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