gtkdoc/fixxref.py

   1 # -*- python -*-
   2 #
   3 # gtk-doc - GTK DocBook documentation generator.
   4 # Copyright (C) 1998  Damon Chaplin
   5 #               2007-2016  Stefan Sauer
   6 #
   7 # This program is free software; you can redistribute it and/or modify
   8 # it under the terms of the GNU General Public License as published by
   9 # the Free Software Foundation; either version 2 of the License, or
  10 # (at your option) any later version.
  11 #
  12 # This program is distributed in the hope that it will be useful,
  13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 # GNU General Public License for more details.
  16 #
  17 # You should have received a copy of the GNU General Public License
  18 # along with this program; if not, write to the Free Software
  19 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20 #
  21
  22 ''"Fix cross-references in the HTML documentation.''"
  23
  24 import logging
  25 import os
  26 import re
  27 import shlex
  28 import subprocess
  29 import sys
  30 import tempfile
  31
  32 from . import common, config
  33
  34 # This contains all the entities and their relative URLs.
  35 Links = {}
  36
  37 # failing link targets we don't warn about even once
  38 NoLinks = {
  39     'char',
  40     'double',
  41     'float',
  42     'int',
  43     'long',
  44     'main',
  45     'signed',
  46     'unsigned',
  47     'va-list',
  48     'void',
  49 }
  50
  51
  52 def Run(options):
  53     logging.info('options: %s', str(options.__dict__))
  54
  55     LoadIndicies(options.module_dir, options.html_dir, options.extra_dir)
  56     ReadSections(options.module)
  57     FixCrossReferences(options.module_dir, options.module, options.src_lang)
  58
  59
  60 # TODO(ensonic): try to refactor so that we get a list of path's and then just
  61 # loop over them.
  62 # - module_dir is by default 'html'
  63 # - html_dir can be set by configure, defaults to $(docdir)
  64 def LoadIndicies(module_dir, html_dir, extra_dirs):
  65     # Cache of dirs we already scanned for index files
  66     dir_cache = {}
  67
  68     path_prefix = ''
  69     m = re.search(r'(.*?)/share/gtk-doc/html', html_dir)
  70     if m:
  71         path_prefix = m.group(1)
  72         logging.info('Path prefix: %s', path_prefix)
  73     prefix_match = r'^' + re.escape(path_prefix) + r'/'
  74
  75     # We scan the directory containing GLib and any directories in GNOME2_PATH
  76     # first, but these will be overriden by any later scans.
  77     dir = common.GetModuleDocDir('glib-2.0')
  78     if dir and os.path.exists(dir):
  79         # Some predefined link targets to get links into type hierarchies as these
  80         # have no targets. These are always absolute for now.
  81         Links['GBoxed'] = dir + '/gobject/gobject-Boxed-Types.html'
  82         Links['GEnum'] = dir + '/gobject/gobject-Enumeration-and-Flag-Types.html'
  83         Links['GFlags'] = dir + '/gobject/gobject-Enumeration-and-Flag-Types.html'
  84         Links['GInterface'] = dir + '/gobject/GTypeModule.html'
  85
  86         if dir != html_dir:
  87             logging.info('Scanning GLib directory: %s', dir)
  88             ScanIndices(dir, (re.search(prefix_match, dir) is None), dir_cache)
  89     else:
  90         NoLinks.add('GBoxed')
  91         NoLinks.add('GEnum')
  92         NoLinks.add('GFlags')
  93         NoLinks.add('GInterface')
  94
  95     path = os.environ.get('GNOME2_PATH')
  96     if path:
  97         for dir in path.split(':'):
  98             dir += 'share/gtk-doc/html'
  99             if os.path.exists(dir) and dir != html_dir:
 100                 logging.info('Scanning GNOME2_PATH directory: %s', dir)
 101                 ScanIndices(dir, (re.search(prefix_match, dir) is None), dir_cache)
 102
 103     logging.info('Scanning HTML_DIR directory: %s', html_dir)
 104     ScanIndices(html_dir, False, dir_cache)
 105     logging.info('Scanning MODULE_DIR directory: %s', module_dir)
 106     ScanIndices(module_dir, False, dir_cache)
 107
 108     # check all extra dirs, but skip already scanned dirs or subdirs of those
 109     for dir in extra_dirs:
 110         dir = dir.rstrip('/')
 111         logging.info('Scanning EXTRA_DIR directory: %s', dir)
 112
 113         # If the --extra-dir option is not relative and is not sharing the same
 114         # prefix as the target directory of the docs, we need to use absolute
 115         # directories for the links
 116         if not dir.startswith('..') and re.search(prefix_match, dir) is None:
 117             ScanIndices(dir, True, dir_cache)
 118         else:
 119             ScanIndices(dir, False, dir_cache)
 120
 121
 122 def ScanIndices(scan_dir, use_absolute_links, dir_cache):
 123     if not scan_dir or scan_dir in dir_cache:
 124         return
 125     dir_cache[scan_dir] = 1
 126
 127     logging.info('Scanning index directory: %s, absolute: %d', scan_dir, use_absolute_links)
 128
 129     # TODO(ensonic): this code is the same as in rebase.py
 130     if not os.path.isdir(scan_dir):
 131         logging.info('Cannot open dir "%s"', scan_dir)
 132         return
 133
 134     subdirs = []
 135     for entry in sorted(os.listdir(scan_dir)):
 136         full_entry = os.path.join(scan_dir, entry)
 137         if os.path.isdir(full_entry):
 138             subdirs.append(full_entry)
 139             continue
 140
 141         if entry.endswith('.devhelp2'):
 142             # if devhelp-file is good don't read index.sgml
 143             ReadDevhelp(full_entry, use_absolute_links)
 144         elif entry == "index.sgml.gz" and not os.path.exists(os.path.join(scan_dir, 'index.sgml')):
 145             # debian/ubuntu started to compress this as index.sgml.gz :/
 146             print(''' Please fix https://bugs.launchpad.net/ubuntu/+source/gtk-doc/+bug/77138 . For now run:
 147 gunzip %s
 148 ''' % full_entry)
 149         elif entry.endswith('.devhelp2.gz') and not os.path.exists(full_entry[:-3]):
 150             # debian/ubuntu started to compress this as *devhelp2.gz :/
 151             print('''Please fix https://bugs.launchpad.net/ubuntu/+source/gtk-doc/+bug/1466210 . For now run:
 152 gunzip %s
 153 ''' % full_entry)
 154         # we could consider supporting: gzip module
 155
 156     # Now recursively scan the subdirectories.
 157     for subdir in subdirs:
 158         ScanIndices(subdir, use_absolute_links, dir_cache)
 159
 160
 161 def ReadDevhelp(file, use_absolute_links):
 162     # Determine the absolute directory, to be added to links in $file
 163     # if we need to use an absolute link.
 164     # $file will be something like /prefix/gnome/share/gtk-doc/html/gtk/$file
 165     # We want the part up to 'html/.*' since the links in $file include
 166     # the rest.
 167     dir = "../"
 168     if use_absolute_links:
 169         # For uninstalled index files we'd need to map the path to where it
 170         # will be installed to
 171         if not file.startswith('./'):
 172             m = re.search(r'(.*\/)(.*?)\/.*?\.devhelp2', file)
 173             dir = m.group(1) + m.group(2) + '/'
 174     else:
 175         m = re.search(r'(.*\/)(.*?)\/.*?\.devhelp2', file)
 176         if m:
 177             dir += m.group(2) + '/'
 178         else:
 179             dir = ''
 180
 181     logging.info('Scanning index file=%s, absolute=%d, dir=%s', file, use_absolute_links, dir)
 182
 183     for line in open(file, 'r', encoding='utf-8'):
 184         m = re.search(r' link="([^#]*)#([^"]*)"', line)
 185         if m:
 186             link = m.group(1) + '#' + m.group(2)
 187             logging.debug('Found id: %s href: %s', m.group(2), link)
 188             Links[m.group(2)] = dir + link
 189
 190
 191 def ReadSections(module):
 192     """We don't warn on missing links to non-public sysmbols."""
 193     for line in open(module + '-sections.txt', 'r', encoding='utf-8'):
 194         m1 = re.search(r'^<SUBSECTION\s*(.*)>', line)
 195         if line.startswith('#') or line.strip() == '':
 196             continue
 197         elif line.startswith('<SECTION>'):
 198             subsection = ''
 199         elif m1:
 200             subsection = m1.group(1)
 201         elif line.startswith('<SUBSECTION>') or line.startswith('</SECTION>'):
 202             continue
 203         elif re.search(r'^<TITLE>(.*)<\/TITLE>', line):
 204             continue
 205         elif re.search(r'^<FILE>(.*)<\/FILE>', line):
 206             continue
 207         elif re.search(r'^<INCLUDE>(.*)<\/INCLUDE>', line):
 208             continue
 209         else:
 210             symbol = line.strip()
 211             if subsection == "Standard" or subsection == "Private":
 212                 NoLinks.add(common.CreateValidSGMLID(symbol))
 213
 214
 215 def FixCrossReferences(module_dir, module, src_lang):
 216     # TODO(ensonic): use glob.glob()?
 217     for entry in sorted(os.listdir(module_dir)):
 218         full_entry = os.path.join(module_dir, entry)
 219         if os.path.isdir(full_entry):
 220             continue
 221         elif entry.endswith('.html') or entry.endswith('.htm'):
 222             FixHTMLFile(src_lang, module, full_entry)
 223
 224
 225 def FixHTMLFile(src_lang, module, file):
 226     logging.info('Fixing file: %s', file)
 227
 228     content = open(file, 'r', encoding='utf-8').read()
 229
 230     if config.highlight:
 231         # FIXME: ideally we'd pass a clue about the example language to the highligher
 232         # unfortunately the "language" attribute is not appearing in the html output
 233         # we could patch the customization to have <code class="xxx"> inside of <pre>
 234         if config.highlight.endswith('vim'):
 235             def repl_func(m):
 236                 return HighlightSourceVim(src_lang, m.group(1), m.group(2))
 237             content = re.sub(
 238                 r'<div class=\"(example-contents|informalexample)\"><pre class=\"programlisting\">(.*?)</pre></div>',
 239                 repl_func, content, flags=re.DOTALL)
 240         else:
 241             def repl_func(m):
 242                 return HighlightSource(src_lang, m.group(1), m.group(2))
 243             content = re.sub(
 244                 r'<div class=\"(example-contents|informalexample)\"><pre class=\"programlisting\">(.*?)</pre></div>',
 245                 repl_func, content, flags=re.DOTALL)
 246
 247         content = re.sub(r'\&lt;GTKDOCLINK\s+HREF=\&quot;(.*?)\&quot;\&gt;(.*?)\&lt;/GTKDOCLINK\&gt;',
 248                          r'\<GTKDOCLINK\ HREF=\"\1\"\>\2\</GTKDOCLINK\>', content, flags=re.DOTALL)
 249
 250         # From the highlighter we get all the functions marked up. Now we can turn them into GTKDOCLINK items
 251         def repl_func(m):
 252             return MakeGtkDocLink(m.group(1), m.group(2), m.group(3))
 253         content = re.sub(r'(<span class=\"function\">)(.*?)(</span>)', repl_func, content, flags=re.DOTALL)
 254         # We can also try the first item in stuff marked up as 'normal'
 255         content = re.sub(
 256             r'(<span class=\"normal\">\s*)(.+?)((\s+.+?)?\s*</span>)', repl_func, content, flags=re.DOTALL)
 257
 258     lines = content.rstrip().split('\n')
 259
 260     def repl_func_with_ix(i):
 261         def repl_func(m):
 262             return MakeXRef(module, file, i + 1, m.group(1), m.group(2))
 263         return repl_func
 264
 265     for i in range(len(lines)):
 266         lines[i] = re.sub(r'<GTKDOCLINK\s+HREF="([^"]*)"\s*>(.*?)</GTKDOCLINK\s*>', repl_func_with_ix(i), lines[i])
 267         if 'GTKDOCLINK' in lines[i]:
 268             logging.info('make xref failed for line %d: "%s"', i, lines[i])
 269
 270     new_file = file + '.new'
 271     content = '\n'.join(lines)
 272     with open(new_file, 'w', encoding='utf-8') as h:
 273         h.write(content)
 274
 275     os.unlink(file)
 276     os.rename(new_file, file)
 277
 278
 279 def GetXRef(id):
 280     href = Links.get(id)
 281     if href:
 282         return (id, href)
 283
 284     # This is a workaround for some inconsistency we have with CreateValidSGMLID
 285     if ':' in id:
 286         tid = id.replace(':', '--')
 287         href = Links.get(tid)
 288         if href:
 289             return (tid, href)
 290
 291     # poor mans plural support
 292     if id.endswith('s'):
 293         tid = id[:-1]
 294         href = Links.get(tid)
 295         if href:
 296             return (tid, href)
 297         tid += '-struct'
 298         href = Links.get(tid)
 299         if href:
 300             return (tid, href)
 301
 302     tid = id + '-struct'
 303     href = Links.get(tid)
 304     if href:
 305         return (tid, href)
 306
 307     return (id, None)
 308
 309
 310 def ReportBadXRef(file, line, id, text):
 311     logging.info('no link for: id=%s, linktext=%s', id, text)
 312
 313     # don't warn multiple times and also skip blacklisted (ctypes)
 314     if id in NoLinks:
 315         return
 316     # if it's a function, don't warn if it does not contain a "_"
 317     # (transformed to "-")
 318     # - gnome coding style would use '_'
 319     # - will avoid wrong warnings for ansi c functions
 320     if re.search(r' class=\"function\"', text) and '-' not in id:
 321         return
 322     # if it's a 'return value', don't warn (implicitly created link)
 323     if re.search(r' class=\"returnvalue\"', text):
 324         return
 325     # if it's a 'type', don't warn if it starts with lowercase
 326     # - gnome coding style would use CamelCase
 327     if re.search(r' class=\"type\"', text) and id[0].islower():
 328         return
 329     # don't warn for self links
 330     if text == id:
 331         return
 332
 333     common.LogWarning(file, line, 'no link for: "%s" -> (%s).' % (id, text))
 334     NoLinks.add(id)
 335
 336
 337 def MakeRelativeXRef(module, href):
 338     # if it is a link to same module, remove path to make it work uninstalled
 339     m = re.search(r'^\.\./' + module + '/(.*)$', href)
 340     if m:
 341         href = m.group(1)
 342     return href
 343
 344
 345 def MakeXRef(module, file, line, id, text):
 346     href = GetXRef(id)[1]
 347
 348     if href:
 349         href = MakeRelativeXRef(module, href)
 350         logging.info('Fixing link: %s, %s, %s', id, href, text)
 351         return "<a href=\"%s\">%s</a>" % (href, text)
 352     else:
 353         ReportBadXRef(file, line, id, text)
 354         return text
 355
 356
 357 def MakeGtkDocLink(pre, symbol, post):
 358     id = common.CreateValidSGMLID(symbol)
 359
 360     # these are implicitely created links in highlighed sources
 361     # we don't want warnings for those if the links cannot be resolved.
 362     NoLinks.add(id)
 363
 364     return pre + '<GTKDOCLINK HREF="' + id + '">' + symbol + '</GTKDOCLINK>' + post
 365
 366
 367 def HighlightSource(src_lang, type, source):
 368     # write source to a temp file
 369     # FIXME: use .c for now to hint the language to the highlighter
 370     with tempfile.NamedTemporaryFile(mode='w+', suffix='.c') as f:
 371         temp_source_file = HighlightSourcePreProcess(f, source)
 372         highlight_options = config.highlight_options.replace('$SRC_LANG', src_lang)
 373
 374         logging.info('running %s %s %s', config.highlight, highlight_options, temp_source_file)
 375
 376         # format source
 377         highlighted_source = subprocess.check_output(
 378             [config.highlight] + shlex.split(highlight_options) + [temp_source_file]).decode('utf-8')
 379         logging.debug('result: [%s]', highlighted_source)
 380         if config.highlight.endswith('/source-highlight'):
 381             highlighted_source = re.sub(r'^<\!-- .*? -->', '', highlighted_source, flags=re.MULTILINE | re.DOTALL)
 382             highlighted_source = re.sub(
 383                 r'<pre><tt>(.*?)</tt></pre>', r'\1', highlighted_source, flags=re.MULTILINE | re.DOTALL)
 384         elif config.highlight.endswith('/highlight'):
 385             # need to rewrite the stylesheet classes
 386             highlighted_source = highlighted_source.replace('<span class="gtkdoc com">', '<span class="comment">')
 387             highlighted_source = highlighted_source.replace('<span class="gtkdoc dir">', '<span class="preproc">')
 388             highlighted_source = highlighted_source.replace('<span class="gtkdoc kwd">', '<span class="function">')
 389             highlighted_source = highlighted_source.replace('<span class="gtkdoc kwa">', '<span class="keyword">')
 390             highlighted_source = highlighted_source.replace('<span class="gtkdoc line">', '<span class="linenum">')
 391             highlighted_source = highlighted_source.replace('<span class="gtkdoc num">', '<span class="number">')
 392             highlighted_source = highlighted_source.replace('<span class="gtkdoc str">', '<span class="string">')
 393             highlighted_source = highlighted_source.replace('<span class="gtkdoc sym">', '<span class="symbol">')
 394             # maybe also do
 395             # highlighted_source = re.sub(r'</span>(.+)<span', '</span><span class="normal">\1</span><span')
 396
 397     return HighlightSourcePostprocess(type, highlighted_source)
 398
 399
 400 def HighlightSourceVim(src_lang, type, source):
 401     # write source to a temp file
 402     with tempfile.NamedTemporaryFile(mode='w+', suffix='.h') as f:
 403         temp_source_file = HighlightSourcePreProcess(f, source)
 404
 405         # format source
 406         # TODO(ensonic): use p.communicate()
 407         script = "echo 'let html_number_lines=0|let html_use_css=1|let html_use_xhtml=1|e %s|syn on|set syntax=%s|run! plugin/tohtml.vim|run! syntax/2html.vim|w! %s.html|qa!' | " % (
 408             temp_source_file, src_lang, temp_source_file)
 409         script += "%s -n -e -u NONE -T xterm >/dev/null" % config.highlight
 410         subprocess.check_call([script], shell=True)
 411
 412         highlighted_source = open(temp_source_file + ".html", 'r', encoding='utf-8').read()
 413         highlighted_source = re.sub(r'.*<pre\b[^>]*>\n', '', highlighted_source, flags=re.DOTALL)
 414         highlighted_source = re.sub(r'</pre>.*', '', highlighted_source, flags=re.DOTALL)
 415
 416         # need to rewrite the stylesheet classes
 417         highlighted_source = highlighted_source.replace('<span class="Comment">', '<span class="comment">')
 418         highlighted_source = highlighted_source.replace('<span class="PreProc">', '<span class="preproc">')
 419         highlighted_source = highlighted_source.replace('<span class="Statement">', '<span class="keyword">')
 420         highlighted_source = highlighted_source.replace('<span class="Identifier">', '<span class="function">')
 421         highlighted_source = highlighted_source.replace('<span class="Constant">', '<span class="number">')
 422         highlighted_source = highlighted_source.replace('<span class="Special">', '<span class="symbol">')
 423         highlighted_source = highlighted_source.replace('<span class="Type">', '<span class="type">')
 424
 425         # remove temp files
 426         os.unlink(temp_source_file + '.html')
 427
 428     return HighlightSourcePostprocess(type, highlighted_source)
 429
 430
 431 def HighlightSourcePreProcess(f, source):
 432     # chop of leading and trailing empty lines, leave leading space in first real line
 433     source = source.strip(' ')
 434     source = source.strip('\n')
 435     source = source.rstrip()
 436
 437     # cut common indent
 438     m = re.search(r'^(\s+)', source)
 439     if m:
 440         source = re.sub(r'^' + m.group(1), '', source, flags=re.MULTILINE)
 441     # avoid double entity replacement
 442     source = source.replace('&lt;', '<')
 443     source = source.replace('&gt;', '>')
 444     source = source.replace('&amp;', '&')
 445     if sys.version_info < (3,):
 446         source = source.encode('utf-8')
 447     f.write(source)
 448     f.flush()
 449     return f.name
 450
 451
 452 def HighlightSourcePostprocess(type, highlighted_source):
 453     # chop of leading and trailing empty lines
 454     highlighted_source = highlighted_source.strip()
 455
 456     # turn common urls in comments into links
 457     highlighted_source = re.sub(r'<span class="url">(.*?)</span>',
 458                                 r'<span class="url"><a href="\1">\1</a></span>',
 459                                 highlighted_source, flags=re.DOTALL)
 460
 461     # we do own line-numbering
 462     line_count = highlighted_source.count('\n')
 463     source_lines = '\n'.join([str(i) for i in range(1, line_count + 2)])
 464
 465     return """<div class="%s">
 466   <table class="listing_frame" border="0" cellpadding="0" cellspacing="0">
 467     <tbody>
 468       <tr>
 469         <td class="listing_lines" align="right"><pre>%s</pre></td>
 470         <td class="listing_code"><pre class="programlisting">%s</pre></td>
 471       </tr>
 472     </tbody>
 473   </table>
 474 </div>
 475 """ % (type, source_lines, highlighted_source)