to enforce this by assertion). This is
1709 because different platforms disagree on whether a filled polygon
1710 should include its boundary line or not, so drawing \e{only} a
1711 filled polygon would have non-portable effects. If you want your
1712 filled polygon not to have a visible outline, you must set
1713 \c{outlinecolour} to the same as \c{fillcolour}.
1715 Some platforms may perform anti-aliasing on this function.
1716 Therefore, do not assume that you can erase a polygon by drawing the
1717 same polygon over it in the background colour. Also, be prepared for
1718 the polygon to extend a pixel beyond its obvious bounding box as a
1719 result of this; if you really need it not to do this to avoid
1720 interfering with other delicate graphics, you should probably use
1721 \cw{clip()} (\k{drawing-clip}).
1723 This function may be used for both drawing and printing.
1725 \S{drawing-draw-circle} \cw{draw_circle()}
1727 \c void draw_circle(drawing *dr, int cx, int cy, int radius,
1728 \c int fillcolour, int outlinecolour);
1730 Draws an outlined or filled circle in the puzzle window.
1732 \c{cx} and \c{cy} give the coordinates of the centre of the circle.
1733 \c{radius} gives its radius. The total horizontal pixel extent of
1734 the circle is from \c{cx-radius+1} to \c{cx+radius-1} inclusive, and
1735 the vertical extent similarly around \c{cy}.
1737 \c{fillcolour} and \c{outlinecolour} are integer indices into the
1738 colours array returned by the back end function \cw{colours()}
1739 (\k{backend-colours}). \c{fillcolour} may also be \cw{-1} to
1740 indicate that the circle should be outlined only.
1742 The circle is first filled in \c{fillcolour}, if specified, and then
1743 outlined in \c{outlinecolour}.
1745 \c{outlinecolour} may \e{not} be \cw{-1}; it must be a valid colour
1746 (and front ends are permitted to enforce this by assertion). This is
1747 because different platforms disagree on whether a filled circle
1748 should include its boundary line or not, so drawing \e{only} a
1749 filled circle would have non-portable effects. If you want your
1750 filled circle not to have a visible outline, you must set
1751 \c{outlinecolour} to the same as \c{fillcolour}.
1753 Some platforms may perform anti-aliasing on this function.
1754 Therefore, do not assume that you can erase a circle by drawing the
1755 same circle over it in the background colour. Also, be prepared for
1756 the circle to extend a pixel beyond its obvious bounding box as a
1757 result of this; if you really need it not to do this to avoid
1758 interfering with other delicate graphics, you should probably use
1759 \cw{clip()} (\k{drawing-clip}).
1761 This function may be used for both drawing and printing.
1763 \S{drawing-draw-text} \cw{draw_text()}
1765 \c void draw_text(drawing *dr, int x, int y, int fonttype,
1766 \c int fontsize, int align, int colour, char *text);
1768 Draws text in the puzzle window.
1770 \c{x} and \c{y} give the coordinates of a point. The relation of
1771 this point to the location of the text is specified by \c{align},
1772 which is a bitwise OR of horizontal and vertical alignment flags:
1774 \dt \cw{ALIGN_VNORMAL}
1776 \dd Indicates that \c{y} is aligned with the baseline of the text.
1778 \dt \cw{ALIGN_VCENTRE}
1780 \dd Indicates that \c{y} is aligned with the vertical centre of the
1781 text. (In fact, it's aligned with the vertical centre of normal
1782 \e{capitalised} text: displaying two pieces of text with
1783 \cw{ALIGN_VCENTRE} at the same \cw{y}-coordinate will cause their
1784 baselines to be aligned with one another, even if one is an ascender
1785 and the other a descender.)
1787 \dt \cw{ALIGN_HLEFT}
1789 \dd Indicates that \c{x} is aligned with the left-hand end of the
1792 \dt \cw{ALIGN_HCENTRE}
1794 \dd Indicates that \c{x} is aligned with the horizontal centre of
1797 \dt \cw{ALIGN_HRIGHT}
1799 \dd Indicates that \c{x} is aligned with the right-hand end of the
1802 \c{fonttype} is either \cw{FONT_FIXED} or \cw{FONT_VARIABLE}, for a
1803 monospaced or proportional font respectively. (No more detail than
1804 that may be specified; it would only lead to portability issues
1805 between different platforms.)
1807 \c{fontsize} is the desired size, in pixels, of the text. This size
1808 corresponds to the overall point size of the text, not to any
1809 internal dimension such as the cap-height.
1811 \c{colour} is an integer index into the colours array returned by
1812 the back end function \cw{colours()} (\k{backend-colours}).
1814 This function may be used for both drawing and printing.
1816 \S{drawing-clip} \cw{clip()}
1818 \c void clip(drawing *dr, int x, int y, int w, int h);
1820 Establishes a clipping rectangle in the puzzle window.
1822 \c{x} and \c{y} give the coordinates of the top left pixel of the
1823 clipping rectangle. \c{w} and \c{h} give its width and height. Thus,
1824 the horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1825 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1826 inclusive. (These are exactly the same semantics as
1829 After this call, no drawing operation will affect anything outside
1830 the specified rectangle. The effect can be reversed by calling
1831 \cw{unclip()} (\k{drawing-unclip}).
1833 Back ends should not assume that a clipping rectangle will be
1834 automatically cleared up by the front end if it's left lying around;
1835 that might work on current front ends, but shouldn't be relied upon.
1836 Always explicitly call \cw{unclip()}.
1838 This function may be used for both drawing and printing.
1840 \S{drawing-unclip} \cw{unclip()}
1842 \c void unclip(drawing *dr);
1844 Reverts the effect of a previous call to \cw{clip()}. After this
1845 call, all drawing operations will be able to affect the entire
1846 puzzle window again.
1848 This function may be used for both drawing and printing.
1850 \S{drawing-draw-update} \cw{draw_update()}
1852 \c void draw_update(drawing *dr, int x, int y, int w, int h);
1854 Informs the front end that a rectangular portion of the puzzle
1855 window has been drawn on and needs to be updated.
1857 \c{x} and \c{y} give the coordinates of the top left pixel of the
1858 update rectangle. \c{w} and \c{h} give its width and height. Thus,
1859 the horizontal extent of the rectangle runs from \c{x} to \c{x+w-1}
1860 inclusive, and the vertical extent from \c{y} to \c{y+h-1}
1861 inclusive. (These are exactly the same semantics as
1864 The back end redraw function \e{must} call this function to report
1865 any changes it has made to the window. Otherwise, those changes may
1866 not become immediately visible, and may then appear at an
1867 unpredictable subsequent time such as the next time the window is
1868 covered and re-exposed.
1870 This function is only important when drawing. It may be called when
1871 printing as well, but doing so is not compulsory, and has no effect.
1872 (So if you have a shared piece of code between the drawing and
1873 printing routines, that code may safely call \cw{draw_update()}.)
1875 \S{drawing-status-bar} \cw{status_bar()}
1877 \c void status_bar(drawing *dr, char *text);
1879 Sets the text in the game's status bar to \c{text}. The text is copied
1880 from the supplied buffer, so the caller is free to deallocate or
1881 modify the buffer after use.
1883 (This function is not exactly a \e{drawing} function, but it shares
1884 with the drawing API the property that it may only be called from
1885 within the back end redraw function, so this is as good a place as
1886 any to document it.)
1888 The supplied text is filtered through the mid-end for optional
1889 rewriting before being passed on to the front end; the mid-end will
1890 prepend the current game time if the game is timed (and may in
1891 future perform other rewriting if it seems like a good idea).
1893 This function is for drawing only; it must never be called during
1896 \S{drawing-blitter} Blitter functions
1898 This section describes a group of related functions which save and
1899 restore a section of the puzzle window. This is most commonly used
1900 to implement user interfaces involving dragging a puzzle element
1901 around the window: at the end of each call to \cw{redraw()}, if an
1902 object is currently being dragged, the back end saves the window
1903 contents under that location and then draws the dragged object, and
1904 at the start of the next \cw{redraw()} the first thing it does is to
1905 restore the background.
1907 The front end defines an opaque type called a \c{blitter}, which is
1908 capable of storing a rectangular area of a specified size.
1910 Blitter functions are for drawing only; they must never be called
1913 \S2{drawing-blitter-new} \cw{blitter_new()}
1915 \c blitter *blitter_new(drawing *dr, int w, int h);
1917 Creates a new blitter object which stores a rectangle of size \c{w}
1918 by \c{h} pixels. Returns a pointer to the blitter object.
1920 Blitter objects are best stored in the \c{game_drawstate}. A good
1921 time to create them is in the \cw{set_size()} function
1922 (\k{backend-set-size}), since it is at this point that you first
1923 know how big a rectangle they will need to save.
1925 \S2{drawing-blitter-free} \cw{blitter_free()}
1927 \c void blitter_free(drawing *dr, blitter *bl);
1929 Disposes of a blitter object. Best called in \cw{free_drawstate()}.
1930 (However, check that the blitter object is not \cw{NULL} before
1931 attempting to free it; it is possible that a draw state might be
1932 created and freed without ever having \cw{set_size()} called on it
1935 \S2{drawing-blitter-save} \cw{blitter_save()}
1937 \c void blitter_save(drawing *dr, blitter *bl, int x, int y);
1939 This is a true drawing API function, in that it may only be called
1940 from within the game redraw routine. It saves a rectangular portion
1941 of the puzzle window into the specified blitter object.
1943 \c{x} and \c{y} give the coordinates of the top left corner of the
1944 saved rectangle. The rectangle's width and height are the ones
1945 specified when the blitter object was created.
1947 This function is required to cope and do the right thing if \c{x}
1948 and \c{y} are out of range. (The right thing probably means saving
1949 whatever part of the blitter rectangle overlaps with the visible
1950 area of the puzzle window.)
1952 \S2{drawing-blitter-load} \cw{blitter_load()}
1954 \c void blitter_load(drawing *dr, blitter *bl, int x, int y);
1956 This is a true drawing API function, in that it may only be called
1957 from within the game redraw routine. It restores a rectangular
1958 portion of the puzzle window from the specified blitter object.
1960 \c{x} and \c{y} give the coordinates of the top left corner of the
1961 rectangle to be restored. The rectangle's width and height are the
1962 ones specified when the blitter object was created.
1964 Alternatively, you can specify both \c{x} and \c{y} as the special
1965 value \cw{BLITTER_FROMSAVED}, in which case the rectangle will be
1966 restored to exactly where it was saved from. (This is probably what
1967 you want to do almost all the time, if you're using blitters to
1968 implement draggable puzzle elements.)
1970 This function is required to cope and do the right thing if \c{x}
1971 and \c{y} (or the equivalent ones saved in the blitter) are out of
1972 range. (The right thing probably means restoring whatever part of
1973 the blitter rectangle overlaps with the visible area of the puzzle
1976 If this function is called on a blitter which had previously been
1977 saved from a partially out-of-range rectangle, then the parts of the
1978 saved bitmap which were not visible at save time are undefined. If
1979 the blitter is restored to a different position so as to make those
1980 parts visible, the effect on the drawing area is undefined.
1982 \S{print-mono-colour} \cw{print_mono_colour()}
1984 \c int print_mono_colour(drawing *dr, int grey);
1986 This function allocates a colour index for a simple monochrome
1987 colour during printing.
1989 \c{grey} must be 0 or 1. If \c{grey} is 0, the colour returned is
1990 black; if \c{grey} is 1, the colour is white.
1992 \S{print-grey-colour} \cw{print_grey_colour()}
1994 \c int print_grey_colour(drawing *dr, float grey);
1996 This function allocates a colour index for a grey-scale colour
1999 \c{grey} may be any number between 0 (black) and 1 (white); for
2000 example, 0.5 indicates a medium grey.
2002 The chosen colour will be rendered to the limits of the printer's
2003 halftoning capability.
2005 \S{print-hatched-colour} \cw{print_hatched_colour()}
2007 \c int print_hatched_colour(drawing *dr, int hatch);
2009 This function allocates a colour index which does not represent a
2010 literal \e{colour}. Instead, regions shaded in this colour will be
2011 hatched with parallel lines. The \c{hatch} parameter defines what
2012 type of hatching should be used in place of this colour:
2014 \dt \cw{HATCH_SLASH}
2016 \dd This colour will be hatched by lines slanting to the right at 45
2019 \dt \cw{HATCH_BACKSLASH}
2021 \dd This colour will be hatched by lines slanting to the left at 45
2024 \dt \cw{HATCH_HORIZ}
2026 \dd This colour will be hatched by horizontal lines.
2030 \dd This colour will be hatched by vertical lines.
2034 \dd This colour will be hatched by criss-crossing horizontal and
2039 \dd This colour will be hatched by criss-crossing diagonal lines.
2041 Colours defined to use hatching may not be used for drawing lines or
2042 text; they may only be used for filling areas. That is, they may be
2043 used as the \c{fillcolour} parameter to \cw{draw_circle()} and
2044 \cw{draw_polygon()}, and as the colour parameter to
2045 \cw{draw_rect()}, but may not be used as the \c{outlinecolour}
2046 parameter to \cw{draw_circle()} or \cw{draw_polygon()}, or with
2047 \cw{draw_line()} or \cw{draw_text()}.
2049 \S{print-rgb-mono-colour} \cw{print_rgb_mono_colour()}
2051 \c int print_rgb_mono_colour(drawing *dr, float r, float g,
2052 \c float b, float grey);
2054 This function allocates a colour index for a fully specified RGB
2055 colour during printing.
2057 \c{r}, \c{g} and \c{b} may each be anywhere in the range from 0 to 1.
2059 If printing in black and white only, these values will be ignored,
2060 and either pure black or pure white will be used instead, according
2061 to the \q{grey} parameter. (The fallback colour is the same as the
2062 one which would be allocated by \cw{print_mono_colour(grey)}.)
2064 \S{print-rgb-grey-colour} \cw{print_rgb_grey_colour()}
2066 \c int print_rgb_grey_colour(drawing *dr, float r, float g,
2067 \c float b, float grey);
2069 This function allocates a colour index for a fully specified RGB
2070 colour during printing.
2072 \c{r}, \c{g} and \c{b} may each be anywhere in the range from 0 to 1.
2074 If printing in black and white only, these values will be ignored,
2075 and a shade of grey given by the \c{grey} parameter will be used
2076 instead. (The fallback colour is the same as the one which would be
2077 allocated by \cw{print_grey_colour(grey)}.)
2079 \S{print-rgb-hatched-colour} \cw{print_rgb_hatched_colour()}
2081 \c int print_rgb_hatched_colour(drawing *dr, float r, float g,
2082 \c float b, float hatched);
2084 This function allocates a colour index for a fully specified RGB
2085 colour during printing.
2087 \c{r}, \c{g} and \c{b} may each be anywhere in the range from 0 to 1.
2089 If printing in black and white only, these values will be ignored,
2090 and a form of cross-hatching given by the \c{hatch} parameter will
2091 be used instead; see \k{print-hatched-colour} for the possible
2092 values of this parameter. (The fallback colour is the same as the
2093 one which would be allocated by \cw{print_hatched_colour(hatch)}.)
2095 \S{print-line-width} \cw{print_line_width()}
2097 \c void print_line_width(drawing *dr, int width);
2099 This function is called to set the thickness of lines drawn during
2100 printing. It is meaningless in drawing: all lines drawn by
2101 \cw{draw_line()}, \cw{draw_circle} and \cw{draw_polygon()} are one
2102 pixel in thickness. However, in printing there is no clear
2103 definition of a pixel and so line widths must be explicitly
2106 The line width is specified in the usual coordinate system. Note,
2107 however, that it is a hint only: the central printing system may
2108 choose to vary line thicknesses at user request or due to printer
2111 \H{drawing-frontend} The drawing API as implemented by the front end
2113 This section describes the drawing API in the function-pointer form
2114 in which it is implemented by a front end.
2116 (It isn't only platform-specific front ends which implement this
2117 API; the platform-independent module \c{ps.c} also provides an
2118 implementation of it which outputs PostScript. Thus, any platform
2119 which wants to do PS printing can do so with minimum fuss.)
2121 The following entries all describe function pointer fields in a
2122 structure called \c{drawing_api}. Each of the functions takes a
2123 \cq{void *} context pointer, which it should internally cast back to
2124 a more useful type. Thus, a drawing \e{object} (\c{drawing *)}
2125 suitable for passing to the back end redraw or printing functions
2126 is constructed by passing a \c{drawing_api} and a \cq{void *} to the
2127 function \cw{drawing_new()} (see \k{drawing-new}).
2129 \S{drawingapi-draw-text} \cw{draw_text()}
2131 \c void (*draw_text)(void *handle, int x, int y, int fonttype,
2132 \c int fontsize, int align, int colour, char *text);
2134 This function behaves exactly like the back end \cw{draw_text()}
2135 function; see \k{drawing-draw-text}.
2137 \S{drawingapi-draw-rect} \cw{draw_rect()}
2139 \c void (*draw_rect)(void *handle, int x, int y, int w, int h,
2142 This function behaves exactly like the back end \cw{draw_rect()}
2143 function; see \k{drawing-draw-rect}.
2145 \S{drawingapi-draw-line} \cw{draw_line()}
2147 \c void (*draw_line)(void *handle, int x1, int y1, int x2, int y2,
2150 This function behaves exactly like the back end \cw{draw_line()}
2151 function; see \k{drawing-draw-line}.
2153 \S{drawingapi-draw-polygon} \cw{draw_polygon()}
2155 \c void (*draw_polygon)(void *handle, int *coords, int npoints,
2156 \c int fillcolour, int outlinecolour);
2158 This function behaves exactly like the back end \cw{draw_polygon()}
2159 function; see \k{drawing-draw-polygon}.
2161 \S{drawingapi-draw-circle} \cw{draw_circle()}
2163 \c void (*draw_circle)(void *handle, int cx, int cy, int radius,
2164 \c int fillcolour, int outlinecolour);
2166 This function behaves exactly like the back end \cw{draw_circle()}
2167 function; see \k{drawing-draw-circle}.
2169 \S{drawingapi-draw-update} \cw{draw_update()}
2171 \c void (*draw_update)(void *handle, int x, int y, int w, int h);
2173 This function behaves exactly like the back end \cw{draw_update()}
2174 function; see \k{drawing-draw-update}.
2176 An implementation of this API which only supports printing is
2177 permitted to define this function pointer to be \cw{NULL} rather
2178 than bothering to define an empty function. The middleware in
2179 \cw{drawing.c} will notice and avoid calling it.
2181 \S{drawingapi-clip} \cw{clip()}
2183 \c void (*clip)(void *handle, int x, int y, int w, int h);
2185 This function behaves exactly like the back end \cw{clip()}
2186 function; see \k{drawing-clip}.
2188 \S{drawingapi-unclip} \cw{unclip()}
2190 \c void (*unclip)(void *handle);
2192 This function behaves exactly like the back end \cw{unclip()}
2193 function; see \k{drawing-unclip}.
2195 \S{drawingapi-start-draw} \cw{start_draw()}
2197 \c void (*start_draw)(void *handle);
2199 This function is called at the start of drawing. It allows the front
2200 end to initialise any temporary data required to draw with, such as
2203 Implementations of this API which do not provide drawing services
2204 may define this function pointer to be \cw{NULL}; it will never be
2205 called unless drawing is attempted.
2207 \S{drawingapi-end-draw} \cw{end_draw()}
2209 \c void (*end_draw)(void *handle);
2211 This function is called at the end of drawing. It allows the front
2212 end to do cleanup tasks such as deallocating device contexts and
2213 scheduling appropriate GUI redraw events.
2215 Implementations of this API which do not provide drawing services
2216 may define this function pointer to be \cw{NULL}; it will never be
2217 called unless drawing is attempted.
2219 \S{drawingapi-status-bar} \cw{status_bar()}
2221 \c void (*status_bar)(void *handle, char *text);
2223 This function behaves exactly like the back end \cw{status_bar()}
2224 function; see \k{drawing-status-bar}.
2226 Front ends implementing this function need not worry about it being
2227 called repeatedly with the same text; the middleware code in
2228 \cw{status_bar()} will take care of this.
2230 Implementations of this API which do not provide drawing services
2231 may define this function pointer to be \cw{NULL}; it will never be
2232 called unless drawing is attempted.
2234 \S{drawingapi-blitter-new} \cw{blitter_new()}
2236 \c blitter *(*blitter_new)(void *handle, int w, int h);
2238 This function behaves exactly like the back end \cw{blitter_new()}
2239 function; see \k{drawing-blitter-new}.
2241 Implementations of this API which do not provide drawing services
2242 may define this function pointer to be \cw{NULL}; it will never be
2243 called unless drawing is attempted.
2245 \S{drawingapi-blitter-free} \cw{blitter_free()}
2247 \c void (*blitter_free)(void *handle, blitter *bl);
2249 This function behaves exactly like the back end \cw{blitter_free()}
2250 function; see \k{drawing-blitter-free}.
2252 Implementations of this API which do not provide drawing services
2253 may define this function pointer to be \cw{NULL}; it will never be
2254 called unless drawing is attempted.
2256 \S{drawingapi-blitter-save} \cw{blitter_save()}
2258 \c void (*blitter_save)(void *handle, blitter *bl, int x, int y);
2260 This function behaves exactly like the back end \cw{blitter_save()}
2261 function; see \k{drawing-blitter-save}.
2263 Implementations of this API which do not provide drawing services
2264 may define this function pointer to be \cw{NULL}; it will never be
2265 called unless drawing is attempted.
2267 \S{drawingapi-blitter-load} \cw{blitter_load()}
2269 \c void (*blitter_load)(void *handle, blitter *bl, int x, int y);
2271 This function behaves exactly like the back end \cw{blitter_load()}
2272 function; see \k{drawing-blitter-load}.
2274 Implementations of this API which do not provide drawing services
2275 may define this function pointer to be \cw{NULL}; it will never be
2276 called unless drawing is attempted.
2278 \S{drawingapi-begin-doc} \cw{begin_doc()}
2280 \c void (*begin_doc)(void *handle, int pages);
2282 This function is called at the beginning of a printing run. It gives
2283 the front end an opportunity to initialise any required printing
2284 subsystem. It also provides the number of pages in advance.
2286 Implementations of this API which do not provide printing services
2287 may define this function pointer to be \cw{NULL}; it will never be
2288 called unless printing is attempted.
2290 \S{drawingapi-begin-page} \cw{begin_page()}
2292 \c void (*begin_page)(void *handle, int number);
2294 This function is called during printing, at the beginning of each
2295 page. It gives the page number (numbered from 1 rather than 0, so
2296 suitable for use in user-visible contexts).
2298 Implementations of this API which do not provide printing services
2299 may define this function pointer to be \cw{NULL}; it will never be
2300 called unless printing is attempted.
2302 \S{drawingapi-begin-puzzle} \cw{begin_puzzle()}
2304 \c void (*begin_puzzle)(void *handle, float xm, float xc,
2305 \c float ym, float yc, int pw, int ph, float wmm);
2307 This function is called during printing, just before printing a
2308 single puzzle on a page. It specifies the size and location of the
2311 \c{xm} and \c{xc} specify the horizontal position of the puzzle on
2312 the page, as a linear function of the page width. The front end is
2313 expected to multiply the page width by \c{xm}, add \c{xc} (measured
2314 in millimetres), and use the resulting x-coordinate as the left edge
2317 Similarly, \c{ym} and \c{yc} specify the vertical position of the
2318 puzzle as a function of the page height: the page height times
2319 \c{ym}, plus \c{yc} millimetres, equals the desired distance from
2320 the top of the page to the top of the puzzle.
2322 (This unwieldy mechanism is required because not all printing
2323 systems can communicate the page size back to the software. The
2324 PostScript back end, for example, writes out PS which determines the
2325 page size at print time by means of calling \cq{clippath}, and
2326 centres the puzzles within that. Thus, exactly the same PS file
2327 works on A4 or on US Letter paper without needing local
2328 configuration, which simplifies matters.)
2330 \cw{pw} and \cw{ph} give the size of the puzzle in drawing API
2331 coordinates. The printing system will subsequently call the puzzle's
2332 own print function, which will in turn call drawing API functions in
2333 the expectation that an area \cw{pw} by \cw{ph} units is available
2334 to draw the puzzle on.
2336 Finally, \cw{wmm} gives the desired width of the puzzle in
2337 millimetres. (The aspect ratio is expected to be preserved, so if
2338 the desired puzzle height is also needed then it can be computed as
2341 Implementations of this API which do not provide printing services
2342 may define this function pointer to be \cw{NULL}; it will never be
2343 called unless printing is attempted.
2345 \S{drawingapi-end-puzzle} \cw{end_puzzle()}
2347 \c void (*end_puzzle)(void *handle);
2349 This function is called after the printing of a specific puzzle is
2352 Implementations of this API which do not provide printing services
2353 may define this function pointer to be \cw{NULL}; it will never be
2354 called unless printing is attempted.
2356 \S{drawingapi-end-page} \cw{end_page()}
2358 \c void (*end_page)(void *handle, int number);
2360 This function is called after the printing of a page is finished.
2362 Implementations of this API which do not provide printing services
2363 may define this function pointer to be \cw{NULL}; it will never be
2364 called unless printing is attempted.
2366 \S{drawingapi-end-doc} \cw{end_doc()}
2368 \c void (*end_doc)(void *handle);
2370 This function is called after the printing of the entire document is
2371 finished. This is the moment to close files, send things to the
2372 print spooler, or whatever the local convention is.
2374 Implementations of this API which do not provide printing services
2375 may define this function pointer to be \cw{NULL}; it will never be
2376 called unless printing is attempted.
2378 \S{drawingapi-line-width} \cw{line_width()}
2380 \c void (*line_width)(void *handle, float width);
2382 This function is called to set the line thickness, during printing
2383 only. Note that the width is a \cw{float} here, where it was an
2384 \cw{int} as seen by the back end. This is because \cw{drawing.c} may
2385 have scaled it on the way past.
2387 However, the width is still specified in the same coordinate system
2388 as the rest of the drawing.
2390 Implementations of this API which do not provide printing services
2391 may define this function pointer to be \cw{NULL}; it will never be
2392 called unless printing is attempted.
2394 \H{drawingapi-frontend} The drawing API as called by the front end
2396 There are a small number of functions provided in \cw{drawing.c}
2397 which the front end needs to \e{call}, rather than helping to
2398 implement. They are described in this section.
2400 \S{drawing-new} \cw{drawing_new()}
2402 \c drawing *drawing_new(const drawing_api *api, midend *me,
2405 This function creates a drawing object. It is passed a
2406 \c{drawing_api}, which is a structure containing nothing but
2407 function pointers; and also a \cq{void *} handle. The handle is
2408 passed back to each function pointer when it is called.
2410 The \c{midend} parameter is used for rewriting the status bar
2411 contents: \cw{status_bar()} (see \k{drawing-status-bar}) has to call
2412 a function in the mid-end which might rewrite the status bar text.
2413 If the drawing object is to be used only for printing, or if the
2414 game is known not to call \cw{status_bar()}, this parameter may be
2417 \S{drawing-free} \cw{drawing_free()}
2419 \c void drawing_free(drawing *dr);
2421 This function frees a drawing object. Note that the \cq{void *}
2422 handle is not freed; if that needs cleaning up it must be done by
2425 \S{drawing-print-get-colour} \cw{print_get_colour()}
2427 \c void print_get_colour(drawing *dr, int colour, int printincolour,
2428 \c int *hatch, float *r, float *g, float *b)
2430 This function is called by the implementations of the drawing API
2431 functions when they are called in a printing context. It takes a
2432 colour index as input, and returns the description of the colour as
2433 requested by the back end.
2435 \c{printincolour} is \cw{TRUE} iff the implementation is printing in
2436 colour. This will alter the results returned if the colour in
2437 question was specified with a black-and-white fallback value.
2439 If the colour should be rendered by hatching, \c{*hatch} is filled
2440 with the type of hatching desired. See \k{print-grey-colour} for
2441 details of the values this integer can take.
2443 If the colour should be rendered as solid colour, \c{*hatch} is
2444 given a negative value, and \c{*r}, \c{*g} and \c{*b} are filled
2445 with the RGB values of the desired colour (if printing in colour),
2446 or all filled with the grey-scale value (if printing in black and
2449 \C{midend} The API provided by the mid-end
2451 This chapter documents the API provided by the mid-end to be called
2452 by the front end. You probably only need to read this if you are a
2453 front end implementor, i.e. you are porting Puzzles to a new
2454 platform. If you're only interested in writing new puzzles, you can
2455 safely skip this chapter.
2457 All the persistent state in the mid-end is encapsulated within a
2458 \c{midend} structure, to facilitate having multiple mid-ends in any
2459 port which supports multiple puzzle windows open simultaneously.
2460 Each \c{midend} is intended to handle the contents of a single
2463 \H{midend-new} \cw{midend_new()}
2465 \c midend *midend_new(frontend *fe, const game *ourgame,
2466 \c const drawing_api *drapi, void *drhandle)
2468 Allocates and returns a new mid-end structure.
2470 The \c{fe} argument is stored in the mid-end. It will be used when
2471 calling back to functions such as \cw{activate_timer()}
2472 (\k{frontend-activate-timer}), and will be passed on to the back end
2473 function \cw{colours()} (\k{backend-colours}).
2475 The parameters \c{drapi} and \c{drhandle} are passed to
2476 \cw{drawing_new()} (\k{drawing-new}) to construct a drawing object
2477 which will be passed to the back end function \cw{redraw()}
2478 (\k{backend-redraw}). Hence, all drawing-related function pointers
2479 defined in \c{drapi} can expect to be called with \c{drhandle} as
2480 their first argument.
2482 The \c{ourgame} argument points to a container structure describing
2483 a game back end. The mid-end thus created will only be capable of
2484 handling that one game. (So even in a monolithic front end
2485 containing all the games, this imposes the constraint that any
2486 individual puzzle window is tied to a single game. Unless, of
2487 course, you feel brave enough to change the mid-end for the window
2488 without closing the window...)
2490 \H{midend-free} \cw{midend_free()}
2492 \c void midend_free(midend *me);
2494 Frees a mid-end structure and all its associated data.
2496 \H{midend-set-params} \cw{midend_set_params()}
2498 \c void midend_set_params(midend *me, game_params *params);
2500 Sets the current game parameters for a mid-end. Subsequent games
2501 generated by \cw{midend_new_game()} (\k{midend-new-game}) will use
2502 these parameters until further notice.
2504 The usual way in which the front end will have an actual
2505 \c{game_params} structure to pass to this function is if it had
2506 previously got it from \cw{midend_fetch_preset()}
2507 (\k{midend-fetch-preset}). Thus, this function is usually called in
2508 response to the user making a selection from the presets menu.
2510 \H{midend-get-params} \cw{midend_get_params()}
2512 \c game_params *midend_get_params(midend *me);
2514 Returns the current game parameters stored in this mid-end.
2516 The returned value is dynamically allocated, and should be freed
2517 when finished with by passing it to the game's own
2518 \cw{free_params()} function (see \k{backend-free-params}).
2520 \H{midend-size} \cw{midend_size()}
2522 \c void midend_size(midend *me, int *x, int *y, int user_size);
2524 Tells the mid-end to figure out its window size.
2526 On input, \c{*x} and \c{*y} should contain the maximum or requested
2527 size for the window. (Typically this will be the size of the screen
2528 that the window has to fit on, or similar.) The mid-end will
2529 repeatedly call the back end function \cw{compute_size()}
2530 (\k{backend-compute-size}), searching for a tile size that best
2531 satisfies the requirements. On exit, \c{*x} and \c{*y} will contain
2532 the size needed for the puzzle window's drawing area. (It is of
2533 course up to the front end to adjust this for any additional window
2534 furniture such as menu bars and window borders, if necessary. The
2535 status bar is also not included in this size.)
2537 Use \c{user_size} to indicate whether \c{*x} and \c{*y} are a
2538 requested size, or just a maximum size.
2540 If \c{user_size} is set to \cw{TRUE}, the mid-end will treat the
2541 input size as a request, and will pick a tile size which
2542 approximates it \e{as closely as possible}, going over the game's
2543 preferred tile size if necessary to achieve this. The mid-end will
2544 also use the resulting tile size as its preferred one until further
2545 notice, on the assumption that this size was explicitly requested
2546 by the user. Use this option if you want your front end to support
2547 dynamic resizing of the puzzle window with automatic scaling of the
2550 If \c{user_size} is set to \cw{FALSE}, then the game's tile size
2551 will never go over its preferred one, although it may go under in
2552 order to fit within the maximum bounds specified by \c{*x} and
2553 \c{*y}. This is the recommended approach when opening a new window
2554 at default size: the game will use its preferred size unless it has
2555 to use a smaller one to fit on the screen. If the tile size is
2556 shrunk for this reason, the change will not persist; if a smaller
2557 grid is subsequently chosen, the tile size will recover.
2559 The mid-end will try as hard as it can to return a size which is
2560 less than or equal to the input size, in both dimensions. In extreme
2561 circumstances it may fail (if even the lowest possible tile size
2562 gives window dimensions greater than the input), in which case it
2563 will return a size greater than the input size. Front ends should be
2564 prepared for this to happen (i.e. don't crash or fail an assertion),
2565 but may handle it in any way they see fit: by rejecting the game
2566 parameters which caused the problem, by opening a window larger than
2567 the screen regardless of inconvenience, by introducing scroll bars
2568 on the window, by drawing on a large bitmap and scaling it into a
2569 smaller window, or by any other means you can think of. It is likely
2570 that when the tile size is that small the game will be unplayable
2571 anyway, so don't put \e{too} much effort into handling it
2574 If your platform has no limit on window size (or if you're planning
2575 to use scroll bars for large puzzles), you can pass dimensions of
2576 \cw{INT_MAX} as input to this function. You should probably not do
2577 that \e{and} set the \c{user_size} flag, though!
2579 \H{midend-new-game} \cw{midend_new_game()}
2581 \c void midend_new_game(midend *me);
2583 Causes the mid-end to begin a new game. Normally the game will be a
2584 new randomly generated puzzle. However, if you have previously
2585 called \cw{midend_game_id()} or \cw{midend_set_config()}, the game
2586 generated might be dictated by the results of those functions. (In
2587 particular, you \e{must} call \cw{midend_new_game()} after calling
2588 either of those functions, or else no immediate effect will be
2591 You will probably need to call \cw{midend_size()} after calling this
2592 function, because if the game parameters have been changed since the
2593 last new game then the window size might need to change. (If you
2594 know the parameters \e{haven't} changed, you don't need to do this.)
2596 This function will create a new \c{game_drawstate}, but does not
2597 actually perform a redraw (since you often need to call
2598 \cw{midend_size()} before the redraw can be done). So after calling
2599 this function and after calling \cw{midend_size()}, you should then
2600 call \cw{midend_redraw()}. (It is not necessary to call
2601 \cw{midend_force_redraw()}; that will discard the draw state and
2602 create a fresh one, which is unnecessary in this case since there's
2603 a fresh one already. It would work, but it's usually excessive.)
2605 \H{midend-restart-game} \cw{midend_restart_game()}
2607 \c void midend_restart_game(midend *me);
2609 This function causes the current game to be restarted. This is done
2610 by placing a new copy of the original game state on the end of the
2611 undo list (so that an accidental restart can be undone).
2613 This function automatically causes a redraw, i.e. the front end can
2614 expect its drawing API to be called from \e{within} a call to this
2617 \H{midend-force-redraw} \cw{midend_force_redraw()}
2619 \c void midend_force_redraw(midend *me);
2621 Forces a complete redraw of the puzzle window, by means of
2622 discarding the current \c{game_drawstate} and creating a new one
2623 from scratch before calling the game's \cw{redraw()} function.
2625 The front end can expect its drawing API to be called from within a
2626 call to this function.
2628 \H{midend-redraw} \cw{midend_redraw()}
2630 \c void midend_redraw(midend *me);
2632 Causes a partial redraw of the puzzle window, by means of simply
2633 calling the game's \cw{redraw()} function. (That is, the only things
2634 redrawn will be things that have changed since the last redraw.)
2636 The front end can expect its drawing API to be called from within a
2637 call to this function.
2639 \H{midend-process-key} \cw{midend_process_key()}
2641 \c int midend_process_key(midend *me, int x, int y, int button);
2643 The front end calls this function to report a mouse or keyboard
2644 event. The parameters \c{x}, \c{y} and \c{button} are almost
2645 identical to the ones passed to the back end function
2646 \cw{interpret_move()} (\k{backend-interpret-move}), except that the
2647 front end is \e{not} required to provide the guarantees about mouse
2648 event ordering. The mid-end will sort out multiple simultaneous
2649 button presses and changes of button; the front end's responsibility
2650 is simply to pass on the mouse events it receives as accurately as
2653 (Some platforms may need to emulate absent mouse buttons by means of
2654 using a modifier key such as Shift with another mouse button. This
2655 tends to mean that if Shift is pressed or released in the middle of
2656 a mouse drag, the mid-end will suddenly stop receiving, say,
2657 \cw{LEFT_DRAG} events and start receiving \cw{RIGHT_DRAG}s, with no
2658 intervening button release or press events. This too is something
2659 which the mid-end will sort out for you; the front end has no
2660 obligation to maintain sanity in this area.)
2662 The front end \e{should}, however, always eventually send some kind
2663 of button release. On some platforms this requires special effort:
2664 Windows, for example, requires a call to the system API function
2665 \cw{SetCapture()} in order to ensure that your window receives a
2666 mouse-up event even if the pointer has left the window by the time
2667 the mouse button is released. On any platform that requires this
2668 sort of thing, the front end \e{is} responsible for doing it.
2670 Calling this function is very likely to result in calls back to the
2671 front end's drawing API and/or \cw{activate_timer()}
2672 (\k{frontend-activate-timer}).
2674 The return value from \cw{midend_process_key()} is non-zero, unless
2675 the effect of the keypress was to request termination of the
2676 program. A front end should shut down the puzzle in response to a
2679 \H{midend-colours} \cw{midend_colours()}
2681 \c float *midend_colours(midend *me, int *ncolours);
2683 Returns an array of the colours required by the game, in exactly the
2684 same format as that returned by the back end function \cw{colours()}
2685 (\k{backend-colours}). Front ends should call this function rather
2686 than calling the back end's version directly, since the mid-end adds
2687 standard customisation facilities. (At the time of writing, those
2688 customisation facilities are implemented hackily by means of
2689 environment variables, but it's not impossible that they may become
2690 more full and formal in future.)
2692 \H{midend-timer} \cw{midend_timer()}
2694 \c void midend_timer(midend *me, float tplus);
2696 If the mid-end has called \cw{activate_timer()}
2697 (\k{frontend-activate-timer}) to request regular callbacks for
2698 purposes of animation or timing, this is the function the front end
2699 should call on a regular basis. The argument \c{tplus} gives the
2700 time, in seconds, since the last time either this function was
2701 called or \cw{activate_timer()} was invoked.
2703 One of the major purposes of timing in the mid-end is to perform
2704 move animation. Therefore, calling this function is very likely to
2705 result in calls back to the front end's drawing API.
2707 \H{midend-num-presets} \cw{midend_num_presets()}
2709 \c int midend_num_presets(midend *me);
2711 Returns the number of game parameter presets supplied by this game.
2712 Front ends should use this function and \cw{midend_fetch_preset()}
2713 to configure their presets menu rather than calling the back end
2714 directly, since the mid-end adds standard customisation facilities.
2715 (At the time of writing, those customisation facilities are
2716 implemented hackily by means of environment variables, but it's not
2717 impossible that they may become more full and formal in future.)
2719 \H{midend-fetch-preset} \cw{midend_fetch_preset()}
2721 \c void midend_fetch_preset(midend *me, int n,
2722 \c char **name, game_params **params);
2724 Returns one of the preset game parameter structures for the game. On
2725 input \c{n} must be a non-negative integer and less than the value
2726 returned from \cw{midend_num_presets()}. On output, \c{*name} is set
2727 to an ASCII string suitable for entering in the game's presets menu,
2728 and \c{*params} is set to the corresponding \c{game_params}
2731 Both of the two output values are dynamically allocated, but they
2732 are owned by the mid-end structure: the front end should not ever
2733 free them directly, because they will be freed automatically during
2736 \H{midend-which-preset} \cw{midend_which_preset()}
2738 \c int midend_which_preset(midend *me);
2740 Returns the numeric index of the preset game parameter structure
2741 which matches the current game parameters, or a negative number if
2742 no preset matches. Front ends could use this to maintain a tick
2743 beside one of the items in the menu (or tick the \q{Custom} option
2744 if the return value is less than zero).
2746 \H{midend-wants-statusbar} \cw{midend_wants_statusbar()}
2748 \c int midend_wants_statusbar(midend *me);
2750 This function returns \cw{TRUE} if the puzzle has a use for a
2751 textual status line (to display score, completion status, currently
2752 active tiles, time, or anything else).
2754 Front ends should call this function rather than talking directly to
2757 \H{midend-get-config} \cw{midend_get_config()}
2759 \c config_item *midend_get_config(midend *me, int which,
2760 \c char **wintitle);
2762 Returns a dialog box description for user configuration.
2764 On input, \cw{which} should be set to one of three values, which
2765 select which of the various dialog box descriptions is returned:
2767 \dt \cw{CFG_SETTINGS}
2769 \dd Requests the GUI parameter configuration box generated by the
2770 puzzle itself. This should be used when the user selects \q{Custom}
2771 from the game types menu (or equivalent). The mid-end passes this
2772 request on to the back end function \cw{configure()}
2773 (\k{backend-configure}).
2777 \dd Requests a box suitable for entering a descriptive game ID (and
2778 viewing the existing one). The mid-end generates this dialog box
2779 description itself. This should be used when the user selects
2780 \q{Specific} from the game menu (or equivalent).
2784 \dd Requests a box suitable for entering a random-seed game ID (and
2785 viewing the existing one). The mid-end generates this dialog box
2786 description itself. This should be used when the user selects
2787 \q{Random Seed} from the game menu (or equivalent).
2789 The returned value is an array of \cw{config_item}s, exactly as
2790 described in \k{backend-configure}. Another returned value is an
2791 ASCII string giving a suitable title for the configuration window,
2794 Both returned values are dynamically allocated and will need to be
2795 freed. The window title can be freed in the obvious way; the
2796 \cw{config_item} array is a slightly complex structure, so a utility
2797 function \cw{free_cfg()} is provided to free it for you. See
2800 (Of course, you will probably not want to free the \cw{config_item}
2801 array until the dialog box is dismissed, because before then you
2802 will probably need to pass it to \cw{midend_set_config}.)
2804 \H{midend-set-config} \cw{midend_set_config()}
2806 \c char *midend_set_config(midend *me, int which,
2807 \c config_item *cfg);
2809 Passes the mid-end the results of a configuration dialog box.
2810 \c{which} should have the same value which it had when
2811 \cw{midend_get_config()} was called; \c{cfg} should be the array of
2812 \c{config_item}s returned from \cw{midend_get_config()}, modified to
2813 contain the results of the user's editing operations.
2815 This function returns \cw{NULL} on success, or otherwise (if the
2816 configuration data was in some way invalid) an ASCII string
2817 containing an error message suitable for showing to the user.
2819 If the function succeeds, it is likely that the game parameters will
2820 have been changed and it is certain that a new game will be
2821 requested. The front end should therefore call
2822 \cw{midend_new_game()}, and probably also re-think the window size
2823 using \cw{midend_size()} and eventually perform a refresh using
2824 \cw{midend_redraw()}.
2826 \H{midend-game-id} \cw{midend_game_id()}
2828 \c char *midend_game_id(midend *me, char *id);
2830 Passes the mid-end a string game ID (of any of the valid forms
2831 \cq{params}, \cq{params:description} or \cq{params#seed}) which the
2832 mid-end will process and use for the next generated game.
2834 This function returns \cw{NULL} on success, or otherwise (if the
2835 configuration data was in some way invalid) an ASCII string
2836 containing an error message (not dynamically allocated) suitable for
2837 showing to the user. In the event of an error, the mid-end's
2838 internal state will be left exactly as it was before the call.
2840 If the function succeeds, it is likely that the game parameters will
2841 have been changed and it is certain that a new game will be
2842 requested. The front end should therefore call
2843 \cw{midend_new_game()}, and probably also re-think the window size
2844 using \cw{midend_size()} and eventually case a refresh using
2845 \cw{midend_redraw()}.
2847 \H{midend-get-game-id} \cw{midend_get_game_id()}
2849 \c char *midend_get_game_id(midend *me)
2851 Returns a descriptive game ID (i.e. one in the form
2852 \cq{params:description}) describing the game currently active in the
2853 mid-end. The returned string is dynamically allocated.
2855 \H{midend-text-format} \cw{midend_text_format()}
2857 \c char *midend_text_format(midend *me);
2859 Formats the current game's current state as ASCII text suitable for
2860 copying to the clipboard. The returned string is dynamically
2863 You should not call this function if the game's
2864 \c{can_format_as_text} flag is \cw{FALSE}.
2866 If the returned string contains multiple lines (which is likely), it
2867 will use the normal C line ending convention (\cw{\\n} only). On
2868 platforms which use a different line ending convention for data in
2869 the clipboard, it is the front end's responsibility to perform the
2872 \H{midend-solve} \cw{midend_solve()}
2874 \c char *midend_solve(midend *me);
2876 Requests the mid-end to perform a Solve operation.
2878 On success, \cw{NULL} is returned. On failure, an error message (not
2879 dynamically allocated) is returned, suitable for showing to the
2882 The front end can expect its drawing API and/or
2883 \cw{activate_timer()} to be called from within a call to this
2886 \H{midend-serialise} \cw{midend_serialise()}
2888 \c void midend_serialise(midend *me,
2889 \c void (*write)(void *ctx, void *buf, int len),
2892 Calling this function causes the mid-end to convert its entire
2893 internal state into a long ASCII text string, and to pass that
2894 string (piece by piece) to the supplied \c{write} function.
2896 Desktop implementations can use this function to save a game in any
2897 state (including half-finished) to a disk file, by supplying a
2898 \c{write} function which is a wrapper on \cw{fwrite()} (or local
2899 equivalent). Other implementations may find other uses for it, such
2900 as compressing the large and sprawling mid-end state into a
2901 manageable amount of memory when a palmtop application is suspended
2902 so that another one can run; in this case \cw{write} might want to
2903 write to a memory buffer rather than a file. There may be other uses
2906 This function will call back to the supplied \c{write} function a
2907 number of times, with the first parameter (\c{ctx}) equal to
2908 \c{wctx}, and the other two parameters pointing at a piece of the
2911 \H{midend-deserialise} \cw{midend_deserialise()}
2913 \c char *midend_deserialise(midend *me,
2914 \c int (*read)(void *ctx, void *buf, int len),
2917 This function is the counterpart to \cw{midend_serialise()}. It
2918 calls the supplied \cw{read} function repeatedly to read a quantity
2919 of data, and attempts to interpret that data as a serialised mid-end
2920 as output by \cw{midend_serialise()}.
2922 The \cw{read} function is called with the first parameter (\c{ctx})
2923 equal to \c{rctx}, and should attempt to read \c{len} bytes of data
2924 into the buffer pointed to by \c{buf}. It should return \cw{FALSE}
2925 on failure or \cw{TRUE} on success. It should not report success
2926 unless it has filled the entire buffer; on platforms which might be
2927 reading from a pipe or other blocking data source, \c{read} is
2928 responsible for looping until the whole buffer has been filled.
2930 If the de-serialisation operation is successful, the mid-end's
2931 internal data structures will be replaced by the results of the
2932 load, and \cw{NULL} will be returned. Otherwise, the mid-end's state
2933 will be completely unchanged and an error message (typically some
2934 variation on \q{save file is corrupt}) will be returned. As usual,
2935 the error message string is not dynamically allocated.
2937 If this function succeeds, it is likely that the game parameters
2938 will have been changed. The front end should therefore probably
2939 re-think the window size using \cw{midend_size()}, and probably
2940 cause a refresh using \cw{midend_redraw()}.
2942 Because each mid-end is tied to a specific game back end, this
2943 function will fail if you attempt to read in a save file generated
2944 by a different game from the one configured in this mid-end, even if
2945 your application is a monolithic one containing all the puzzles. (It
2946 would be pretty easy to write a function which would look at a save
2947 file and determine which game it was for; any front end implementor
2948 who needs such a function can probably be accommodated.)
2950 \H{frontend-backend} Direct reference to the back end structure by
2953 Although \e{most} things the front end needs done should be done by
2954 calling the mid-end, there are a few situations in which the front
2955 end needs to refer directly to the game back end structure.
2957 The most obvious of these is
2959 \b passing the game back end as a parameter to \cw{midend_new()}.
2961 There are a few other back end features which are not wrapped by the
2962 mid-end because there didn't seem much point in doing so:
2964 \b fetching the \c{name} field to use in window titles and similar
2966 \b reading the \c{can_configure}, \c{can_solve} and
2967 \c{can_format_as_text} fields to decide whether to add those items
2968 to the menu bar or equivalent
2970 \b reading the \c{winhelp_topic} field (Windows only)
2972 \b the GTK front end provides a \cq{--generate} command-line option
2973 which directly calls the back end to do most of its work. This is
2974 not really part of the main front end code, though, and I'm not sure
2977 In order to find the game back end structure, the front end does one
2980 \b If the particular front end is compiling a separate binary per
2981 game, then the back end structure is a global variable with the
2982 standard name \cq{thegame}:
2986 \c extern const game thegame;
2990 \b If the front end is compiled as a monolithic application
2991 containing all the puzzles together (in which case the preprocessor
2992 symbol \cw{COMBINED} must be defined when compiling most of the code
2993 base), then there will be two global variables defined:
2997 \c extern const game *gamelist[];
2998 \c extern const int gamecount;
3000 \c{gamelist} will be an array of \c{gamecount} game structures,
3001 declared in the automatically constructed source module \c{list.c}.
3002 The application should search that array for the game it wants,
3003 probably by reaching into each game structure and looking at its
3008 \H{frontend-api} Mid-end to front-end calls
3010 This section describes the small number of functions which a front
3011 end must provide to be called by the mid-end or other standard
3014 \H{frontend-get-random-seed} \cw{get_random_seed()}
3016 \c void get_random_seed(void **randseed, int *randseedsize);
3018 This function is called by a new mid-end, and also occasionally by
3019 game back ends. Its job is to return a piece of data suitable for
3020 using as a seed for initialisation of a new \c{random_state}.
3022 On exit, \c{*randseed} should be set to point at a newly allocated
3023 piece of memory containing some seed data, and \c{*randseedsize}
3024 should be set to the length of that data.
3026 A simple and entirely adequate implementation is to return a piece
3027 of data containing the current system time at the highest
3028 conveniently available resolution.
3030 \H{frontend-activate-timer} \cw{activate_timer()}
3032 \c void activate_timer(frontend *fe);
3034 This is called by the mid-end to request that the front end begin
3035 calling it back at regular intervals.
3037 The timeout interval is left up to the front end; the finer it is,
3038 the smoother move animations will be, but the more CPU time will be
3039 used. Current front ends use values around 20ms (i.e. 50Hz).
3041 After this function is called, the mid-end will expect to receive
3042 calls to \cw{midend_timer()} on a regular basis.
3044 \H{frontend-deactivate-timer} \cw{deactivate_timer()}
3046 \c void deactivate_timer(frontend *fe);
3048 This is called by the mid-end to request that the front end stop
3049 calling \cw{midend_timer()}.
3051 \H{frontend-fatal} \cw{fatal()}
3053 \c void fatal(char *fmt, ...);
3055 This is called by some utility functions if they encounter a
3056 genuinely fatal error such as running out of memory. It is a
3057 variadic function in the style of \cw{printf()}, and is expected to
3058 show the formatted error message to the user any way it can and then
3059 terminate the application. It must not return.
3061 \H{frontend-default-colour} \cw{frontend_default_colour()}
3063 \c void frontend_default_colour(frontend *fe, float *output);
3065 This function expects to be passed a pointer to an array of three
3066 \cw{float}s. It returns the platform's local preferred background
3067 colour in those three floats, as red, green and blue values (in that
3068 order) ranging from \cw{0.0} to \cw{1.0}.
3070 This function should only ever be called by the back end function
3071 \cw{colours()} (\k{backend-colours}). (Thus, it isn't a
3072 \e{midend}-to-frontend function as such, but there didn't seem to be
3073 anywhere else particularly good to put it. Sorry.)
3075 \C{utils} Utility APIs
3077 This chapter documents a variety of utility APIs provided for the
3078 general use of the rest of the Puzzles code.
3080 \H{utils-random} Random number generation
3082 Platforms' local random number generators vary widely in quality and
3083 seed size. Puzzles therefore supplies its own high-quality random
3084 number generator, with the additional advantage of giving the same
3085 results if fed the same seed data on different platforms. This
3086 allows game random seeds to be exchanged between different ports of
3087 Puzzles and still generate the same games.
3089 Unlike the ANSI C \cw{rand()} function, the Puzzles random number
3090 generator has an \e{explicit} state object called a
3091 \c{random_state}. One of these is managed by each mid-end, for
3092 example, and passed to the back end to generate a game with.
3094 \S{utils-random-init} \cw{random_new()}
3096 \c random_state *random_new(char *seed, int len);
3098 Allocates, initialises and returns a new \c{random_state}. The input
3099 data is used as the seed for the random number stream (i.e. using
3100 the same seed at a later time will generate the same stream).
3102 The seed data can be any data at all; there is no requirement to use
3103 printable ASCII, or NUL-terminated strings, or anything like that.
3105 \S{utils-random-copy} \cw{random_copy()}
3107 \c random_state *random_copy(random_state *tocopy);
3109 Allocates a new \c{random_state}, copies the contents of another
3110 \c{random_state} into it, and returns the new state. If exactly the
3111 same sequence of functions is subseqently called on both the copy and
3112 the original, the results will be identical. This may be useful for
3113 speculatively performing some operation using a given random state,
3114 and later replaying that operation precisely.
3116 \S{utils-random-free} \cw{random_free()}
3118 \c void random_free(random_state *state);
3120 Frees a \c{random_state}.
3122 \S{utils-random-bits} \cw{random_bits()}
3124 \c unsigned long random_bits(random_state *state, int bits);
3126 Returns a random number from 0 to \cw{2^bits-1} inclusive. \c{bits}
3127 should be between 1 and 32 inclusive.
3129 \S{utils-random-upto} \cw{random_upto()}
3131 \c unsigned long random_upto(random_state *state, unsigned long limit);
3133 Returns a random number from 0 to \cw{limit-1} inclusive.
3135 \S{utils-random-state-encode} \cw{random_state_encode()}
3137 \c char *random_state_encode(random_state *state);
3139 Encodes the entire contents of a \c{random_state} in printable
3140 ASCII. Returns a dynamically allocated string containing that
3141 encoding. This can subsequently be passed to
3142 \cw{random_state_decode()} to reconstruct the same \c{random_state}.
3144 \S{utils-random-state-decode} \cw{random_state_decode()}
3146 \c random_state *random_state_decode(char *input);
3148 Decodes a string generated by \cw{random_state_encode()} and
3149 reconstructs an equivalent \c{random_state} to the one encoded, i.e.
3150 it should produce the same stream of random numbers.
3152 This function has no error reporting; if you pass it an invalid
3153 string it will simply generate an arbitrary random state, which may
3154 turn out to be noticeably non-random.
3156 \S{utils-shuffle} \cw{shuffle()}
3158 \c void shuffle(void *array, int nelts, int eltsize, random_state *rs);
3160 Shuffles an array into a random order. The interface is much like
3161 ANSI C \cw{qsort()}, except that there's no need for a compare
3164 \c{array} is a pointer to the first element of the array. \c{nelts}
3165 is the number of elements in the array; \c{eltsize} is the size of a
3166 single element (typically measured using \c{sizeof}). \c{rs} is a
3167 \c{random_state} used to generate all the random numbers for the
3170 \H{utils-alloc} Memory allocation
3172 Puzzles has some central wrappers on the standard memory allocation
3173 functions, which provide compile-time type checking, and run-time
3174 error checking by means of quitting the application if it runs out
3175 of memory. This doesn't provide the best possible recovery from
3176 memory shortage, but on the other hand it greatly simplifies the
3177 rest of the code, because nothing else anywhere needs to worry about
3178 \cw{NULL} returns from allocation.
3180 \S{utils-snew} \cw{snew()}
3182 \c var = snew(type);
3185 This macro takes a single argument which is a \e{type name}. It
3186 allocates space for one object of that type. If allocation fails it
3187 will call \cw{fatal()} and not return; so if it does return, you can
3188 be confident that its return value is non-\cw{NULL}.
3190 The return value is cast to the specified type, so that the compiler
3191 will type-check it against the variable you assign it into. Thus,
3192 this ensures you don't accidentally allocate memory the size of the
3193 wrong type and assign it into a variable of the right one (or vice
3196 \S{utils-snewn} \cw{snewn()}
3198 \c var = snewn(n, type);
3201 This macro is the array form of \cw{snew()}. It takes two arguments;
3202 the first is a number, and the second is a type name. It allocates
3203 space for that many objects of that type, and returns a type-checked
3204 non-\cw{NULL} pointer just as \cw{snew()} does.
3206 \S{utils-sresize} \cw{sresize()}
3208 \c var = sresize(var, n, type);
3211 This macro is a type-checked form of \cw{realloc()}. It takes three
3212 arguments: an input memory block, a new size in elements, and a
3213 type. It re-sizes the input memory block to a size sufficient to
3214 contain that many elements of that type. It returns a type-checked
3215 non-\cw{NULL} pointer, like \cw{snew()} and \cw{snewn()}.
3217 The input memory block can be \cw{NULL}, in which case this function
3218 will behave exactly like \cw{snewn()}. (In principle any
3219 ANSI-compliant \cw{realloc()} implementation ought to cope with
3220 this, but I've never quite trusted it to work everywhere.)
3222 \S{utils-sfree} \cw{sfree()}
3224 \c void sfree(void *p);
3226 This function is pretty much equivalent to \cw{free()}. It is
3227 provided with a dynamically allocated block, and frees it.
3229 The input memory block can be \cw{NULL}, in which case this function
3230 will do nothing. (In principle any ANSI-compliant \cw{free()}
3231 implementation ought to cope with this, but I've never quite trusted
3232 it to work everywhere.)
3234 \S{utils-dupstr} \cw{dupstr()}
3236 \c char *dupstr(const char *s);
3238 This function dynamically allocates a duplicate of a C string. Like
3239 the \cw{snew()} functions, it guarantees to return non-\cw{NULL} or
3242 (Many platforms provide the function \cw{strdup()}. As well as
3243 guaranteeing never to return \cw{NULL}, my version has the advantage
3244 of being defined \e{everywhere}, rather than inconveniently not
3247 \S{utils-free-cfg} \cw{free_cfg()}
3249 \c void free_cfg(config_item *cfg);
3251 This function correctly frees an array of \c{config_item}s,
3252 including walking the array until it gets to the end and freeing
3253 precisely those \c{sval} fields which are expected to be dynamically
3256 (See \k{backend-configure} for details of the \c{config_item}
3259 \H{utils-tree234} Sorted and counted tree functions
3261 Many games require complex algorithms for generating random puzzles,
3262 and some require moderately complex algorithms even during play. A
3263 common requirement during these algorithms is for a means of
3264 maintaining sorted or unsorted lists of items, such that items can
3265 be removed and added conveniently.
3267 For general use, Puzzles provides the following set of functions
3268 which maintain 2-3-4 trees in memory. (A 2-3-4 tree is a balanced
3269 tree structure, with the property that all lookups, insertions,
3270 deletions, splits and joins can be done in \cw{O(log N)} time.)
3272 All these functions expect you to be storing a tree of \c{void *}
3273 pointers. You can put anything you like in those pointers.
3275 By the use of per-node element counts, these tree structures have
3276 the slightly unusual ability to look elements up by their numeric
3277 index within the list represented by the tree. This means that they
3278 can be used to store an unsorted list (in which case, every time you
3279 insert a new element, you must explicitly specify the position where
3280 you wish to insert it). They can also do numeric lookups in a sorted
3281 tree, which might be useful for (for example) tracking the median of
3282 a changing data set.
3284 As well as storing sorted lists, these functions can be used for
3285 storing \q{maps} (associative arrays), by defining each element of a
3286 tree to be a (key, value) pair.
3288 \S{utils-newtree234} \cw{newtree234()}
3290 \c tree234 *newtree234(cmpfn234 cmp);
3292 Creates a new empty tree, and returns a pointer to it.
3294 The parameter \c{cmp} determines the sorting criterion on the tree.
3297 \c typedef int (*cmpfn234)(void *, void *);
3299 If you want a sorted tree, you should provide a function matching
3300 this prototype, which returns like \cw{strcmp()} does (negative if
3301 the first argument is smaller than the second, positive if it is
3302 bigger, zero if they compare equal). In this case, the function
3303 \cw{addpos234()} will not be usable on your tree (because all
3304 insertions must respect the sorting order).
3306 If you want an unsorted tree, pass \cw{NULL}. In this case you will
3307 not be able to use either \cw{add234()} or \cw{del234()}, or any
3308 other function such as \cw{find234()} which depends on a sorting
3309 order. Your tree will become something more like an array, except
3310 that it will efficiently support insertion and deletion as well as
3311 lookups by numeric index.
3313 \S{utils-freetree234} \cw{freetree234()}
3315 \c void freetree234(tree234 *t);
3317 Frees a tree. This function will not free the \e{elements} of the
3318 tree (because they might not be dynamically allocated, or you might
3319 be storing the same set of elements in more than one tree); it will
3320 just free the tree structure itself. If you want to free all the
3321 elements of a tree, you should empty it before passing it to
3322 \cw{freetree234()}, by means of code along the lines of
3324 \c while ((element = delpos234(tree, 0)) != NULL)
3325 \c sfree(element); /* or some more complicated free function */
3326 \e iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
3328 \S{utils-add234} \cw{add234()}
3330 \c void *add234(tree234 *t, void *e);
3332 Inserts a new element \c{e} into the tree \c{t}. This function
3333 expects the tree to be sorted; the new element is inserted according
3336 If an element comparing equal to \c{e} is already in the tree, then
3337 the insertion will fail, and the return value will be the existing
3338 element. Otherwise, the insertion succeeds, and \c{e} is returned.
3340 \S{utils-addpos234} \cw{addpos234()}
3342 \c void *addpos234(tree234 *t, void *e, int index);
3344 Inserts a new element into an unsorted tree. Since there is no
3345 sorting order to dictate where the new element goes, you must
3346 specify where you want it to go. Setting \c{index} to zero puts the
3347 new element right at the start of the list; setting \c{index} to the
3348 current number of elements in the tree puts the new element at the
3351 Return value is \c{e}, in line with \cw{add234()} (although this
3352 function cannot fail except by running out of memory, in which case
3353 it will bomb out and die rather than returning an error indication).
3355 \S{utils-index234} \cw{index234()}
3357 \c void *index234(tree234 *t, int index);
3359 Returns a pointer to the \c{index}th element of the tree, or
3360 \cw{NULL} if \c{index} is out of range. Elements of the tree are
3363 \S{utils-find234} \cw{find234()}
3365 \c void *find234(tree234 *t, void *e, cmpfn234 cmp);
3367 Searches for an element comparing equal to \c{e} in a sorted tree.
3369 If \c{cmp} is \cw{NULL}, the tree's ordinary comparison function
3370 will be used to perform the search. However, sometimes you don't
3371 want that; suppose, for example, each of your elements is a big
3372 structure containing a \c{char *} name field, and you want to find
3373 the element with a given name. You \e{could} achieve this by
3374 constructing a fake element structure, setting its name field
3375 appropriately, and passing it to \cw{find234()}, but you might find
3376 it more convenient to pass \e{just} a name string to \cw{find234()},
3377 supplying an alternative comparison function which expects one of
3378 its arguments to be a bare name and the other to be a large
3379 structure containing a name field.
3381 Therefore, if \c{cmp} is not \cw{NULL}, then it will be used to
3382 compare \c{e} to elements of the tree. The first argument passed to
3383 \c{cmp} will always be \c{e}; the second will be an element of the
3386 (See \k{utils-newtree234} for the definition of the \c{cmpfn234}
3387 function pointer type.)
3389 The returned value is the element found, or \cw{NULL} if the search
3392 \S{utils-findrel234} \cw{findrel234()}
3394 \c void *findrel234(tree234 *t, void *e, cmpfn234 cmp, int relation);
3396 This function is like \cw{find234()}, but has the additional ability
3397 to do a \e{relative} search. The additional parameter \c{relation}
3398 can be one of the following values:
3402 \dd Find only an element that compares equal to \c{e}. This is
3403 exactly the behaviour of \cw{find234()}.
3407 \dd Find the greatest element that compares strictly less than
3408 \c{e}. \c{e} may be \cw{NULL}, in which case it finds the greatest
3409 element in the whole tree (which could also be done by
3410 \cw{index234(t, count234(t)-1)}).
3414 \dd Find the greatest element that compares less than or equal to
3415 \c{e}. (That is, find an element that compares equal to \c{e} if
3416 possible, but failing that settle for something just less than it.)
3420 \dd Find the smallest element that compares strictly greater than
3421 \c{e}. \c{e} may be \cw{NULL}, in which case it finds the smallest
3422 element in the whole tree (which could also be done by
3423 \cw{index234(t, 0)}).
3427 \dd Find the smallest element that compares greater than or equal to
3428 \c{e}. (That is, find an element that compares equal to \c{e} if
3429 possible, but failing that settle for something just bigger than
3432 Return value, as before, is the element found or \cw{NULL} if no
3433 element satisfied the search criterion.
3435 \S{utils-findpos234} \cw{findpos234()}
3437 \c void *findpos234(tree234 *t, void *e, cmpfn234 cmp, int *index);
3439 This function is like \cw{find234()}, but has the additional feature
3440 of returning the index of the element found in the tree; that index
3441 is written to \c{*index} in the event of a successful search (a
3442 non-\cw{NULL} return value).
3444 \c{index} may be \cw{NULL}, in which case this function behaves
3445 exactly like \cw{find234()}.
3447 \S{utils-findrelpos234} \cw{findrelpos234()}
3449 \c void *findrelpos234(tree234 *t, void *e, cmpfn234 cmp, int relation,
3452 This function combines all the features of \cw{findrel234()} and
3455 \S{utils-del234} \cw{del234()}
3457 \c void *del234(tree234 *t, void *e);
3459 Finds an element comparing equal to \c{e} in the tree, deletes it,
3462 The input tree must be sorted.
3464 The element found might be \c{e} itself, or might merely compare
3467 Return value is \cw{NULL} if no such element is found.
3469 \S{utils-delpos234} \cw{delpos234()}
3471 \c void *delpos234(tree234 *t, int index);
3473 Deletes the element at position \c{index} in the tree, and returns
3476 Return value is \cw{NULL} if the index is out of range.
3478 \S{utils-count234} \cw{count234()}
3480 \c int count234(tree234 *t);
3482 Returns the number of elements currently in the tree.
3484 \S{utils-splitpos234} \cw{splitpos234()}
3486 \c tree234 *splitpos234(tree234 *t, int index, int before);
3488 Splits the input tree into two pieces at a given position, and
3489 creates a new tree containing all the elements on one side of that
3492 If \c{before} is \cw{TRUE}, then all the items at or after position
3493 \c{index} are left in the input tree, and the items before that
3494 point are returned in the new tree. Otherwise, the reverse happens:
3495 all the items at or after \c{index} are moved into the new tree, and
3496 those before that point are left in the old one.
3498 If \c{index} is equal to 0 or to the number of elements in the input
3499 tree, then one of the two trees will end up empty (and this is not
3500 an error condition). If \c{index} is further out of range in either
3501 direction, the operation will fail completely and return \cw{NULL}.
3503 This operation completes in \cw{O(log N)} time, no matter how large
3504 the tree or how balanced or unbalanced the split.
3506 \S{utils-split234} \cw{split234()}
3508 \c tree234 *split234(tree234 *t, void *e, cmpfn234 cmp, int rel);
3510 Splits a sorted tree according to its sort order.
3512 \c{rel} can be any of the relation constants described in
3513 \k{utils-findrel234}, \e{except} for \cw{REL234_EQ}. All the
3514 elements having that relation to \c{e} will be transferred into the
3515 new tree; the rest will be left in the old one.
3517 The parameter \c{cmp} has the same semantics as it does in
3518 \cw{find234()}: if it is not \cw{NULL}, it will be used in place of
3519 the tree's own comparison function when comparing elements to \c{e},
3520 in such a way that \c{e} itself is always the first of its two
3523 Again, this operation completes in \cw{O(log N)} time, no matter how
3524 large the tree or how balanced or unbalanced the split.
3526 \S{utils-join234} \cw{join234()}
3528 \c tree234 *join234(tree234 *t1, tree234 *t2);
3530 Joins two trees together by concatenating the lists they represent.
3531 All the elements of \c{t2} are moved into \c{t1}, in such a way that
3532 they appear \e{after} the elements of \c{t1}. The tree \c{t2} is
3533 freed; the return value is \c{t1}.
3535 If you apply this function to a sorted tree and it violates the sort
3536 order (i.e. the smallest element in \c{t2} is smaller than or equal
3537 to the largest element in \c{t1}), the operation will fail and
3540 This operation completes in \cw{O(log N)} time, no matter how large
3541 the trees being joined together.
3543 \S{utils-join234r} \cw{join234r()}
3545 \c tree234 *join234r(tree234 *t1, tree234 *t2);
3547 Joins two trees together in exactly the same way as \cw{join234()},
3548 but this time the combined tree is returned in \c{t2}, and \c{t1} is
3549 destroyed. The elements in \c{t1} still appear before those in
3552 Again, this operation completes in \cw{O(log N)} time, no matter how
3553 large the trees being joined together.
3555 \S{utils-copytree234} \cw{copytree234()}
3557 \c tree234 *copytree234(tree234 *t, copyfn234 copyfn,
3558 \c void *copyfnstate);
3560 Makes a copy of an entire tree.
3562 If \c{copyfn} is \cw{NULL}, the tree will be copied but the elements
3563 will not be; i.e. the new tree will contain pointers to exactly the
3564 same physical elements as the old one.
3566 If you want to copy each actual element during the operation, you
3567 can instead pass a function in \c{copyfn} which makes a copy of each
3568 element. That function has the prototype
3570 \c typedef void *(*copyfn234)(void *state, void *element);
3572 and every time it is called, the \c{state} parameter will be set to
3573 the value you passed in as \c{copyfnstate}.
3575 \H{utils-misc} Miscellaneous utility functions and macros
3577 This section contains all the utility functions which didn't
3578 sensibly fit anywhere else.
3580 \S{utils-truefalse} \cw{TRUE} and \cw{FALSE}
3582 The main Puzzles header file defines the macros \cw{TRUE} and
3583 \cw{FALSE}, which are used throughout the code in place of 1 and 0
3584 (respectively) to indicate that the values are in a boolean context.
3585 For code base consistency, I'd prefer it if submissions of new code
3586 followed this convention as well.
3588 \S{utils-maxmin} \cw{max()} and \cw{min()}
3590 The main Puzzles header file defines the pretty standard macros
3591 \cw{max()} and \cw{min()}, each of which is given two arguments and
3592 returns the one which compares greater or less respectively.
3594 These macros may evaluate their arguments multiple times. Avoid side
3597 \S{utils-pi} \cw{PI}
3599 The main Puzzles header file defines a macro \cw{PI} which expands
3600 to a floating-point constant representing pi.
3602 (I've never understood why ANSI's \cw{<math.h>} doesn't define this.
3605 \S{utils-obfuscate-bitmap} \cw{obfuscate_bitmap()}
3607 \c void obfuscate_bitmap(unsigned char *bmp, int bits, int decode);
3609 This function obscures the contents of a piece of data, by
3610 cryptographic methods. It is useful for games of hidden information
3611 (such as Mines, Guess or Black Box), in which the game ID
3612 theoretically reveals all the information the player is supposed to
3613 be trying to guess. So in order that players should be able to send
3614 game IDs to one another without accidentally spoiling the resulting
3615 game by looking at them, these games obfuscate their game IDs using
3618 Although the obfuscation function is cryptographic, it cannot
3619 properly be called encryption because it has no key. Therefore,
3620 anybody motivated enough can re-implement it, or hack it out of the
3621 Puzzles source, and strip the obfuscation off one of these game IDs
3622 to see what lies beneath. (Indeed, they could usually do it much
3623 more easily than that, by entering the game ID into their own copy
3624 of the puzzle and hitting Solve.) The aim is not to protect against
3625 a determined attacker; the aim is simply to protect people who
3626 wanted to play the game honestly from \e{accidentally} spoiling
3629 The input argument \c{bmp} points at a piece of memory to be
3630 obfuscated. \c{bits} gives the length of the data. Note that that
3631 length is in \e{bits} rather than bytes: if you ask for obfuscation
3632 of a partial number of bytes, then you will get it. Bytes are
3633 considered to be used from the top down: thus, for example, setting
3634 \c{bits} to 10 will cover the whole of \cw{bmp[0]} and the \e{top
3635 two} bits of \cw{bmp[1]}. The remainder of a partially used byte is
3636 undefined (i.e. it may be corrupted by the function).
3638 The parameter \c{decode} is \cw{FALSE} for an encoding operation,
3639 and \cw{TRUE} for a decoding operation. Each is the inverse of the
3640 other. (There's no particular reason you shouldn't obfuscate by
3641 decoding and restore cleartext by encoding, if you really wanted to;
3642 it should still work.)
3644 The input bitmap is processed in place.
3646 \S{utils-bin2hex} \cw{bin2hex()}
3648 \c char *bin2hex(const unsigned char *in, int inlen);
3650 This function takes an input byte array and converts it into an
3651 ASCII string encoding those bytes in (lower-case) hex. It returns a
3652 dynamically allocated string containing that encoding.
3654 This function is useful for encoding the result of
3655 \cw{obfuscate_bitmap()} in printable ASCII for use in game IDs.
3657 \S{utils-hex2bin} \cw{hex2bin()}
3659 \c unsigned char *hex2bin(const char *in, int outlen);
3661 This function takes an ASCII string containing hex digits, and
3662 converts it back into a byte array of length \c{outlen}. If there
3663 aren't enough hex digits in the string, the contents of the
3664 resulting array will be undefined.
3666 This function is the inverse of \cw{bin2hex()}.
3668 \S{utils-game-mkhighlight} \cw{game_mkhighlight()}
3670 \c void game_mkhighlight(frontend *fe, float *ret,
3671 \c int background, int highlight, int lowlight);
3673 It's reasonably common for a puzzle game's graphics to use
3674 highlights and lowlights to indicate \q{raised} or \q{lowered}
3675 sections. Fifteen, Sixteen and Twiddle are good examples of this.
3677 Puzzles using this graphical style are running a risk if they just
3678 use whatever background colour is supplied to them by the front end,
3679 because that background colour might be too light to see any
3680 highlights on at all. (In particular, it's not unheard of for the
3681 front end to specify a default background colour of white.)
3683 Therefore, such puzzles can call this utility function from their
3684 \cw{colours()} routine (\k{backend-colours}). You pass it your front
3685 end handle, a pointer to the start of your return array, and three
3686 colour indices. It will:
3688 \b call \cw{frontend_default_colour()} (\k{frontend-default-colour})
3689 to fetch the front end's default background colour
3691 \b alter the brightness of that colour if it's unsuitable
3693 \b define brighter and darker variants of the colour to be used as
3694 highlights and lowlights
3696 \b write those results into the relevant positions in the \c{ret}
3699 Thus, \cw{ret[background*3]} to \cw{ret[background*3+2]} will be set
3700 to RGB values defining a sensible background colour, and similary
3701 \c{highlight} and \c{lowlight} will be set to sensible colours.
3703 \C{writing} How to write a new puzzle
3705 This chapter gives a guide to how to actually write a new puzzle:
3706 where to start, what to do first, how to solve common problems.
3708 The previous chapters have been largely composed of facts. This one
3711 \H{writing-editorial} Choosing a puzzle
3713 Before you start writing a puzzle, you have to choose one. Your
3714 taste in puzzle games is up to you, of course; and, in fact, you're
3715 probably reading this guide because you've \e{already} thought of a
3716 game you want to write. But if you want to get it accepted into the
3717 official Puzzles distribution, then there's a criterion it has to
3720 The current Puzzles editorial policy is that all games should be
3721 \e{fair}. A fair game is one which a player can only fail to
3722 complete through demonstrable lack of skill \dash that is, such that
3723 a better player in the same situation would have \e{known} to do
3724 something different.
3726 For a start, that means every game presented to the user must have
3727 \e{at least one solution}. Giving the unsuspecting user a puzzle
3728 which is actually impossible is not acceptable. (There is an
3729 exception: if the user has selected some non-default option which is
3730 clearly labelled as potentially unfair, \e{then} you're allowed to
3731 generate possibly insoluble puzzles, because the user isn't
3732 unsuspecting any more. Same Game and Mines both have options of this
3735 Also, this actually \e{rules out} games such as Klondike, or the
3736 normal form of Mahjong Solitaire. Those games have the property that
3737 even if there is a solution (i.e. some sequence of moves which will
3738 get from the start state to the solved state), the player doesn't
3739 necessarily have enough information to \e{find} that solution. In
3740 both games, it is possible to reach a dead end because you had an
3741 arbitrary choice to make and made it the wrong way. This violates
3742 the fairness criterion, because a better player couldn't have known
3743 they needed to make the other choice.
3745 (GNOME has a variant on Mahjong Solitaire which makes it fair: there
3746 is a Shuffle operation which randomly permutes all the remaining
3747 tiles without changing their positions, which allows you to get out
3748 of a sticky situation. Using this operation adds a 60-second penalty
3749 to your solution time, so it's to the player's advantage to try to
3750 minimise the chance of having to use it. It's still possible to
3751 render the game uncompletable if you end up with only two tiles
3752 vertically stacked, but that's easy to foresee and avoid using a
3753 shuffle operation. This form of the game \e{is} fair. Implementing
3754 it in Puzzles would require an infrastructure change so that the
3755 back end could communicate time penalties to the mid-end, but that
3756 would be easy enough.)
3758 Providing a \e{unique} solution is a little more negotiable; it
3759 depends on the puzzle. Solo would have been of unacceptably low
3760 quality if it didn't always have a unique solution, whereas Twiddle
3761 inherently has multiple solutions by its very nature and it would
3762 have been meaningless to even \e{suggest} making it uniquely
3763 soluble. Somewhere in between, Flip could reasonably be made to have
3764 unique solutions (by enforcing a zero-dimension kernel in every
3765 generated matrix) but it doesn't seem like a serious quality problem
3768 Of course, you don't \e{have} to care about all this. There's
3769 nothing stopping you implementing any puzzle you want to if you're
3770 happy to maintain your puzzle yourself, distribute it from your own
3771 web site, fork the Puzzles code completely, or anything like that.
3772 It's free software; you can do what you like with it. But any game
3773 that you want to be accepted into \e{my} Puzzles code base has to
3774 satisfy the fairness criterion, which means all randomly generated
3775 puzzles must have a solution (unless the user has deliberately
3776 chosen otherwise) and it must be possible \e{in theory} to find that
3777 solution without having to guess.
3779 \H{writing-gs} Getting started
3781 The simplest way to start writing a new puzzle is to copy
3782 \c{nullgame.c}. This is a template puzzle source file which does
3783 almost nothing, but which contains all the back end function
3784 prototypes and declares the back end data structure correctly. It is
3785 built every time the rest of Puzzles is built, to ensure that it
3786 doesn't get out of sync with the code and remains buildable.
3788 So start by copying \c{nullgame.c} into your new source file. Then
3789 you'll gradually add functionality until the very boring Null Game
3790 turns into your real game.
3792 Next you'll need to add your puzzle to the Makefiles, in order to
3793 compile it conveniently. \e{Do not edit the Makefiles}: they are
3794 created automatically by the script \c{mkfiles.pl}, from the file
3795 called \c{Recipe}. Edit \c{Recipe}, and then re-run \c{mkfiles.pl}.
3797 Also, don't forget to add your puzzle to \c{list.c}: if you don't,
3798 then it will still run fine on platforms which build each puzzle
3799 separately, but Mac OS X and other monolithic platforms will not
3800 include your new puzzle in their single binary.
3802 Once your source file is building, you can move on to the fun bit.
3804 \S{writing-generation} Puzzle generation
3806 Randomly generating instances of your puzzle is almost certain to be
3807 the most difficult part of the code, and also the task with the
3808 highest chance of turning out to be completely infeasible. Therefore
3809 I strongly recommend doing it \e{first}, so that if it all goes
3810 horribly wrong you haven't wasted any more time than you absolutely
3811 had to. What I usually do is to take an unmodified \c{nullgame.c},
3812 and start adding code to \cw{new_game_desc()} which tries to
3813 generate a puzzle instance and print it out using \cw{printf()}.
3814 Once that's working, \e{then} I start connecting it up to the return
3815 value of \cw{new_game_desc()}, populating other structures like
3816 \c{game_params}, and generally writing the rest of the source file.
3818 There are many ways to generate a puzzle which is known to be
3819 soluble. In this section I list all the methods I currently know of,
3820 in case any of them can be applied to your puzzle. (Not all of these
3821 methods will work, or in some cases even make sense, for all
3824 Some puzzles are mathematically tractable, meaning you can work out
3825 in advance which instances are soluble. Sixteen, for example, has a
3826 parity constraint in some settings which renders exactly half the
3827 game space unreachable, but it can be mathematically proved that any
3828 position not in that half \e{is} reachable. Therefore, Sixteen's
3829 grid generation simply consists of selecting at random from a well
3830 defined subset of the game space. Cube in its default state is even
3831 easier: \e{every} possible arrangement of the blue squares and the
3832 cube's starting position is soluble!
3834 Another option is to redefine what you mean by \q{soluble}. Black
3835 Box takes this approach. There are layouts of balls in the box which
3836 are completely indistinguishable from one another no matter how many
3837 beams you fire into the box from which angles, which would normally
3838 be grounds for declaring those layouts unfair; but fortunately,
3839 detecting that indistinguishability is computationally easy. So
3840 Black Box doesn't demand that your ball placements match its own; it
3841 merely demands that your ball placements be \e{indistinguishable}
3842 from the ones it was thinking of. If you have an ambiguous puzzle,
3843 then any of the possible answers is considered to be a solution.
3844 Having redefined the rules in that way, any puzzle is soluble again.
3846 Those are the simple techniques. If they don't work, you have to get
3849 One way to generate a soluble puzzle is to start from the solved
3850 state and make inverse moves until you reach a starting state. Then
3851 you know there's a solution, because you can just list the inverse
3852 moves you made and make them in the opposite order to return to the
3855 This method can be simple and effective for puzzles where you get to
3856 decide what's a starting state and what's not. In Pegs, for example,
3857 the generator begins with one peg in the centre of the board and
3858 makes inverse moves until it gets bored; in this puzzle, valid
3859 inverse moves are easy to detect, and \e{any} state that's reachable
3860 from the solved state by inverse moves is a reasonable starting
3861 position. So Pegs just continues making inverse moves until the
3862 board satisfies some criteria about extent and density, and then
3863 stops and declares itself done.
3865 For other puzzles, it can be a lot more difficult. Same Game uses
3866 this strategy too, and it's lucky to get away with it at all: valid
3867 inverse moves aren't easy to find (because although it's easy to
3868 insert additional squares in a Same Game position, it's difficult to
3869 arrange that \e{after} the insertion they aren't adjacent to any
3870 other squares of the same colour), so you're constantly at risk of
3871 running out of options and having to backtrack or start again. Also,
3872 Same Game grids never start off half-empty, which means you can't
3873 just stop when you run out of moves \dash you have to find a way to
3874 fill the grid up \e{completely}.
3876 The other way to generate a puzzle that's soluble is to start from
3877 the other end, and actually write a \e{solver}. This tends to ensure
3878 that a puzzle has a \e{unique} solution over and above having a
3879 solution at all, so it's a good technique to apply to puzzles for
3880 which that's important.
3882 One theoretical drawback of generating soluble puzzles by using a
3883 solver is that your puzzles are restricted in difficulty to those
3884 which the solver can handle. (Most solvers are not fully general:
3885 many sets of puzzle rules are NP-complete or otherwise nasty, so
3886 most solvers can only handle a subset of the theoretically soluble
3887 puzzles.) It's been my experience in practice, however, that this
3888 usually isn't a problem; computers are good at very different things
3889 from humans, and what the computer thinks is nice and easy might
3890 still be pleasantly challenging for a human. For example, when
3891 solving Dominosa puzzles I frequently find myself using a variety of
3892 reasoning techniques that my solver doesn't know about; in
3893 principle, therefore, I should be able to solve the puzzle using
3894 only those techniques it \e{does} know about, but this would involve
3895 repeatedly searching the entire grid for the one simple deduction I
3896 can make. Computers are good at this sort of exhaustive search, but
3897 it's been my experience that human solvers prefer to do more complex
3898 deductions than to spend ages searching for simple ones. So in many
3899 cases I don't find my own playing experience to be limited by the
3900 restrictions on the solver.
3902 (This isn't \e{always} the case. Solo is a counter-example;
3903 generating Solo puzzles using a simple solver does lead to
3904 qualitatively easier puzzles. Therefore I had to make the Solo
3905 solver rather more advanced than most of them.)
3907 There are several different ways to apply a solver to the problem of
3908 generating a soluble puzzle. I list a few of them below.
3910 The simplest approach is brute force: randomly generate a puzzle,
3911 use the solver to see if it's soluble, and if not, throw it away and
3912 try again until you get lucky. This is often a viable technique if
3913 all else fails, but it tends not to scale well: for many puzzle
3914 types, the probability of finding a uniquely soluble instance
3915 decreases sharply as puzzle size goes up, so this technique might
3916 work reasonably fast for small puzzles but take (almost) forever at
3917 larger sizes. Still, if there's no other alternative it can be
3918 usable: Pattern and Dominosa both use this technique. (However,
3919 Dominosa has a means of tweaking the randomly generated grids to
3920 increase the \e{probability} of them being soluble, by ruling out
3921 one of the most common ambiguous cases. This improved generation
3922 speed by over a factor of 10 on the highest preset!)
3924 An approach which can be more scalable involves generating a grid
3925 and then tweaking it to make it soluble. This is the technique used
3926 by Mines and also by Net: first a random puzzle is generated, and
3927 then the solver is run to see how far it gets. Sometimes the solver
3928 will get stuck; when that happens, examine the area it's having
3929 trouble with, and make a small random change in that area to allow
3930 it to make more progress. Continue solving (possibly even without
3931 restarting the solver), tweaking as necessary, until the solver
3932 finishes. Then restart the solver from the beginning to ensure that
3933 the tweaks haven't caused new problems in the process of solving old
3934 ones (which can sometimes happen).
3936 This strategy works well in situations where the usual solver
3937 failure mode is to get stuck in an easily localised spot. Thus it
3938 works well for Net and Mines, whose most common failure mode tends
3939 to be that most of the grid is fine but there are a few widely
3940 separated ambiguous sections; but it would work less well for
3941 Dominosa, in which the way you get stuck is to have scoured the
3942 whole grid and not found anything you can deduce \e{anywhere}. Also,
3943 it relies on there being a low probability that tweaking the grid
3944 introduces a new problem at the same time as solving the old one;
3945 Mines and Net also have the property that most of their deductions
3946 are local, so that it's very unlikely for a tweak to affect
3947 something half way across the grid from the location where it was
3948 applied. In Dominosa, by contrast, a lot of deductions use
3949 information about half the grid (\q{out of all the sixes, only one
3950 is next to a three}, which can depend on the values of up to 32 of
3951 the 56 squares in the default setting!), so this tweaking strategy
3952 would be rather less likely to work well.
3954 A more specialised strategy is that used in Solo and Slant. These
3955 puzzles have the property that they derive their difficulty from not
3956 presenting all the available clues. (In Solo's case, if all the
3957 possible clues were provided then the puzzle would already be
3958 solved; in Slant it would still require user action to fill in the
3959 lines, but it would present no challenge at all). Therefore, a
3960 simple generation technique is to leave the decision of which clues
3961 to provide until the last minute. In other words, first generate a
3962 random \e{filled} grid with all possible clues present, and then
3963 gradually remove clues for as long as the solver reports that it's
3964 still soluble. Unlike the methods described above, this technique
3965 \e{cannot} fail \dash once you've got a filled grid, nothing can
3966 stop you from being able to convert it into a viable puzzle.
3967 However, it wouldn't even be meaningful to apply this technique to
3968 (say) Pattern, in which clues can never be left out, so the only way
3969 to affect the set of clues is by altering the solution.
3971 (Unfortunately, Solo is complicated by the need to provide puzzles
3972 at varying difficulty levels. It's easy enough to generate a puzzle
3973 of \e{at most} a given level of difficulty; you just have a solver
3974 with configurable intelligence, and you set it to a given level and
3975 apply the above technique, thus guaranteeing that the resulting grid
3976 is solvable by someone with at most that much intelligence. However,
3977 generating a puzzle of \e{at least} a given level of difficulty is
3978 rather harder; if you go for \e{at most} Intermediate level, you're
3979 likely to find that you've accidentally generated a Trivial grid a
3980 lot of the time, because removing just one number is sufficient to
3981 take the puzzle from Trivial straight to Ambiguous. In that
3982 situation Solo has no remaining options but to throw the puzzle away
3985 A final strategy is to use the solver \e{during} puzzle
3986 construction: lay out a bit of the grid, run the solver to see what
3987 it allows you to deduce, and then lay out a bit more to allow the
3988 solver to make more progress. There are articles on the web that
3989 recommend constructing Sudoku puzzles by this method (which is
3990 completely the opposite way round to how Solo does it); for Sudoku
3991 it has the advantage that you get to specify your clue squares in
3992 advance (so you can have them make pretty patterns).
3994 Rectangles uses a strategy along these lines. First it generates a
3995 grid by placing the actual rectangles; then it has to decide where
3996 in each rectangle to place a number. It uses a solver to help it
3997 place the numbers in such a way as to ensure a unique solution. It
3998 does this by means of running a test solver, but it runs the solver
3999 \e{before} it's placed any of the numbers \dash which means the
4000 solver must be capable of coping with uncertainty about exactly
4001 where the numbers are! It runs the solver as far as it can until it
4002 gets stuck; then it narrows down the possible positions of a number
4003 in order to allow the solver to make more progress, and so on. Most
4004 of the time this process terminates with the grid fully solved, at
4005 which point any remaining number-placement decisions can be made at
4006 random from the options not so far ruled out. Note that unlike the
4007 Net/Mines tweaking strategy described above, this algorithm does not
4008 require a checking run after it completes: if it finishes
4009 successfully at all, then it has definitely produced a uniquely
4012 Most of the strategies described above are not 100% reliable. Each
4013 one has a failure rate: every so often it has to throw out the whole
4014 grid and generate a fresh one from scratch. (Solo's strategy would
4015 be the exception, if it weren't for the need to provide configurable
4016 difficulty levels.) Occasional failures are not a fundamental
4017 problem in this sort of work, however: it's just a question of
4018 dividing the grid generation time by the success rate (if it takes
4019 10ms to generate a candidate grid and 1/5 of them work, then it will
4020 take 50ms on average to generate a viable one), and seeing whether
4021 the expected time taken to \e{successfully} generate a puzzle is
4022 unacceptably slow. Dominosa's generator has a very low success rate
4023 (about 1 out of 20 candidate grids turn out to be usable, and if you
4024 think \e{that's} bad then go and look at the source code and find
4025 the comment showing what the figures were before the generation-time
4026 tweaks!), but the generator itself is very fast so this doesn't
4027 matter. Rectangles has a slower generator, but fails well under 50%
4030 So don't be discouraged if you have an algorithm that doesn't always
4031 work: if it \e{nearly} always works, that's probably good enough.
4032 The one place where reliability is important is that your algorithm
4033 must never produce false positives: it must not claim a puzzle is
4034 soluble when it isn't. It can produce false negatives (failing to
4035 notice that a puzzle is soluble), and it can fail to generate a
4036 puzzle at all, provided it doesn't do either so often as to become
4039 One last piece of advice: for grid-based puzzles, when writing and
4040 testing your generation algorithm, it's almost always a good idea
4041 \e{not} to test it initially on a grid that's square (i.e.
4042 \cw{w==h}), because if the grid is square then you won't notice if
4043 you mistakenly write \c{h} instead of \c{w} (or vice versa)
4044 somewhere in the code. Use a rectangular grid for testing, and any
4045 size of grid will be likely to work after that.
4047 \S{writing-textformats} Designing textual description formats
4049 Another aspect of writing a puzzle which is worth putting some
4050 thought into is the design of the various text description formats:
4051 the format of the game parameter encoding, the game description
4052 encoding, and the move encoding.
4054 The first two of these should be reasonably intuitive for a user to
4055 type in; so provide some flexibility where possible. Suppose, for
4056 example, your parameter format consists of two numbers separated by
4057 an \c{x} to specify the grid dimensions (\c{10x10} or \c{20x15}),
4058 and then has some suffixes to specify other aspects of the game
4059 type. It's almost always a good idea in this situation to arrange
4060 that \cw{decode_params()} can handle the suffixes appearing in any
4061 order, even if \cw{encode_params()} only ever generates them in one
4064 These formats will also be expected to be reasonably stable: users
4065 will expect to be able to exchange game IDs with other users who
4066 aren't running exactly the same version of your game. So make them
4067 robust and stable: don't build too many assumptions into the game ID
4068 format which will have to be changed every time something subtle
4069 changes in the puzzle code.
4071 \H{writing-howto} Common how-to questions
4073 This section lists some common things people want to do when writing
4074 a puzzle, and describes how to achieve them within the Puzzles
4077 \S{writing-howto-cursor} Drawing objects at only one position
4079 A common phenomenon is to have an object described in the
4080 \c{game_state} or the \c{game_ui} which can only be at one position.
4081 A cursor \dash probably specified in the \c{game_ui} \dash is a good
4084 In the \c{game_ui}, it would \e{obviously} be silly to have an array
4085 covering the whole game grid with a boolean flag stating whether the
4086 cursor was at each position. Doing that would waste space, would
4087 make it difficult to find the cursor in order to do anything with
4088 it, and would introduce the potential for synchronisation bugs in
4089 which you ended up with two cursors or none. The obviously sensible
4090 way to store a cursor in the \c{game_ui} is to have fields directly
4091 encoding the cursor's coordinates.
4093 However, it is a mistake to assume that the same logic applies to
4094 the \c{game_drawstate}. If you replicate the cursor position fields
4095 in the draw state, the redraw code will get very complicated. In the
4096 draw state, in fact, it \e{is} probably the right thing to have a
4097 cursor flag for every position in the grid. You probably have an
4098 array for the whole grid in the drawstate already (stating what is
4099 currently displayed in the window at each position); the sensible
4100 approach is to add a \q{cursor} flag to each element of that array.
4101 Then the main redraw loop will look something like this
4104 \c for (y = 0; y < h; y++) {
4105 \c for (x = 0; x < w; x++) {
4106 \c int value = state->symbol_at_position[y][x];
4107 \c if (x == ui->cursor_x && y == ui->cursor_y)
4109 \c if (ds->symbol_at_position[y][x] != value) {
4110 \c symbol_drawing_subroutine(dr, ds, x, y, value);
4111 \c ds->symbol_at_position[y][x] = value;
4116 This loop is very simple, pretty hard to get wrong, and
4117 \e{automatically} deals both with erasing the previous cursor and
4118 drawing the new one, with no special case code required.
4120 This type of loop is generally a sensible way to write a redraw
4121 function, in fact. The best thing is to ensure that the information
4122 stored in the draw state for each position tells you \e{everything}
4123 about what was drawn there. A good way to ensure that is to pass
4124 precisely the same information, and \e{only} that information, to a
4125 subroutine that does the actual drawing; then you know there's no
4126 additional information which affects the drawing but which you don't
4129 \S{writing-keyboard-cursor} Implementing a keyboard-controlled cursor
4131 It is often useful to provide a keyboard control method in a
4132 basically mouse-controlled game. A keyboard-controlled cursor is
4133 best implemented by storing its location in the \c{game_ui} (since
4134 if it were in the \c{game_state} then the user would have to
4135 separately undo every cursor move operation). So the procedure would
4138 \b Put cursor position fields in the \c{game_ui}.
4140 \b \cw{interpret_move()} responds to arrow keys by modifying the
4141 cursor position fields and returning \cw{""}.
4143 \b \cw{interpret_move()} responds to some sort of fire button by
4144 actually performing a move based on the current cursor location.
4146 \b You might want an additional \c{game_ui} field stating whether
4147 the cursor is currently visible, and having it disappear when a
4148 mouse action occurs (so that it doesn't clutter the display when not
4151 \b You might also want to automatically hide the cursor in
4152 \cw{changed_state()} when the current game state changes to one in
4153 which there is no move to make (which is the case in some types of
4156 \b \cw{redraw()} draws the cursor using the technique described in
4157 \k{writing-howto-cursor}.
4159 \S{writing-howto-dragging} Implementing draggable sprites
4161 Some games have a user interface which involves dragging some sort
4162 of game element around using the mouse. If you need to show a
4163 graphic moving smoothly over the top of other graphics, use a
4164 blitter (see \k{drawing-blitter} for the blitter API) to save the
4165 background underneath it. The typical scenario goes:
4167 \b Have a blitter field in the \c{game_drawstate}.
4169 \b Set the blitter field to \cw{NULL} in the game's
4170 \cw{new_drawstate()} function, since you don't yet know how big the
4171 piece of saved background needs to be.
4173 \b In the game's \cw{set_size()} function, once you know the size of
4174 the object you'll be dragging around the display and hence the
4175 required size of the blitter, actually allocate the blitter.
4177 \b In \cw{free_drawstate()}, free the blitter if it's not \cw{NULL}.
4179 \b In \cw{interpret_move()}, respond to mouse-down and mouse-drag
4180 events by updating some fields in the \cw{game_ui} which indicate
4181 that a drag is in progress.
4183 \b At the \e{very end} of \cw{redraw()}, after all other drawing has
4184 been done, draw the moving object if there is one. First save the
4185 background under the object in the blitter; then set a clip
4186 rectangle covering precisely the area you just saved (just in case
4187 anti-aliasing or some other error causes your drawing to go beyond
4188 the area you saved). Then draw the object, and call \cw{unclip()}.
4189 Finally, set a flag in the \cw{game_drawstate} that indicates that
4190 the blitter needs restoring.
4192 \b At the very start of \cw{redraw()}, before doing anything else at
4193 all, check the flag in the \cw{game_drawstate}, and if it says the
4194 blitter needs restoring then restore it. (Then clear the flag, so
4195 that this won't happen again in the next redraw if no moving object
4196 is drawn this time.)
4198 This way, you will be able to write the rest of the redraw function
4199 completely ignoring the dragged object, as if it were floating above
4200 your bitmap and being completely separate.
4202 \S{writing-ref-counting} Sharing large invariant data between all
4205 In some puzzles, there is a large amount of data which never changes
4206 between game states. The array of numbers in Dominosa is a good
4209 You \e{could} dynamically allocate a copy of that array in every
4210 \c{game_state}, and have \cw{dup_game()} make a fresh copy of it for
4211 every new \c{game_state}; but it would waste memory and time. A
4212 more efficient way is to use a reference-counted structure.
4214 \b Define a structure type containing the data in question, and also
4215 containing an integer reference count.
4217 \b Have a field in \c{game_state} which is a pointer to this
4220 \b In \cw{new_game()}, when creating a fresh game state at the start
4221 of a new game, create an instance of this structure, initialise it
4222 with the invariant data, and set its reference count to 1.
4224 \b In \cw{dup_game()}, rather than making a copy of the structure
4225 for the new game state, simply set the new game state to point at
4226 the same copy of the structure, and increment its reference count.
4228 \b In \cw{free_game()}, decrement the reference count in the
4229 structure pointed to by the game state; if the count reaches zero,
4232 This way, the invariant data will persist for only as long as it's
4233 genuinely needed; \e{as soon} as the last game state for a
4234 particular puzzle instance is freed, the invariant data for that
4235 puzzle will vanish as well. Reference counting is a very efficient
4236 form of garbage collection, when it works at all. (Which it does in
4237 this instance, of course, because there's no possibility of circular
4240 \S{writing-flash-types} Implementing multiple types of flash
4242 In some games you need to flash in more than one different way.
4243 Mines, for example, flashes white when you win, and flashes red when
4244 you tread on a mine and die.
4246 The simple way to do this is:
4248 \b Have a field in the \c{game_ui} which describes the type of flash.
4250 \b In \cw{flash_length()}, examine the old and new game states to
4251 decide whether a flash is required and what type. Write the type of
4252 flash to the \c{game_ui} field whenever you return non-zero.
4254 \b In \cw{redraw()}, when you detect that \c{flash_time} is
4255 non-zero, examine the field in \c{game_ui} to decide which type of
4258 \cw{redraw()} will never be called with \c{flash_time} non-zero
4259 unless \cw{flash_length()} was first called to tell the mid-end that
4260 a flash was required; so whenever \cw{redraw()} notices that
4261 \c{flash_time} is non-zero, you can be sure that the field in
4262 \c{game_ui} is correctly set.
4264 \S{writing-move-anim} Animating game moves
4266 A number of puzzle types benefit from a quick animation of each move
4269 For some games, such as Fifteen, this is particularly easy. Whenever
4270 \cw{redraw()} is called with \c{oldstate} non-\cw{NULL}, Fifteen
4271 simply compares the position of each tile in the two game states,
4272 and if the tile is not in the same place then it draws it some
4273 fraction of the way from its old position to its new position. This
4274 method copes automatically with undo.
4276 Other games are less obvious. In Sixteen, for example, you can't
4277 just draw each tile a fraction of the way from its old to its new
4278 position: if you did that, the end tile would zip very rapidly past
4279 all the others to get to the other end and that would look silly.
4280 (Worse, it would look inconsistent if the end tile was drawn on top
4281 going one way and on the bottom going the other way.)
4283 A useful trick here is to define a field or two in the game state
4284 that indicates what the last move was.
4286 \b Add a \q{last move} field to the \c{game_state} (or two or more
4287 fields if the move is complex enough to need them).
4289 \b \cw{new_game()} initialises this field to a null value for a new
4292 \b \cw{execute_move()} sets up the field to reflect the move it just
4295 \b \cw{redraw()} now needs to examine its \c{dir} parameter. If
4296 \c{dir} is positive, it determines the move being animated by
4297 looking at the last-move field in \c{newstate}; but if \c{dir} is
4298 negative, it has to look at the last-move field in \c{oldstate}, and
4299 invert whatever move it finds there.
4301 Note also that Sixteen needs to store the \e{direction} of the move,
4302 because you can't quite determine it by examining the row or column
4303 in question. You can in almost all cases, but when the row is
4304 precisely two squares long it doesn't work since a move in either
4305 direction looks the same. (You could argue that since moving a
4306 2-element row left and right has the same effect, it doesn't matter
4307 which one you animate; but in fact it's very disorienting to click
4308 the arrow left and find the row moving right, and almost as bad to
4309 undo a move to the right and find the game animating \e{another}
4312 \S{writing-conditional-anim} Animating drag operations
4314 In Untangle, moves are made by dragging a node from an old position
4315 to a new position. Therefore, at the time when the move is initially
4316 made, it should not be animated, because the node has already been
4317 dragged to the right place and doesn't need moving there. However,
4318 it's nice to animate the same move if it's later undone or redone.
4319 This requires a bit of fiddling.
4321 The obvious approach is to have a flag in the \c{game_ui} which
4322 inhibits move animation, and to set that flag in
4323 \cw{interpret_move()}. The question is, when would the flag be reset
4324 again? The obvious place to do so is \cw{changed_state()}, which
4325 will be called once per move. But it will be called \e{before}
4326 \cw{anim_length()}, so if it resets the flag then \cw{anim_length()}
4327 will never see the flag set at all.
4329 The solution is to have \e{two} flags in a queue.
4331 \b Define two flags in \c{game_ui}; let's call them \q{current} and
4334 \b Set both to \cw{FALSE} in \c{new_ui()}.
4336 \b When a drag operation completes in \cw{interpret_move()}, set the
4337 \q{next} flag to \cw{TRUE}.
4339 \b Every time \cw{changed_state()} is called, set the value of
4340 \q{current} to the value in \q{next}, and then set the value of
4341 \q{next} to \cw{FALSE}.
4343 \b That way, \q{current} will be \cw{TRUE} \e{after} a call to
4344 \cw{changed_state()} if and only if that call to
4345 \cw{changed_state()} was the result of a drag operation processed by
4346 \cw{interpret_move()}. Any other call to \cw{changed_state()}, due
4347 to an Undo or a Redo or a Restart or a Solve, will leave \q{current}
4350 \b So now \cw{anim_length()} can request a move animation if and
4351 only if the \q{current} flag is \e{not} set.
4353 \S{writing-cheating} Inhibiting the victory flash when Solve is used
4355 Many games flash when you complete them, as a visual congratulation
4356 for having got to the end of the puzzle. It often seems like a good
4357 idea to disable that flash when the puzzle is brought to a solved
4358 state by means of the Solve operation.
4360 This is easily done:
4362 \b Add a \q{cheated} flag to the \c{game_state}.
4364 \b Set this flag to \cw{FALSE} in \cw{new_game()}.
4366 \b Have \cw{solve()} return a move description string which clearly
4367 identifies the move as a solve operation.
4369 \b Have \cw{execute_move()} respond to that clear identification by
4370 setting the \q{cheated} flag in the returned \c{game_state}. The
4371 flag will then be propagated to all subsequent game states, even if
4372 the user continues fiddling with the game after it is solved.
4374 \b \cw{flash_length()} now returns non-zero if \c{oldstate} is not
4375 completed and \c{newstate} is, \e{and} neither state has the
4376 \q{cheated} flag set.
4378 \H{writing-testing} Things to test once your puzzle is written
4380 Puzzle implementations written in this framework are self-testing as
4381 far as I could make them.
4383 Textual game and move descriptions, for example, are generated and
4384 parsed as part of the normal process of play. Therefore, if you can
4385 make moves in the game \e{at all} you can be reasonably confident
4386 that the mid-end serialisation interface will function correctly and
4387 you will be able to save your game. (By contrast, if I'd stuck with
4388 a single \cw{make_move()} function performing the jobs of both
4389 \cw{interpret_move()} and \cw{execute_move()}, and had separate
4390 functions to encode and decode a game state in string form, then
4391 those functions would not be used during normal play; so they could
4392 have been completely broken, and you'd never know it until you tried
4393 to save the game \dash which would have meant you'd have to test
4394 game saving \e{extensively} and make sure to test every possible
4395 type of game state. As an added bonus, doing it the way I did leads
4396 to smaller save files.)
4398 There is one exception to this, which is the string encoding of the
4399 \c{game_ui}. Most games do not store anything permanent in the
4400 \c{game_ui}, and hence do not need to put anything in its encode and
4401 decode functions; but if there is anything in there, you do need to
4402 test game loading and saving to ensure those functions work
4405 It's also worth testing undo and redo of all operations, to ensure
4406 that the redraw and the animations (if any) work properly. Failing
4407 to animate undo properly seems to be a common error.
4409 Other than that, just use your common sense.