*
* - linearised PDF, perhaps?
*
- * - we should use PDFDocEncoding or Unicode for outline strings,
- * now that I actually know how to do them. Probably easiest if
- * I do this _after_ bringing in libcharset, since I can simply
- * supply PDFDocEncoding in there.
- *
* - I'm uncertain of whether I need to include a ToUnicode CMap
* in each of my font definitions in PDF. Currently things (by
* which I mean cut and paste out of acroread) seem to be
*
* - title pages
*
- * - ability to import other Type 1 fonts
+ * - ability to use Type 1 fonts without AFM files
* * we need to parse the font to extract its metrics
- * * then we pass the font bodily to both PS and PDF so it can
- * be included in the output file
*
* - character substitution for better typography?
- * * fi, fl, ffi, ffl ligatures
* * use real ellipsis rather than ...
* * a hyphen in a word by itself might prefer to be an en-dash
* * (Americans might even want a convenient way to use an
#include <assert.h>
#include <stdio.h>
#include <stdarg.h>
+#include <stdlib.h>
#include "halibut.h"
#include "paper.h"
int pagenum_fontsize;
int footer_distance;
wchar_t *lquote, *rquote, *bullet;
+ wchar_t *contents_text, *index_text;
/* These are derived from the above */
int base_width;
int page_height;
word_PageXref = word_NotWordType + 1
};
+/* Flags for render_string() */
+#define RS_NOLIG 1
+
static font_data *make_std_font(font_list *fontlist, char const *name);
static void wrap_paragraph(para_data *pdata, word *words,
int w, int i1, int i2, paper_conf *conf);
static page_data *page_breaks(line_data *first, line_data *last,
int page_height, int ncols, int headspace);
static int render_string(page_data *page, font_data *font, int fontsize,
- int x, int y, wchar_t *str);
+ int x, int y, wchar_t *str, unsigned flags);
static int render_line(line_data *ldata, int left_x, int top_y,
xref_dest *dest, keywordlist *keywords, indexdata *idx,
paper_conf *conf);
static void render_para(para_data *pdata, paper_conf *conf,
keywordlist *keywords, indexdata *idx,
paragraph *index_placeholder, page_data *index_page);
-static int string_width(font_data *font, wchar_t const *string, int *errs);
+static int string_width(font_data *font, wchar_t const *string, int *errs,
+ unsigned flags);
static int paper_width_simple(para_data *pdata, word *text, paper_conf *conf);
static para_data *code_paragraph(int indent, word *words, paper_conf *conf);
static para_data *rule_paragraph(int indent, paper_conf *conf);
va_start(ap, string);
while ( (font = va_arg(ap, font_data *)) != NULL) {
int errs;
- (void) string_width(font, string, &errs);
+ (void) string_width(font, string, &errs, 0);
if (errs) {
ret = FALSE;
break;
ret.lquote = L"\x2018\0\x2019\0'\0'\0\0";
ret.rquote = uadv(ret.lquote);
ret.bullet = L"\x2022\0-\0\0";
+ ret.contents_text = L"Contents";
+ ret.index_text = L"Index";
/*
* Two-pass configuration so that we can pick up global config
ret.lquote = uadv(p->keyword);
ret.rquote = uadv(ret.lquote);
}
+ } else if (!ustricmp(p->keyword, L"contents")) {
+ ret.contents_text = uadv(p->keyword);
+ } else if (!ustricmp(p->keyword, L"index")) {
+ ret.index_text = uadv(p->keyword);
} else if (!ustricmp(p->keyword, L"paper-bullet")) {
ret.bullet = uadv(p->keyword);
} else if (!ustricmp(p->keyword, L"paper-page-width")) {
paper_cfg_fonts(ret.ftitle.fonts, fontlist, uadv(p->keyword),
&p->fpos);
} else if (!ustricmp(p->keyword, L"paper-chapter-font-size")) {
- ret.ftitle.font_size = utoi(uadv(p->keyword));
+ ret.fchapter.font_size = utoi(uadv(p->keyword));
} else if (!ustricmp(p->keyword, L"paper-chapter-fonts")) {
- paper_cfg_fonts(ret.ftitle.fonts, fontlist, uadv(p->keyword),
+ paper_cfg_fonts(ret.fchapter.fonts, fontlist, uadv(p->keyword),
&p->fpos);
} else if (!ustricmp(p->keyword, L"paper-section-font-size")) {
wchar_t *q = uadv(p->keyword);
* but must be in the title and body fonts. */
while (*uadv(ret.rquote) && *uadv(uadv(ret.rquote))) {
int n;
- if (!fonts_ok(ret.lquote,
- ret.fbase.fonts[FONT_NORMAL],
- ret.fbase.fonts[FONT_EMPH],
- ret.ftitle.fonts[FONT_NORMAL],
- ret.ftitle.fonts[FONT_EMPH],
- ret.fchapter.fonts[FONT_NORMAL],
- ret.fchapter.fonts[FONT_EMPH], NULL) ||
- !fonts_ok(ret.rquote,
- ret.fbase.fonts[FONT_NORMAL],
- ret.fbase.fonts[FONT_EMPH],
- ret.ftitle.fonts[FONT_NORMAL],
- ret.ftitle.fonts[FONT_EMPH],
- ret.fchapter.fonts[FONT_NORMAL],
- ret.fchapter.fonts[FONT_EMPH], NULL))
- break;
- for (n = 0; n < ret.nfsect; n++)
- if (!fonts_ok(ret.lquote,
- ret.fsect[n].fonts[FONT_NORMAL],
- ret.fsect[n].fonts[FONT_EMPH], NULL) ||
- !fonts_ok(ret.rquote,
- ret.fsect[n].fonts[FONT_NORMAL],
- ret.fsect[n].fonts[FONT_EMPH], NULL))
+ if (fonts_ok(ret.lquote,
+ ret.fbase.fonts[FONT_NORMAL],
+ ret.fbase.fonts[FONT_EMPH],
+ ret.ftitle.fonts[FONT_NORMAL],
+ ret.ftitle.fonts[FONT_EMPH],
+ ret.fchapter.fonts[FONT_NORMAL],
+ ret.fchapter.fonts[FONT_EMPH], NULL) &&
+ fonts_ok(ret.rquote,
+ ret.fbase.fonts[FONT_NORMAL],
+ ret.fbase.fonts[FONT_EMPH],
+ ret.ftitle.fonts[FONT_NORMAL],
+ ret.ftitle.fonts[FONT_EMPH],
+ ret.fchapter.fonts[FONT_NORMAL],
+ ret.fchapter.fonts[FONT_EMPH], NULL)) {
+ for (n = 0; n < ret.nfsect; n++)
+ if (!fonts_ok(ret.lquote,
+ ret.fsect[n].fonts[FONT_NORMAL],
+ ret.fsect[n].fonts[FONT_EMPH], NULL) ||
+ !fonts_ok(ret.rquote,
+ ret.fsect[n].fonts[FONT_NORMAL],
+ ret.fsect[n].fonts[FONT_EMPH], NULL))
+ break;
+ if (n == ret.nfsect)
break;
+ }
ret.lquote = uadv(ret.rquote);
ret.rquote = uadv(ret.lquote);
}
paragraph index_placeholder_para;
page_data *first_index_page;
+ init_std_fonts();
fontlist = snew(font_list);
fontlist->head = fontlist->tail = NULL;
*/
{
word *contents_title;
- contents_title = fake_word(L"Contents");
+ contents_title = fake_word(conf->contents_text);
firstcont = make_para_data(para_UnnumberedChapter, 0, 0, 0,
NULL, NULL, contents_title, conf);
if (has_index) {
pdata = make_para_data(para_Normal, 0, 0,
conf->contents_margin,
- NULL, NULL, fake_word(L"Index"), conf);
+ NULL, NULL,
+ fake_word(conf->index_text), conf);
pdata->next = NULL;
pdata->contents_entry = &index_placeholder_para;
lastcont->next = pdata;
/*
* Create a set of paragraphs for the index.
*/
- index_title = fake_word(L"Index");
+ index_title = fake_word(conf->index_text);
firstidx = make_para_data(para_UnnumberedChapter, 0, 0, 0,
NULL, NULL, index_title, conf);
width = conf->pagenum_fontsize *
string_width(conf->fbase.fonts[FONT_NORMAL], page->number,
- NULL);
+ NULL, 0);
render_string(page, conf->fbase.fonts[FONT_NORMAL],
conf->pagenum_fontsize,
conf->left_margin + (conf->base_width - width)/2,
conf->bottom_margin - conf->footer_distance,
- page->number);
+ page->number, 0);
}
}
pdata->rect_type = RECT_NONE;
pdata->contents_entry = NULL;
pdata->justification = JUST;
+ pdata->extraflags = 0;
/*
* Choose fonts for this paragraph.
fe->free_pos = 0x21;
for (i = 0; i < 256; i++) {
- fe->vector[i] = NULL;
- fe->indices[i] = -1;
+ fe->vector[i] = NOGLYPH;
fe->to_unicode[i] = 0xFFFF;
}
return fe;
}
-static int kern_cmp(void *a, void *b)
+static subfont_map_entry *encode_glyph_at(glyph g, wchar_t u,
+ font_encoding *fe, int pos)
+{
+ subfont_map_entry *sme = snew(subfont_map_entry);
+
+ sme->subfont = fe;
+ sme->position = pos;
+ fe->vector[pos] = g;
+ fe->to_unicode[pos] = u;
+ add234(fe->font->subfont_map, sme);
+ return sme;
+}
+
+static int new_sfmap_cmp(void *a, void *b)
+{
+ glyph ga = *(glyph *)a;
+ subfont_map_entry *sb = b;
+ glyph gb = sb->subfont->vector[sb->position];
+
+ if (ga < gb) return -1;
+ if (ga > gb) return 1;
+ return 0;
+}
+
+static subfont_map_entry *encode_glyph(glyph g, wchar_t u, font_data *font)
+{
+ subfont_map_entry *sme;
+ int c;
+
+ sme = find234(font->subfont_map, &g, new_sfmap_cmp);
+ if (sme) return sme;
+
+ /*
+ * This character is not yet in a subfont. Assign one.
+ */
+ if (font->latest_subfont->free_pos >= 0x100)
+ font->latest_subfont = new_font_encoding(font);
+
+ c = font->latest_subfont->free_pos++;
+ if (font->latest_subfont->free_pos == 0x7F)
+ font->latest_subfont->free_pos = 0xA1;
+
+ return encode_glyph_at(g, u, font->latest_subfont, c);
+}
+
+static int sfmap_cmp(void *a, void *b)
+{
+ subfont_map_entry *sa = a, *sb = b;
+ glyph ga = sa->subfont->vector[sa->position];
+ glyph gb = sb->subfont->vector[sb->position];
+
+ if (ga < gb) return -1;
+ if (ga > gb) return 1;
+ return 0;
+}
+
+int width_cmp(void *a, void *b)
+{
+ glyph_width const *wa = a, *wb = b;
+
+ if (wa->glyph < wb->glyph)
+ return -1;
+ if (wa->glyph > wb->glyph)
+ return 1;
+ return 0;
+}
+
+int kern_cmp(void *a, void *b)
{
kern_pair const *ka = a, *kb = b;
return 0;
}
+int lig_cmp(void *a, void *b)
+{
+ ligature const *la = a, *lb = b;
+
+ if (la->left < lb->left)
+ return -1;
+ if (la->left > lb->left)
+ return 1;
+ if (la->right < lb->right)
+ return -1;
+ if (la->right > lb->right)
+ return 1;
+ return 0;
+}
+
+static int utoglyph(font_info const *fi, wchar_t u) {
+ return (u < 0 || u > 0xFFFF ? NOGLYPH : fi->bmp[u]);
+}
+
+void listfonts(void) {
+ font_info const *fi;
+
+ init_std_fonts();
+ for (fi = all_fonts; fi; fi = fi->next)
+ printf("%s\n", fi->name);
+}
+
static font_data *make_std_font(font_list *fontlist, char const *name)
{
- const int *widths;
- const kern_pair *kerns;
- int nglyphs;
+ font_info const *fi;
font_data *f;
font_encoding *fe;
int i;
for (fe = fontlist->head; fe; fe = fe->next)
- if (strcmp(fe->font->name, name) == 0)
+ if (strcmp(fe->font->info->name, name) == 0)
return fe->font;
- /* XXXKERN */
- widths = ps_std_font_widths(name);
- kerns = ps_std_font_kerns(name);
- if (!widths || !kerns)
- return NULL;
-
- for (nglyphs = 0; ps_std_glyphs[nglyphs] != NULL; nglyphs++);
+ for (fi = all_fonts; fi; fi = fi->next)
+ if (strcmp(fi->name, name) == 0) break;
+ if (!fi) return NULL;
f = snew(font_data);
f->list = fontlist;
- f->name = name;
- f->nglyphs = nglyphs;
- f->glyphs = ps_std_glyphs;
- f->widths = widths;
- f->kerns = newtree234(kern_cmp);
- for (;kerns->left != 0xFFFF; kerns++)
- add234(f->kerns, (void *)kerns);
- f->subfont_map = snewn(nglyphs, subfont_map_entry);
+ f->info = fi;
+ f->subfont_map = newtree234(sfmap_cmp);
/*
* Our first subfont will contain all of US-ASCII. This isn't
fe->free_pos = 0xA1; /* only the top half is free */
f->latest_subfont = fe;
- for (i = 0; i < (int)lenof(f->bmp); i++)
- f->bmp[i] = 0xFFFF;
-
- for (i = 0; i < nglyphs; i++) {
- wchar_t ucs;
- ucs = ps_glyph_to_unicode(f->glyphs[i]);
- assert(ucs != 0xFFFF);
- f->bmp[ucs] = i;
- if (ucs >= 0x20 && ucs <= 0x7E) {
- fe->vector[ucs] = f->glyphs[i];
- fe->indices[ucs] = i;
- fe->to_unicode[ucs] = ucs;
- f->subfont_map[i].subfont = fe;
- f->subfont_map[i].position = ucs;
- } else {
- /*
- * This character is not yet assigned to a subfont.
- */
- f->subfont_map[i].subfont = NULL;
- f->subfont_map[i].position = 0;
- }
+ for (i = 0x20; i <= 0x7E; i++) {
+ glyph g = utoglyph(fi, i);
+ if (g != NOGLYPH)
+ encode_glyph_at(g, i, fe, i);
}
return f;
}
/* NB: arguments are glyph numbers from font->bmp. */
+int find_width(font_data *font, glyph index)
+{
+ glyph_width wantw;
+ glyph_width const *w;
+
+ wantw.glyph = index;
+ w = find234(font->info->widths, &wantw, NULL);
+ if (!w) return 0;
+ return w->width;
+}
+
static int find_kern(font_data *font, int lindex, int rindex)
{
kern_pair wantkp;
kern_pair const *kp;
- if (lindex == 0xFFFF || rindex == 0xFFFF)
+ if (lindex == NOGLYPH || rindex == NOGLYPH)
return 0;
wantkp.left = lindex;
wantkp.right = rindex;
- kp = find234(font->kerns, &wantkp, NULL);
+ kp = find234(font->info->kerns, &wantkp, NULL);
if (kp == NULL)
return 0;
return kp->kern;
}
-static int string_width(font_data *font, wchar_t const *string, int *errs)
+static int find_lig(font_data *font, int lindex, int rindex)
+{
+ ligature wantlig;
+ ligature const *lig;
+
+ if (lindex == NOGLYPH || rindex == NOGLYPH)
+ return NOGLYPH;
+ wantlig.left = lindex;
+ wantlig.right = rindex;
+ lig = find234(font->info->ligs, &wantlig, NULL);
+ if (lig == NULL)
+ return NOGLYPH;
+ return lig->lig;
+}
+
+static int string_width(font_data *font, wchar_t const *string, int *errs,
+ unsigned flags)
{
int width = 0;
- int index, oindex;
+ int nindex, index, oindex, lindex;
if (errs)
*errs = 0;
- oindex = 0xFFFF;
+ oindex = NOGLYPH;
+ index = utoglyph(font->info, *string);
for (; *string; string++) {
- index = (*string < 0 || *string > 0xFFFF ? 0xFFFF :
- font->bmp[*string]);
+ nindex = utoglyph(font->info, string[1]);
- if (index == 0xFFFF) {
+ if (index == NOGLYPH) {
if (errs)
*errs = 1;
} else {
- width += find_kern(font, oindex, index) + font->widths[index];
+ if (!(flags & RS_NOLIG) &&
+ (lindex = find_lig(font, index, nindex)) != NOGLYPH) {
+ index = lindex;
+ continue;
+ }
+ width += find_kern(font, oindex, index) + find_width(font, index);
}
oindex = index;
+ index = nindex;
}
return width;
struct paper_width_ctx *ctx = (struct paper_width_ctx *)vctx;
int style, type, findex, width, errs;
wchar_t *str;
+ unsigned flags = 0;
switch (word->type) {
case word_HyperLink:
style == word_Emph ? FONT_EMPH :
FONT_CODE);
+ if (style == word_Code || style == word_WeakCode) flags |= RS_NOLIG;
+
if (type == word_Normal) {
str = word->text;
} else if (type == word_WhiteSpace) {
str = ctx->conf->rquote;
}
- width = string_width(ctx->pdata->fonts[findex], str, &errs);
+ width = string_width(ctx->pdata->fonts[findex], str, &errs, flags);
if (errs && word->alt)
return paper_width_list(vctx, word->alt, NULL, nspaces);
ctx.pdata = pdata;
ctx.minspacewidth =
(pdata->sizes[FONT_NORMAL] *
- string_width(pdata->fonts[FONT_NORMAL], L" ", NULL));
+ string_width(pdata->fonts[FONT_NORMAL], L" ", NULL, 0));
ctx.conf = conf;
return paper_width_list(&ctx, text, NULL, NULL);
}
spacewidth = (pdata->sizes[FONT_NORMAL] *
- string_width(pdata->fonts[FONT_NORMAL], L" ", NULL));
+ string_width(pdata->fonts[FONT_NORMAL], L" ", NULL, 0));
if (spacewidth == 0) {
/*
* A font without a space?! Disturbing. I hope this never
* Returns the updated x coordinate.
*/
static int render_string(page_data *page, font_data *font, int fontsize,
- int x, int y, wchar_t *str)
+ int x, int y, wchar_t *str, unsigned flags)
{
char *text;
- int textpos, textwid, kern, glyph, oglyph;
+ int textpos, textwid, kern, nglyph, glyph, oglyph, lig;
font_encoding *subfont = NULL, *sf;
+ subfont_map_entry *sme;
text = snewn(1 + ustrlen(str), char);
textpos = textwid = 0;
- glyph = 0xFFFF;
+ glyph = NOGLYPH;
+ nglyph = utoglyph(font->info, *str);
while (*str) {
oglyph = glyph;
- glyph = (*str < 0 || *str > 0xFFFF ? 0xFFFF :
- font->bmp[*str]);
+ glyph = nglyph;
+ nglyph = utoglyph(font->info, str[1]);
- if (glyph == 0xFFFF) {
+ if (glyph == NOGLYPH) {
str++;
continue; /* nothing more we can do here */
}
+ if (!(flags & RS_NOLIG) &&
+ (lig = find_lig(font, glyph, nglyph)) != NOGLYPH) {
+ nglyph = lig;
+ str++;
+ continue;
+ }
+
/*
* Find which subfont this character is going in.
*/
- sf = font->subfont_map[glyph].subfont;
-
- if (!sf) {
- int c;
-
- /*
- * This character is not yet in a subfont. Assign one.
- */
- if (font->latest_subfont->free_pos >= 0x100)
- font->latest_subfont = new_font_encoding(font);
-
- c = font->latest_subfont->free_pos++;
- if (font->latest_subfont->free_pos == 0x7F)
- font->latest_subfont->free_pos = 0xA1;
-
- font->subfont_map[glyph].subfont = font->latest_subfont;
- font->subfont_map[glyph].position = c;
- font->latest_subfont->vector[c] = font->glyphs[glyph];
- font->latest_subfont->indices[c] = glyph;
- font->latest_subfont->to_unicode[c] = *str;
-
- sf = font->latest_subfont;
- }
+ sme = encode_glyph(glyph, *str, font);
+ sf = sme->subfont;
kern = find_kern(font, oglyph, glyph) * fontsize;
subfont = sf;
}
- text[textpos++] = font->subfont_map[glyph].position;
- textwid += font->widths[glyph] * fontsize;
+ text[textpos++] = sme->position;
+ textwid += find_width(font, glyph) * fontsize;
str++;
}
int style, type, findex, errs;
wchar_t *str;
xref_dest dest;
+ unsigned flags = 0;
switch (text->type) {
/*
style == word_Emph ? FONT_EMPH :
FONT_CODE);
+ if (style == word_Code || style == word_WeakCode) flags |= RS_NOLIG;
+ flags |= pdata->extraflags;
+
if (type == word_Normal) {
str = text->text;
} else if (type == word_WhiteSpace) {
x += pdata->sizes[findex] *
- string_width(pdata->fonts[findex], L" ", NULL);
+ string_width(pdata->fonts[findex], L" ", NULL, 0);
if (nspaces && findex != FONT_CODE) {
x += (*nspace+1) * shortfall / nspaces;
x -= *nspace * shortfall / nspaces;
str = conf->rquote;
}
- (void) string_width(pdata->fonts[findex], str, &errs);
+ (void) string_width(pdata->fonts[findex], str, &errs, flags);
if (errs && text->alt)
x = render_text(page, pdata, ldata, x, y, text->alt, NULL,
conf);
else
x = render_string(page, pdata->fonts[findex],
- pdata->sizes[findex], x, y, str);
+ pdata->sizes[findex], x, y, str, flags);
if (*xr)
(*xr)->rx = x;
pdata->sizes[FONT_NORMAL],
conf->left_margin + x,
(conf->paper_height - conf->top_margin -
- pdata->last->ypos), L".");
+ pdata->last->ypos), L".", 0);
render_string(pdata->last->page,
pdata->fonts[FONT_NORMAL],
pdata->sizes[FONT_NORMAL],
conf->paper_width - conf->right_margin - wid,
(conf->paper_height - conf->top_margin -
- pdata->last->ypos), num);
+ pdata->last->ypos), num, 0);
}
/*
pdata->rect_type = RECT_NONE;
pdata->contents_entry = NULL;
pdata->justification = LEFT;
+ pdata->extraflags = RS_NOLIG;
for (; words; words = words->next) {
wchar_t *t, *e, *start;
pdata->rect_type = RECT_RULE;
pdata->contents_entry = NULL;
pdata->justification = LEFT;
+ pdata->extraflags = 0;
standard_line_spacing(pdata, conf);