| 1 | /* |
| 2 | * misc.c: miscellaneous useful items |
| 3 | */ |
| 4 | |
| 5 | #include "halibut.h" |
| 6 | |
| 7 | struct stackTag { |
| 8 | void **data; |
| 9 | int sp; |
| 10 | int size; |
| 11 | }; |
| 12 | |
| 13 | stack stk_new(void) { |
| 14 | stack s; |
| 15 | |
| 16 | s = mknew(struct stackTag); |
| 17 | s->sp = 0; |
| 18 | s->size = 0; |
| 19 | s->data = NULL; |
| 20 | |
| 21 | return s; |
| 22 | } |
| 23 | |
| 24 | void stk_free(stack s) { |
| 25 | sfree(s->data); |
| 26 | sfree(s); |
| 27 | } |
| 28 | |
| 29 | void stk_push(stack s, void *item) { |
| 30 | if (s->size <= s->sp) { |
| 31 | s->size = s->sp + 32; |
| 32 | s->data = resize(s->data, s->size); |
| 33 | } |
| 34 | s->data[s->sp++] = item; |
| 35 | } |
| 36 | |
| 37 | void *stk_pop(stack s) { |
| 38 | if (s->sp > 0) |
| 39 | return s->data[--s->sp]; |
| 40 | else |
| 41 | return NULL; |
| 42 | } |
| 43 | |
| 44 | /* |
| 45 | * Small routines to amalgamate a string from an input source. |
| 46 | */ |
| 47 | const rdstring empty_rdstring = {0, 0, NULL}; |
| 48 | const rdstringc empty_rdstringc = {0, 0, NULL}; |
| 49 | |
| 50 | void rdadd(rdstring *rs, wchar_t c) { |
| 51 | if (rs->pos >= rs->size-1) { |
| 52 | rs->size = rs->pos + 128; |
| 53 | rs->text = resize(rs->text, rs->size); |
| 54 | } |
| 55 | rs->text[rs->pos++] = c; |
| 56 | rs->text[rs->pos] = 0; |
| 57 | } |
| 58 | void rdadds(rdstring *rs, wchar_t *p) { |
| 59 | int len = ustrlen(p); |
| 60 | if (rs->pos >= rs->size - len) { |
| 61 | rs->size = rs->pos + len + 128; |
| 62 | rs->text = resize(rs->text, rs->size); |
| 63 | } |
| 64 | ustrcpy(rs->text + rs->pos, p); |
| 65 | rs->pos += len; |
| 66 | } |
| 67 | wchar_t *rdtrim(rdstring *rs) { |
| 68 | rs->text = resize(rs->text, rs->pos + 1); |
| 69 | return rs->text; |
| 70 | } |
| 71 | |
| 72 | void rdaddc(rdstringc *rs, char c) { |
| 73 | if (rs->pos >= rs->size-1) { |
| 74 | rs->size = rs->pos + 128; |
| 75 | rs->text = resize(rs->text, rs->size); |
| 76 | } |
| 77 | rs->text[rs->pos++] = c; |
| 78 | rs->text[rs->pos] = 0; |
| 79 | } |
| 80 | void rdaddsc(rdstringc *rs, char *p) { |
| 81 | int len = strlen(p); |
| 82 | if (rs->pos >= rs->size - len) { |
| 83 | rs->size = rs->pos + len + 128; |
| 84 | rs->text = resize(rs->text, rs->size); |
| 85 | } |
| 86 | strcpy(rs->text + rs->pos, p); |
| 87 | rs->pos += len; |
| 88 | } |
| 89 | char *rdtrimc(rdstringc *rs) { |
| 90 | rs->text = resize(rs->text, rs->pos + 1); |
| 91 | return rs->text; |
| 92 | } |
| 93 | |
| 94 | int compare_wordlists(word *a, word *b) { |
| 95 | int t; |
| 96 | while (a && b) { |
| 97 | if (a->type != b->type) |
| 98 | return (a->type < b->type ? -1 : +1); /* FIXME? */ |
| 99 | t = a->type; |
| 100 | if ((t != word_Normal && t != word_Code && |
| 101 | t != word_WeakCode && t != word_Emph) || |
| 102 | a->alt || b->alt) { |
| 103 | int c; |
| 104 | if (a->text && b->text) { |
| 105 | c = ustricmp(a->text, b->text); |
| 106 | if (c) |
| 107 | return c; |
| 108 | } |
| 109 | c = compare_wordlists(a->alt, b->alt); |
| 110 | if (c) |
| 111 | return c; |
| 112 | a = a->next; |
| 113 | b = b->next; |
| 114 | } else { |
| 115 | wchar_t *ap = a->text, *bp = b->text; |
| 116 | while (*ap && *bp) { |
| 117 | wchar_t ac = utolower(*ap), bc = utolower(*bp); |
| 118 | if (ac != bc) |
| 119 | return (ac < bc ? -1 : +1); |
| 120 | if (!*++ap && a->next && a->next->type == t && !a->next->alt) |
| 121 | a = a->next, ap = a->text; |
| 122 | if (!*++bp && b->next && b->next->type == t && !b->next->alt) |
| 123 | b = b->next, bp = b->text; |
| 124 | } |
| 125 | if (*ap || *bp) |
| 126 | return (*ap ? +1 : -1); |
| 127 | a = a->next; |
| 128 | b = b->next; |
| 129 | } |
| 130 | } |
| 131 | |
| 132 | if (a || b) |
| 133 | return (a ? +1 : -1); |
| 134 | else |
| 135 | return 0; |
| 136 | } |
| 137 | |
| 138 | void mark_attr_ends(paragraph *sourceform) { |
| 139 | paragraph *p; |
| 140 | word *w, *wp; |
| 141 | for (p = sourceform; p; p = p->next) { |
| 142 | wp = NULL; |
| 143 | for (w = p->words; w; w = w->next) { |
| 144 | if (isattr(w->type)) { |
| 145 | int before = (wp && isattr(wp->type) && |
| 146 | sameattr(wp->type, w->type)); |
| 147 | int after = (w->next && isattr(w->next->type) && |
| 148 | sameattr(w->next->type, w->type)); |
| 149 | w->aux |= (before ? |
| 150 | (after ? attr_Always : attr_Last) : |
| 151 | (after ? attr_First : attr_Only)); |
| 152 | } |
| 153 | wp = w; |
| 154 | } |
| 155 | } |
| 156 | } |
| 157 | |
| 158 | wrappedline *wrap_para(word *text, int width, int subsequentwidth, |
| 159 | int (*widthfn)(word *)) { |
| 160 | wrappedline *head = NULL, **ptr = &head; |
| 161 | int nwords, wordsize; |
| 162 | struct wrapword { |
| 163 | word *begin, *end; |
| 164 | int width; |
| 165 | int spacewidth; |
| 166 | int cost; |
| 167 | int nwords; |
| 168 | } *wrapwords; |
| 169 | int i, j, n; |
| 170 | |
| 171 | /* |
| 172 | * Break the line up into wrappable components. |
| 173 | */ |
| 174 | nwords = wordsize = 0; |
| 175 | wrapwords = NULL; |
| 176 | while (text) { |
| 177 | if (nwords >= wordsize) { |
| 178 | wordsize = nwords + 64; |
| 179 | wrapwords = srealloc(wrapwords, wordsize * sizeof(*wrapwords)); |
| 180 | } |
| 181 | wrapwords[nwords].width = 0; |
| 182 | wrapwords[nwords].begin = text; |
| 183 | while (text) { |
| 184 | wrapwords[nwords].width += widthfn(text); |
| 185 | wrapwords[nwords].end = text->next; |
| 186 | if (text->next && (text->next->type == word_WhiteSpace || |
| 187 | text->next->type == word_EmphSpace || |
| 188 | text->breaks)) |
| 189 | break; |
| 190 | text = text->next; |
| 191 | } |
| 192 | if (text && text->next && (text->next->type == word_WhiteSpace || |
| 193 | text->next->type == word_EmphSpace)) { |
| 194 | wrapwords[nwords].spacewidth = widthfn(text->next); |
| 195 | text = text->next; |
| 196 | } else { |
| 197 | wrapwords[nwords].spacewidth = 0; |
| 198 | } |
| 199 | nwords++; |
| 200 | if (text) |
| 201 | text = text->next; |
| 202 | } |
| 203 | |
| 204 | /* |
| 205 | * Perform the dynamic wrapping algorithm: work backwards from |
| 206 | * nwords-1, determining the optimal wrapping for each terminal |
| 207 | * subsequence of the paragraph. |
| 208 | */ |
| 209 | for (i = nwords; i-- ;) { |
| 210 | int best = -1; |
| 211 | int bestcost = 0; |
| 212 | int cost; |
| 213 | int linelen = 0, spacewidth = 0; |
| 214 | int seenspace; |
| 215 | int thiswidth = (i == 0 ? width : subsequentwidth); |
| 216 | |
| 217 | j = 0; |
| 218 | seenspace = 0; |
| 219 | while (i+j < nwords) { |
| 220 | /* |
| 221 | * See what happens if we put j+1 words on this line. |
| 222 | */ |
| 223 | if (spacewidth) |
| 224 | seenspace = 1; |
| 225 | linelen += spacewidth + wrapwords[i+j].width; |
| 226 | spacewidth = wrapwords[i+j].spacewidth; |
| 227 | j++; |
| 228 | if (linelen > thiswidth) { |
| 229 | /* |
| 230 | * If we're over the width limit, abandon ship, |
| 231 | * _unless_ there is no best-effort yet (which will |
| 232 | * only happen if the first word is too long all by |
| 233 | * itself). |
| 234 | */ |
| 235 | if (best > 0) |
| 236 | break; |
| 237 | } |
| 238 | if (i+j == nwords) { |
| 239 | /* |
| 240 | * Special case: if we're at the very end of the |
| 241 | * paragraph, we don't score penalty points for the |
| 242 | * white space left on the line. |
| 243 | */ |
| 244 | cost = 0; |
| 245 | } else { |
| 246 | cost = (thiswidth-linelen) * (thiswidth-linelen); |
| 247 | cost += wrapwords[i+j].cost; |
| 248 | } |
| 249 | /* |
| 250 | * We compare bestcost >= cost, not bestcost > cost, |
| 251 | * because in cases where the costs are identical we |
| 252 | * want to try to look like the greedy algorithm, |
| 253 | * because readers are likely to have spent a lot of |
| 254 | * time looking at greedy-wrapped paragraphs and |
| 255 | * there's no point violating the Principle of Least |
| 256 | * Surprise if it doesn't actually gain anything. |
| 257 | */ |
| 258 | if (best < 0 || bestcost >= cost) { |
| 259 | bestcost = cost; |
| 260 | best = j; |
| 261 | } |
| 262 | } |
| 263 | /* |
| 264 | * Now we know the optimal answer for this terminal |
| 265 | * subsequence, so put it in wrapwords. |
| 266 | */ |
| 267 | wrapwords[i].cost = bestcost; |
| 268 | wrapwords[i].nwords = best; |
| 269 | } |
| 270 | |
| 271 | /* |
| 272 | * We've wrapped the paragraph. Now build the output |
| 273 | * `wrappedline' list. |
| 274 | */ |
| 275 | i = 0; |
| 276 | while (i < nwords) { |
| 277 | wrappedline *w = mknew(wrappedline); |
| 278 | *ptr = w; |
| 279 | ptr = &w->next; |
| 280 | w->next = NULL; |
| 281 | |
| 282 | n = wrapwords[i].nwords; |
| 283 | w->begin = wrapwords[i].begin; |
| 284 | w->end = wrapwords[i+n-1].end; |
| 285 | |
| 286 | /* |
| 287 | * Count along the words to find nspaces and shortfall. |
| 288 | */ |
| 289 | w->nspaces = 0; |
| 290 | w->shortfall = width; |
| 291 | for (j = 0; j < n; j++) { |
| 292 | w->shortfall -= wrapwords[i+j].width; |
| 293 | if (j < n-1 && wrapwords[i+j].spacewidth) { |
| 294 | w->nspaces++; |
| 295 | w->shortfall -= wrapwords[i+j].spacewidth; |
| 296 | } |
| 297 | } |
| 298 | i += n; |
| 299 | } |
| 300 | |
| 301 | sfree(wrapwords); |
| 302 | |
| 303 | return head; |
| 304 | } |
| 305 | |
| 306 | void wrap_free(wrappedline *w) { |
| 307 | while (w) { |
| 308 | wrappedline *t = w->next; |
| 309 | sfree(w); |
| 310 | w = t; |
| 311 | } |
| 312 | } |