20
21
#include "system.h"
24
26
#include "makeinfo.h"
25
28
#include "sectioning.h"
31
/* Append CHAR to BUFFER, (re)allocating as necessary. We don't handle
36
unsigned size; /* allocated */
37
unsigned length; /* used */
44
buffer_type *buf = xmalloc (sizeof (buffer_type));
53
append_char (buffer_type *buf, int c)
56
if (buf->length >= buf->size)
59
buf->buffer = xrealloc (buf->buffer, buf->size);
61
buf->buffer[buf->length - 1] = c;
62
buf->buffer[buf->length] = 0;
65
/* Read the cascading style-sheet file FILENAME. Write out any @import
66
commands, which must come first, by the definition of css. If the
67
file contains any actual css code following the @imports, return it;
70
process_css_file (char *filename)
75
buffer_type *import_text = init_buffer ();
76
buffer_type *inline_text = init_buffer ();
78
enum { null_state, comment_state, import_state, inline_state } state
79
= null_state, prev_state;
81
prev_state = null_state;
83
/* read from stdin if `-' is the filename. */
84
f = STREQ (filename, "-") ? stdin : fopen (filename, "r");
87
error (_("%s: could not open --css-file: %s"), progname, filename);
91
/* Read the file. The @import statements must come at the beginning,
92
with only whitespace and comments allowed before any inline css code. */
93
while ((c = getc (f)) >= 0)
100
case null_state: /* between things */
102
{ /* Only @import and @charset should switch into
103
import_state, other @-commands, such as @media, should
104
put us into inline_state. I don't think any other css
105
@-commands start with `i' or `c', although of course
106
this will break when such a command is defined. */
107
int nextchar = getc (f);
108
if (nextchar == 'i' || nextchar == 'c')
110
append_char (import_text, c);
111
state = import_state;
115
ungetc (nextchar, f); /* wasn't an @import */
116
state = inline_state;
120
{ /* possible start of a comment */
121
int nextchar = getc (f);
123
state = comment_state;
126
ungetc (nextchar, f); /* wasn't a comment */
127
state = inline_state;
130
else if (isspace (c))
131
; /* skip whitespace; maybe should use c_isspace? */
134
/* not an @import, not a comment, not whitespace: we must
135
have started the inline text. */
136
state = inline_state;
138
if (state == inline_state)
139
append_char (inline_text, c);
141
if (state != null_state)
142
prev_state = null_state;
146
if (c == '/' && lastchar == '*')
147
state = prev_state; /* end of comment */
148
break; /* else ignore this comment char */
151
append_char (import_text, c); /* include this import char */
153
{ /* done with @import */
154
append_char (import_text, '\n'); /* make the output nice */
156
prev_state = import_state;
161
/* No harm in writing out comments, so don't bother parsing
162
them out, just append everything. */
163
append_char (inline_text, c);
170
/* Reached the end of the file. We should not be still in a comment. */
171
if (state == comment_state)
172
warning (_("%s:%d: --css-file ended in comment"), filename, lineno);
174
/* Write the @import text, if any. */
175
if (import_text->buffer)
177
add_word (import_text->buffer);
178
free (import_text->buffer);
182
/* We're wasting the buffer struct memory, but so what. */
183
return inline_text->buffer;
186
HSTACK *htmlstack = NULL;
28
189
int html_output_head_p = 0;
190
int html_title_written = 0;
193
html_output_head (void)
33
static char *html_title = NULL;
34
static int html_title_written = 0;
195
static const char *html_title = NULL;
36
198
if (html_output_head_p)
38
200
html_output_head_p = 1;
202
encoding = current_document_encoding ();
40
204
/* The <title> should not have markup, so use text_expansion. */
42
html_title = title ? text_expansion (title) : _("Untitled");
44
add_word_args ("<html lang=\"%s\">\n<head>\n<title>%s</title>\n",
45
language_table[language_code].abbrev, html_title);
206
html_title = escape_string (title ?
207
text_expansion (title) : (char *) _("Untitled"));
209
/* Make sure this is the very first string of the output document. */
210
output_paragraph_offset = 0;
212
add_html_block_elt_args ("<html lang=\"%s\">\n<head>\n",
213
language_table[language_code].abbrev);
215
/* When splitting, add current node's name to title if it's available and not
217
if (splitting && current_node && !STREQ (current_node, "Top"))
218
add_word_args ("<title>%s - %s</title>\n",
219
escape_string (xstrdup (current_node)), html_title);
221
add_word_args ("<title>%s</title>\n", html_title);
47
223
add_word ("<meta http-equiv=\"Content-Type\" content=\"text/html");
48
if (document_encoding_code != no_encoding)
49
add_word_args ("; charset=%s",
50
encoding_table[document_encoding_code].ecname);
224
if (encoding && *encoding)
225
add_word_args ("; charset=%s", encoding);
51
227
add_word ("\">\n");
53
229
if (!document_description)
54
230
document_description = html_title;
56
add_word_args ("<meta name=description content=\"%s\">\n",
232
add_word_args ("<meta name=\"description\" content=\"%s\">\n",
57
233
document_description);
58
add_word_args ("<meta name=generator content=\"makeinfo %s\">\n", VERSION);
59
add_word ("<link href=\"http://texinfo.org/\" rel=generator-home>\n");
234
add_word_args ("<meta name=\"generator\" content=\"makeinfo %s\">\n",
237
/* Navigation bar links. */
239
add_word ("<link title=\"Top\" rel=\"top\" href=\"#Top\">\n");
242
/* Always put a top link. */
243
add_word ("<link title=\"Top\" rel=\"start\" href=\"index.html#Top\">\n");
245
/* We already have a top link, avoid duplication. */
246
if (tag_table->up && !STREQ (tag_table->up, "Top"))
247
add_link (tag_table->up, "rel=\"up\"");
250
add_link (tag_table->prev, "rel=\"prev\"");
253
add_link (tag_table->next, "rel=\"next\"");
255
/* fixxme: Look for a way to put links to various indices in the
256
document. Also possible candidates to be added here are First and
261
/* We are splitting, but we neither have a tag_table. So this must be
262
index.html. So put a link to Top. */
263
add_word ("<link title=\"Top\" rel=\"start\" href=\"#Top\">\n");
266
add_word ("<link href=\"http://www.gnu.org/software/texinfo/\" \
267
rel=\"generator-home\" title=\"Texinfo Homepage\">\n");
270
{ /* It is not ideal that we include the html markup here within
271
<head>, so we use text_expansion. */
272
insert_string ("<!--\n");
273
insert_string (text_expansion (copying_text));
274
insert_string ("-->\n");
277
/* Put the style definitions in a comment for the sake of browsers
278
that don't support <style>. */
279
add_word ("<meta http-equiv=\"Content-Style-Type\" content=\"text/css\">\n");
280
add_word ("<style type=\"text/css\"><!--\n");
283
char *css_inline = NULL;
286
/* This writes out any @import commands from the --css-file,
287
and returns any actual css code following the imports. */
288
css_inline = process_css_file (css_include);
290
/* This seems cleaner than adding <br>'s at the end of each line for
291
these "roman" displays. It's hardly the end of the world if the
292
browser doesn't do <style>s, in any case; they'll just come out in
294
#define CSS_FONT_INHERIT "font-family:inherit"
295
add_word_args (" pre.display { %s }\n", CSS_FONT_INHERIT);
296
add_word_args (" pre.format { %s }\n", CSS_FONT_INHERIT);
298
/* Alternatively, we could do <font size=-1> in insertion.c, but this
299
way makes it easier to override. */
300
#define CSS_FONT_SMALLER "font-size:smaller"
301
add_word_args (" pre.smalldisplay { %s; %s }\n", CSS_FONT_INHERIT,
303
add_word_args (" pre.smallformat { %s; %s }\n", CSS_FONT_INHERIT,
305
add_word_args (" pre.smallexample { %s }\n", CSS_FONT_SMALLER);
306
add_word_args (" pre.smalllisp { %s }\n", CSS_FONT_SMALLER);
308
/* Since HTML doesn't have a sc element, we use span with a bit of
309
CSS spice instead. */
310
#define CSS_FONT_SMALL_CAPS "font-variant:small-caps"
311
add_word_args (" span.sc { %s }\n", CSS_FONT_SMALL_CAPS);
313
/* Roman (default) font class, closest we can come. */
314
#define CSS_FONT_ROMAN "font-family:serif; font-weight:normal;"
315
add_word_args (" span.roman { %s } \n", CSS_FONT_ROMAN);
317
/* Sans serif font class. */
318
#define CSS_FONT_SANSSERIF "font-family:sans-serif; font-weight:normal;"
319
add_word_args (" span.sansserif { %s } \n", CSS_FONT_SANSSERIF);
321
/* Write out any css code from the user's --css-file. */
323
insert_string (css_inline);
325
add_word ("--></style>\n");
60
328
add_word ("</head>\n<body>\n");
62
if (title && !html_title_written)
330
if (title && !html_title_written && titlepage_cmd_present)
64
add_word_args ("<h1>%s</h1>\n", html_title);
332
add_word_args ("<h1 class=\"settitle\">%s</h1>\n", html_title);
65
333
html_title_written = 1;
70
339
/* Escape HTML special characters in the string if necessary,
71
340
returning a pointer to a possibly newly-allocated one. */
73
escape_string (string)
342
escape_string (char *string)
345
int i = 0, newlen = 0;
81
349
/* Find how much to allocate. */
82
350
switch (string[i])
353
newlen += 6; /* `"' */
85
356
newlen += 5; /* `&' */
124
399
return newstring - newlen;
402
/* Save current tag. */
404
push_tag (char *tag, char *attribs)
406
HSTACK *newstack = xmalloc (sizeof (HSTACK));
409
newstack->attribs = xstrdup (attribs);
410
newstack->next = htmlstack;
411
htmlstack = newstack;
418
HSTACK *tos = htmlstack;
422
line_error (_("[unexpected] no html tag to pop"));
426
free (htmlstack->attribs);
428
htmlstack = htmlstack->next;
432
/* Check if tag is an empty or a whitespace only element.
433
If so, remove it, keeping whitespace intact. */
435
rollback_empty_tag (char *tag)
437
int check_position = output_paragraph_offset;
438
int taglen = strlen (tag);
439
int rollback_happened = 0;
441
char *contents_canon_white = "";
443
/* If output_paragraph is empty, we cannot rollback :-\ */
444
if (output_paragraph_offset <= 0)
447
/* Find the end of the previous tag. */
448
while (output_paragraph[check_position-1] != '>' && check_position > 0)
451
/* Save stuff between tag's end to output_paragraph's end. */
452
if (check_position != output_paragraph_offset)
454
contents = xmalloc (output_paragraph_offset - check_position + 1);
455
memcpy (contents, output_paragraph + check_position,
456
output_paragraph_offset - check_position);
458
contents[output_paragraph_offset - check_position] = '\0';
460
contents_canon_white = xstrdup (contents);
461
canon_white (contents_canon_white);
464
/* Find the start of the previous tag. */
465
while (output_paragraph[check_position-1] != '<' && check_position > 0)
468
/* Check to see if this is the tag. */
469
if (strncmp ((char *) output_paragraph + check_position, tag, taglen) == 0
470
&& (whitespace (output_paragraph[check_position + taglen])
471
|| output_paragraph[check_position + taglen] == '>'))
473
if (!contents_canon_white || !*contents_canon_white)
475
/* Empty content after whitespace removal, so roll it back. */
476
output_paragraph_offset = check_position - 1;
477
rollback_happened = 1;
479
/* Original contents may not be empty (whitespace.) */
480
if (contents && *contents)
482
insert_string (contents);
488
return rollback_happened;
127
491
/* Open or close TAG according to START_OR_END. */
129
insert_html_tag (start_or_end, tag)
493
#if defined (VA_FPRINTF) && __STDC__
494
insert_html_tag_with_attribute (int start_or_end, char *tag, char *format, ...)
496
insert_html_tag_with_attribute (start_or_end, tag, format, va_alist)
130
497
int start_or_end;
133
if (!paragraph_is_open && (start_or_end == START))
135
/* Need to compensate for the <p> we are about to insert, or
136
else cm_xxx functions that call us will get wrong text
137
between START and END. */
138
adjust_braces_following (output_paragraph_offset, 3);
503
char *old_tag = NULL;
504
char *old_attribs = NULL;
505
char formatted_attribs[2000]; /* xx no fixed limits */
507
extern int in_html_elt;
142
509
if (start_or_end != START)
514
old_tag = htmlstack->tag;
515
old_attribs = htmlstack->attribs;
524
VA_START (ap, format);
526
VA_SPRINTF (formatted_attribs, format, ap);
528
sprintf (formatted_attribs, format, a1, a2, a3, a4, a5, a6, a7, a8);
533
formatted_attribs[0] = '\0';
535
/* Exception: can nest multiple spans. */
537
&& STREQ (htmlstack->tag, tag)
538
&& !(STREQ (tag, "span") && STREQ (old_attribs, formatted_attribs)))
541
if (start_or_end == START)
542
push_tag (tag, formatted_attribs);
549
/* texinfo.tex doesn't support more than one font attribute
551
if ((start_or_end == START) && old_tag && *old_tag
552
&& !rollback_empty_tag (old_tag))
553
add_word_args ("</%s>", old_tag);
557
if (start_or_end == START)
558
add_word_args (format ? "<%s %s>" : "<%s>", tag, formatted_attribs);
559
else if (!rollback_empty_tag (tag))
560
/* Insert close tag only if we didn't rollback,
561
in which case the opening tag is removed. */
562
add_word_args ("</%s>", tag);
565
if ((start_or_end != START) && old_tag && *old_tag)
566
add_word_args (strlen (old_attribs) > 0 ? "<%s %s>" : "<%s>",
567
old_tag, old_attribs);
573
insert_html_tag (int start_or_end, char *tag)
575
insert_html_tag_with_attribute (start_or_end, tag, NULL);
148
578
/* Output an HTML <link> to the filename for NODE, including the
149
579
other string as extra attributes. */
151
add_link (nodename, attributes)
152
char *nodename, *attributes;
581
add_link (char *nodename, char *attributes)
157
586
add_word_args ("%s", attributes);
158
587
add_word_args (" href=\"");
159
588
add_anchor_name (nodename, 1);
160
add_word ("\"></a>\n");
589
add_word_args ("\" title=\"%s\">\n", nodename);
164
593
/* Output NAME with characters escaped as appropriate for an anchor
165
name, i.e., escape URL special characters as %<n>. */
594
name, i.e., escape URL special characters with our _00hh convention
595
if OLD is zero. (See the manual for details on the new scheme.)
597
If OLD is nonzero, generate the node name with the 4.6-and-earlier
598
convention of %hh (and more special characters output as-is, notably
599
- and *). This is only so that external references to old names can
600
still work with HTML generated by the new makeinfo; the gcc folks
601
needed this. Our own HTML does not refer to these names. */
167
add_escaped_anchor_name (name)
604
add_escaped_anchor_name (char *name, int old)
608
if (!old && !strchr ("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ",
610
{ /* XHTML does not allow anything but an ASCII letter to start an
611
identifier. Therefore kludge in this constant string if we
170
616
for (; *name; name++)
174
else if (! URL_SAFE_CHAR (*name))
618
if (cr_or_whitespace (*name))
621
else if (!old && !URL_SAFE_CHAR (*name))
175
622
/* Cast so characters with the high bit set are treated as >128,
176
623
for example o-umlaut should be 246, not -10. */
624
add_word_args ("_00%x", (unsigned char) *name);
626
else if (old && !URL_SAFE_CHAR (*name) && !OLD_URL_SAFE_CHAR (*name))
627
/* Different output convention, but still cast as above. */
177
628
add_word_args ("%%%x", (unsigned char) *name);
179
631
add_char (*name);
183
635
/* Insert the text for the name of a reference in an HTML anchor
184
appropriate for NODENAME. If HREF is nonzero, it will be
185
appropriate for a href= attribute, rather than name= i.e., including
186
the `#' if it's an internal reference. */
636
appropriate for NODENAME.
638
If HREF is zero, generate text for name= in the new node name
639
conversion convention.
640
If HREF is negative, generate text for name= in the old convention.
641
If HREF is positive, generate the name for an href= attribute, i.e.,
642
including the `#' if it's an internal reference. */
188
add_anchor_name (nodename, href)
644
add_anchor_name (char *nodename, int href)
195
649
add_url_name (nodename, href);
202
656
if (strcasecmp (nodename, "(dir)") == 0)
203
657
/* Strip the parens, but keep the original letter-case. */
204
658
add_word_args ("%.3s", nodename + 1);
659
else if (strcasecmp (nodename, "top") == 0)
206
add_escaped_anchor_name (nodename);
662
add_escaped_anchor_name (nodename, href < 0);
209
665
/* Insert the text for the name of a reference in an HTML url, aprropriate
212
add_url_name (nodename, href)
668
add_url_name (char *nodename, int href)
216
670
add_nodename_to_filename (nodename, href);
219
/* Only allow [-0-9a-zA-Z_.] when nodifying filenames. This may
220
result in filename clashes; e.g.,
225
both map to Foo--.html. If that happens, cm_node will put all
226
the nodes whose file names clash on the same file. */
228
fix_filename (filename)
673
/* Convert non [A-Za-z0-9] to _00xx, where xx means the hexadecimal
674
representation of the ASCII character. Also convert spaces and
675
newlines to dashes. */
677
fix_filename (char *filename)
232
for (p = filename; *p; p++)
680
int len = strlen (filename);
681
char *oldname = xstrdup (filename);
685
for (i = 0; i < len; i++)
234
if (!(isalnum (*p) || strchr ("-._", *p)))
687
if (cr_or_whitespace (oldname[i]))
688
strcat (filename, "-");
689
else if (URL_SAFE_CHAR (oldname[i]))
690
strncat (filename, (char *) oldname + i, 1);
693
char *hexchar = xmalloc (6 * sizeof (char));
694
sprintf (hexchar, "_00%x", (unsigned char) oldname[i]);
695
strcat (filename, hexchar);
699
/* Check if we are nearing boundaries. */
700
if (strlen (filename) >= PATH_MAX - 20)
239
707
/* As we can't look-up a (forward-referenced) nodes' html filename
added
removed
makeinfo/html.c
