/[gentoo]/xml/htdocs/proj/en/glep/glep-0025.html
Gentoo

Contents of /xml/htdocs/proj/en/glep/glep-0025.html

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.2 - (hide annotations) (download) (as text)
Thu Nov 11 21:34:45 2004 UTC (9 years, 10 months ago) by g2boojum
Branch: MAIN
Changes since 1.1: +49 -44 lines
File MIME type: text/html
update

1 g2boojum 1.1 <?xml version="1.0" encoding="utf-8" ?>
2     <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
3     <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
4     <!--
5     This HTML is auto-generated. DO NOT EDIT THIS FILE! If you are writing a new
6     PEP, see http://www.python.org/peps/pep-0001.html for instructions and links
7     to templates. DO NOT USE THIS HTML FILE AS YOUR TEMPLATE!
8     -->
9     <head>
10     <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
11 g2boojum 1.2 <meta name="generator" content="Docutils 0.3.3: http://docutils.sourceforge.net/" />
12 g2boojum 1.1 <title>GLEP 25 -- Distfile Patching Support</title>
13     <link rel="stylesheet" href="tools/glep.css" type="text/css" />
14     </head>
15     <body bgcolor="white">
16     <table class="navigation" cellpadding="0" cellspacing="0"
17     width="100%" border="0">
18     <tr><td class="navicon" width="150" height="35">
19     <a href="http://www.gentoo.org/" title="Gentoo Linux Home Page">
20     <img src="http://www.gentoo.org/images/gentoo-new.gif" alt="[Gentoo]"
21     border="0" width="150" height="35" /></a></td>
22     <td class="textlinks" align="left">
23     [<b><a href="http://www.gentoo.org/">Gentoo Linux Home</a></b>]
24     [<b><a href="http://www.gentoo.org/proj/en/glep">GLEP Index</a></b>]
25 g2boojum 1.2 [<b><a href="./glep-0025.txt">GLEP Source</a></b>]
26 g2boojum 1.1 </td></tr></table>
27     <div class="document">
28     <table class="rfc2822 field-list" frame="void" rules="none">
29     <col class="field-name" />
30     <col class="field-body" />
31     <tbody valign="top">
32     <tr class="field"><th class="field-name">GLEP:</th><td class="field-body">25</td>
33     </tr>
34     <tr class="field"><th class="field-name">Title:</th><td class="field-body">Distfile Patching Support</td>
35     </tr>
36 g2boojum 1.2 <tr class="field"><th class="field-name">Version:</th><td class="field-body">1.2</td>
37 g2boojum 1.1 </tr>
38 g2boojum 1.2 <tr class="field"><th class="field-name">Last-Modified:</th><td class="field-body"><a class="reference" href="http://www.gentoo.org/cgi-bin/viewcvs/xml/htdocs/proj/en/glep/glep-0025.txt?cvsroot=gentoo">2004/11/11 21:34:36</a></td>
39 g2boojum 1.1 </tr>
40     <tr class="field"><th class="field-name">Author:</th><td class="field-body">Brian Harring &lt;ferringb&#32;&#97;t&#32;gentoo.org&gt;</td>
41     </tr>
42 g2boojum 1.2 <tr class="field"><th class="field-name">Status:</th><td class="field-body">deferred</td>
43 g2boojum 1.1 </tr>
44     <tr class="field"><th class="field-name">Type:</th><td class="field-body">Standards Track</td>
45     </tr>
46 g2boojum 1.2 <tr class="field"><th class="field-name">Content-Type:</th><td class="field-body"><a class="reference" href="glep-0012.html">text/x-rst</a></td>
47 g2boojum 1.1 </tr>
48     <tr class="field"><th class="field-name">Created:</th><td class="field-body">6-Mar-2004</td>
49     </tr>
50 g2boojum 1.2 <tr class="field"><th class="field-name">Post-History:</th><td class="field-body">4-Apr-2004, 11-Nov-2004</td>
51 g2boojum 1.1 </tr>
52     </tbody>
53     </table>
54     <hr />
55     <div class="contents topic" id="contents">
56 g2boojum 1.2 <p class="topic-title first"><a name="contents">Contents</a></p>
57 g2boojum 1.1 <ul class="simple">
58     <li><a class="reference" href="#abstract" id="id7" name="id7">Abstract</a></li>
59 g2boojum 1.2 <li><a class="reference" href="#status" id="id8" name="id8">Status</a></li>
60     <li><a class="reference" href="#motivation" id="id9" name="id9">Motivation</a></li>
61     <li><a class="reference" href="#binary-patches-vs-gnudiff-patches" id="id10" name="id10">Binary patches vs GNUDiff patches</a></li>
62     <li><a class="reference" href="#rationale" id="id11" name="id11">Rationale</a></li>
63     <li><a class="reference" href="#specification" id="id12" name="id12">Specification</a><ul>
64     <li><a class="reference" href="#additions-to-the-tree" id="id13" name="id13">Additions to the tree</a></li>
65     <li><a class="reference" href="#portage-implementation" id="id14" name="id14">Portage Implementation</a><ul>
66     <li><a class="reference" href="#fetching" id="id15" name="id15">Fetching</a></li>
67     <li><a class="reference" href="#reconstruction" id="id16" name="id16">Reconstruction</a></li>
68     <li><a class="reference" href="#compressed-md5sums" id="id17" name="id17">Compressed MD5sums:</a><ul>
69     <li><a class="reference" href="#the-problem-in-detail" id="id18" name="id18">The Problem in Detail</a></li>
70     <li><a class="reference" href="#the-proposed-solution" id="id19" name="id19">The Proposed Solution</a></li>
71 g2boojum 1.1 </ul>
72     </li>
73     </ul>
74     </li>
75 g2boojum 1.2 <li><a class="reference" href="#distfile-mirror-additions" id="id20" name="id20">Distfile Mirror Additions</a></li>
76     <li><a class="reference" href="#patch-creation" id="id21" name="id21">Patch Creation</a></li>
77 g2boojum 1.1 </ul>
78     </li>
79 g2boojum 1.2 <li><a class="reference" href="#backwards-compatability" id="id22" name="id22">Backwards Compatability</a></li>
80     <li><a class="reference" href="#reference-implementation" id="id23" name="id23">Reference Implementation</a></li>
81     <li><a class="reference" href="#references" id="id24" name="id24">References</a></li>
82     <li><a class="reference" href="#copyright" id="id25" name="id25">Copyright</a></li>
83 g2boojum 1.1 </ul>
84     </div>
85     <div class="section" id="abstract">
86     <h1><a class="toc-backref" href="#id7" name="abstract">Abstract</a></h1>
87     <p>The intention of this GLEP is to propose the creation of patching support for
88     portage, and iron out the implementation details.</p>
89     </div>
90 g2boojum 1.2 <div class="section" id="status">
91     <h1><a class="toc-backref" href="#id8" name="status">Status</a></h1>
92     <p>Timed out</p>
93     </div>
94 g2boojum 1.1 <div class="section" id="motivation">
95 g2boojum 1.2 <h1><a class="toc-backref" href="#id9" name="motivation">Motivation</a></h1>
96 g2boojum 1.1 <p>Reduce the bandwidth load placed on our mirrors by decreasing the amount of
97     bytes transferred when upgrading between versions. Side benefit of this is to
98     significantly decrease the download requirements for users lacking broadband.</p>
99     </div>
100     <div class="section" id="binary-patches-vs-gnudiff-patches">
101 g2boojum 1.2 <h1><a class="toc-backref" href="#id10" name="binary-patches-vs-gnudiff-patches">Binary patches vs GNUDiff patches</a></h1>
102 g2boojum 1.1 <p>Most people are familiar with diff patches (unified diff for example)- this
103     glep is specifically proposing the use of an actual binary differencer. The
104     reason for this is that diff patches are line based- you change a single
105     character in a line, and the whole line must be included in the patch. Binary
106     differencers work at the byte level- it encodes just that byte. In that
107     respect binary patches are often much more efficient then diff patches.</p>
108     <p>Further, the ability to reverse a unified patch is due to the fact the diff
109     includes <strong>both</strong> the original line, and the modified line. The author isn't
110     aware of any binary differencer that is able to create patches the can be
111     reversed- basically they're unidirectional, the patch that is generated can
112     only be used to upgrade or downgrade the version, not both. The plus side of
113     this limitation is a significantly decreased patch size.</p>
114     <p>The choice of binary patches over diff patches pretty much comes down to the
115     fact they're smaller- example being a kdelibs binary patch for 3.1.4-&gt;3.1.5 is
116     75kb, the equivalent diff patch is 123kb, and is unable to result in a correct
117     md5 <a class="footnote-reference" href="#id4" id="id1" name="id1">[1]</a>.</p>
118     <p>Currently, this glep is proposing only the usage of binary patches- that's not
119     to say (with a fair amount of work) it couldn't be extended to support
120     standard diffs.</p>
121     </div>
122     <div class="section" id="rationale">
123 g2boojum 1.2 <h1><a class="toc-backref" href="#id11" name="rationale">Rationale</a></h1>
124 g2boojum 1.1 <p>The difference between source releases typically isn't very large, especially
125     for minor releases. As an example, kdelibs-3.1.4.tar.bz2 is 10.53 MB, and
126     kdelibs-3.1.5.tar.bz2 is 10.54 MB. A bzip2'ed patch between those versions is
127     75.6 kb <a class="footnote-reference" href="#id5" id="id2" name="id2">[2]</a>, less then 1% the size of 3.1.5's tbz2.</p>
128     </div>
129     <div class="section" id="specification">
130 g2boojum 1.2 <h1><a class="toc-backref" href="#id12" name="specification">Specification</a></h1>
131 g2boojum 1.1 <p>Quite a few sections of gentoo are affected- mirroring, the portage tree, and
132     portage itself.</p>
133     <div class="section" id="additions-to-the-tree">
134 g2boojum 1.2 <h2><a class="toc-backref" href="#id13" name="additions-to-the-tree">Additions to the tree</a></h2>
135 g2boojum 1.1 <p>For adding patch info into the tree, this glep proposes a global patch list
136     (stored in profiles as patches.global), and individual patch lists stored in
137     relevant package directories (named patches). Using the kernel packages as an
138     example, a global list of patches enables us to create a patch once, add an
139     entry, and have all kernel packages benefit from that single entry. Both
140     patches.global, and individual package patch files share the same format:</p>
141     <pre class="literal-block">
142     MD5 md5-value patch-url size MD5 md5-value ref-file size UMD5 md5-value new-file size
143     </pre>
144     <p>For those familiar with digest file layout, this should look familiar.
145     Essentially, chksum type, value, filename, size. The UMD5 chksum type is just
146     the uncompressed md5/size of the file- so if the UMD5 were for a bzip2
147     compressed file, it would be the md5 value/size of the uncompressed file.
148     And an example:</p>
149     <pre class="literal-block">
150     MD5 ccd5411b3558326cbce0306fcae32e26 http://dev.gentoo.org/~ferringb/patches/kdelibs-3.1.4-3.1.5.patch.bz2 75687 MD5 82c265de78d53c7060a09c5cb1a78942 kdelibs-3.1.4.tar.bz2 10537433 UMD5 0b1908a51e739c07ff5a88e189d2f7a9 kdelibs-3.1.5.tar.bz2 48056320
151     </pre>
152     <p>In the above example, the md5sum of
153     <a class="reference" href="http://dev.gentoo.org/~ferringb/patches/kdelibs-3.1.4-3.1.5.patch.bz2">http://dev.gentoo.org/~ferringb/patches/kdelibs-3.1.4-3.1.5.patch.bz2</a> is
154     calculated, compared to the stored value, and then the file size is checked.
155     The one difference is the UMD5 checksum type- the md5 value and the size are
156     specific to the <em>uncompressed</em> file. Continuing, for cases where the patch
157     will reside on one of our mirrors, the patch filename would be sufficient.</p>
158     <p>Finally, note that this is a unidirectional patch- using the above example,
159     kdelibs-3.1.4-3.1.5 can <strong>only</strong> be used to upgrade from 3.1.4 to 3.1.5, not
160     in reverse (originally explained in <a class="reference" href="#binary-patches-vs-gnudiff-patches">Binary patches vs GNUDiff patches</a>).</p>
161     </div>
162     <div class="section" id="portage-implementation">
163 g2boojum 1.2 <h2><a class="toc-backref" href="#id14" name="portage-implementation">Portage Implementation</a></h2>
164 g2boojum 1.1 <p>This glep proposes the patching support should be (at this stage) optional-
165     specifically, enabled via FEATURES=&quot;patching&quot;.</p>
166     <div class="section" id="fetching">
167 g2boojum 1.2 <h3><a class="toc-backref" href="#id15" name="fetching">Fetching</a></h3>
168 g2boojum 1.1 <p>When patching is enabled, the global patch list is read, and the packages
169     patch list is read. From there, portage determines what files could be used
170     as a base for patching to the desired file- further, determining if it's
171     actually worth patching (case where it wouldn't be is when the target file is
172     less then the sum of the patches needed). Any patches to be used are fetched,
173     and md5 verified.</p>
174     </div>
175     <div class="section" id="reconstruction">
176 g2boojum 1.2 <h3><a class="toc-backref" href="#id16" name="reconstruction">Reconstruction</a></h3>
177 g2boojum 1.1 <p>Upon fetching and md5 verification of patch(es), the desired file is
178     reconstructed. Assuming reconstruction didn't return any errors, the target
179     file has its uncompressed md5sum calculated and verified, then is recompressed
180     and the compressed md5sum calculated. At this point, if the compressed md5
181     matches the md5 stored in the tree, then portage transfers the file into
182     distfiles, and continues on it's merry way.</p>
183     <p>If the compressed md5 is different from the tree's value, then the (proposed)
184     md5 database is updated with new compressed md5. Details of this database
185     (and the issue it addresses) follow.</p>
186     </div>
187     <div class="section" id="compressed-md5sums">
188 g2boojum 1.2 <h3><a class="toc-backref" href="#id17" name="compressed-md5sums">Compressed MD5sums:</a></h3>
189 g2boojum 1.1 <p>There will be instances where a file is reconstructed perfectly, recompressed,
190     and the recompressed md5sum differs from what is stored in the tree- the
191     problem is that the md5sum of a compressed file is inherently tied to the
192     compressor version/options used to compress the original source.</p>
193     <div class="section" id="the-problem-in-detail">
194 g2boojum 1.2 <h4><a class="toc-backref" href="#id18" name="the-problem-in-detail">The Problem in Detail</a></h4>
195 g2boojum 1.1 <p>A good example of this problem is related to bzip2 versions used for
196     compression. Between bzip2 0.9x and bzip2 1.x, there was a subtle change in
197     the compressor resulting in a slightly better compression result- end result
198     being a different file, eg a different md5sum. Assuming compressor versions
199     are the same, there also is the issue of what compression level the target
200     source was originally compressed at- was it compressed with -9, -8 or -7?
201     That's just a sampling of the various original settings that must be accounted
202     for, and that's limited to gzip/bzip2; other compressors will add to the
203     number of variables to be accounted for to produce an exact recreation of the
204     compressed md5sum.</p>
205     <p>Tracking the compressor version and options originally used isn't really a
206     valid option- assuming all options were accounted for, clients would still be
207     required to have multiple versions of the same compressor installed just for
208     the sake of recreating a compressed md5sum <em>even though</em> the uncompressed
209     source's md5 has already been verified.</p>
210     </div>
211     <div class="section" id="the-proposed-solution">
212 g2boojum 1.2 <h4><a class="toc-backref" href="#id19" name="the-proposed-solution">The Proposed Solution</a></h4>
213 g2boojum 1.1 <p>The creation of a clientside flatfile/db of valid alternate md5/size pairs
214     would enable portage to handle perfectly reconstructed files, that have a
215     different md5sum due to compression differences. The proposed format is thus:</p>
216     <pre class="literal-block">
217     MD5 md5sum orig-file size MD5 md5sum [ optional new-name ] size
218     </pre>
219     <p>Example:</p>
220     <pre class="literal-block">
221     MD5 984146931906a7d53300b29f58f6a899 OOo_1.0.3_source.tar.bz2 165475319 MD5 0733dd85ed44d88d1eabed704d579721 165444187
222     </pre>
223     <p>An alternate md5/size pair for a file would be added <strong>only</strong> when the
224     uncompressed source's md5/size has been verified, yet upon recompression the
225     md5 differs. For cleansing of older md5/size pairs from this db, a utility
226     would be required- the author suggests the addition of a distfiles-cleaning
227     utility to portage, with the ability to also cleanse old md5/size pairs when
228     the file the pair was created for no longer exists in distfiles.</p>
229     <p>Where to store the database is debatable- /etc/portage or /var/cache/edb are
230     definite options.</p>
231     <p>The reasoning for allowing for an optional new-name is that it provides needed
232     functionality should anyone attempt to extend portage to allow for clients to
233     change the compression used for a source (eg, recompress all gzip files as
234     bzip2). Granted, no such code or attempt has been made, but nothing is lost
235     by leaving the option open should the request/attempt be made.</p>
236     <p>A potential gotcha of adding this support is that in environments where the
237     distfiles directory is shared out to multiple systems, this db must be shared
238     also.</p>
239     </div>
240     </div>
241     </div>
242     <div class="section" id="distfile-mirror-additions">
243 g2boojum 1.2 <h2><a class="toc-backref" href="#id20" name="distfile-mirror-additions">Distfile Mirror Additions</a></h2>
244 g2boojum 1.1 <p>One issue of contention is where these files will actually be stored. As of
245     the writing of this glep, a full distfiles mirror is roughly around 40 gb- a
246     rough estimate by the author places the space requirements for patches for
247     each version at a total of around 4gb. Note this isn't even remotely a hard
248     figure yet, and a better figure is being checked into currently.</p>
249     <p>Regardless of the exact space figure, finding a place to store the patches
250     will be problematic. Expansion of the required mirror space (essentially just
251     swallowing the patches storage requirement) is unlikely, since it was one of
252     the main arguements against the now defunct glep9 attempt <a class="footnote-reference" href="#id5" id="id3" name="id3">[2]</a>. A couple of
253     ideas that have been put forth to handle the additional space requirements are
254     as follows-</p>
255     <p>1) Identification of mirrors willing to handle the extra space requirements-
256     essentially create an additional patch mirror tier.</p>
257     <p>2) Mirroring only a patch for certain package versions, rather then full
258     source. Using kdelibs-3.1.5 as an example, only the patch would be mirrored
259     (rather then the full 10.53 MB source). Downside to this approach is that a
260     user who is downloading kdelibs for the first time would either need to pull
261     it from the original SRC_URI (placing the burden onto the upstream mirror), or
262     pull the 3.1.4 version, and the patch- pulling 63k more then if they had just
263     pulled the full version. The kdelibs 3.1.4/3.1.5 example is something of an
264     optimal case- not all versions will have such miniscule patches.</p>
265     <p>3) A variation on the idea above, essentially mirroring only the patch for
266     the oldest version(s) of a package; eg, kdelibs currently has version 3.05,
267     3.1.5, 3.2.0, and 3.2.1- the mirrors would only carry a patch for 3.05, not
268     full source (think RESTRICT=&quot;fetch&quot;). One plus to this is that patches to
269     downgrade in version are smaller then the patches to upgrade in version- there
270     are exceptions to this, but they're hard to find. A major downside to this
271     approach is A) a user would have to sync up to get the patchlists for that
272     version, B) creation of a set of patches to go backwards in version (see
273     <a class="reference" href="#binary-patches-vs-gnudiff-patches">Binary patches vs GNUDiff patches</a>)..</p>
274     <p>Of the options listed above, the first is the easiest, although the second
275     could be made to work. Feedback and any possible alternatives would be
276     greatly appreciated.</p>
277     </div>
278     <div class="section" id="patch-creation">
279 g2boojum 1.2 <h2><a class="toc-backref" href="#id21" name="patch-creation">Patch Creation</a></h2>
280 g2boojum 1.1 <p>Maintenance of patch lists, and the actual patch creation ought to be managed
281     by a high level script- essentally a dev says &quot;I want a patch between this
282     version, and that version: make it so&quot;, the script churns away
283     creating/updating the patch list, and generating the patch locally. The
284     utility next uploads the new patch to /space/distfiles-local on dev.gentoo.org
285     (exempting if it's not a locally generated patch), and repoman is used to
286     commit the updated patch list.</p>
287     <p>What would be preferable (although possibly wishful thinking), is if hardware
288     could be co-opted for automatic patch generation, rather then forcing it upon
289     the devs- something akin to how files are pulled onto the mirror automatically
290     for new ebuilds.</p>
291     <p>The initial bulk of patches to get will be generated by the author, to ease
292     the transition and offer patches for people to test out.</p>
293     </div>
294     </div>
295     <div class="section" id="backwards-compatability">
296 g2boojum 1.2 <h1><a class="toc-backref" href="#id22" name="backwards-compatability">Backwards Compatability</a></h1>
297 g2boojum 1.1 <p>As noted in <a class="reference" href="#the-proposed-solution">The Proposed Solution</a>, a system using patching and sharing out
298     it's distfiles must share out it's alternate md5 db. Any system that uses the
299     distfiles share must support the alternate md5 db also. If this is considered
300     enough of an issue, it is conceivable to place reconstructed sources with an
301     alternate md5 into a subdirectory of distdir- portage only looks within
302     distdir, unwilling to descend into subdirectories.</p>
303     <p>Also note that <a class="reference" href="#distfile-mirror-additions">Distfile Mirror Additions</a> may add additional backwards
304     compatability issues, depending on what solution is accepted.</p>
305     </div>
306     <div class="section" id="reference-implementation">
307 g2boojum 1.2 <h1><a class="toc-backref" href="#id23" name="reference-implementation">Reference Implementation</a></h1>
308 g2boojum 1.1 <p>TODO</p>
309     </div>
310     <div class="section" id="references">
311 g2boojum 1.2 <h1><a class="toc-backref" href="#id24" name="references">References</a></h1>
312 g2boojum 1.1 <table class="footnote" frame="void" id="id4" rules="none">
313     <colgroup><col class="label" /><col /></colgroup>
314     <tbody valign="top">
315     <tr><td class="label"><a class="fn-backref" href="#id1" name="id4">[1]</a></td><td><a class="reference" href="http://dev.gentoo.org/~ferringb/patches/kdelibs-3.1.4-3.1.5">http://dev.gentoo.org/~ferringb/patches/kdelibs-3.1.4-3.1.5</a>.{patch,diff}.bz2.</td></tr>
316     </tbody>
317     </table>
318     <table class="footnote" frame="void" id="id5" rules="none">
319     <colgroup><col class="label" /><col /></colgroup>
320     <tbody valign="top">
321     <tr><td class="label"><a name="id5">[2]</a></td><td><em>(<a class="fn-backref" href="#id2">1</a>, <a class="fn-backref" href="#id3">2</a>)</em> kdelibs-3.1.4-3.1.5.patch.bz2, switching format patch, created via diffball-0.4_pre4 (diffball is available at <a class="reference" href="http://sourceforge.net/projects/diffball">http://sourceforge.net/projects/diffball</a>)
322     Bzip2 -9 compressed, the patch is 75,687 bytes, uncompressed it is 337,649 bytes. The patch is available at <a class="reference" href="http://dev.gentoo.org/~ferringb/kdelibs-3.1.4-3.1.5.patch.bz2">http://dev.gentoo.org/~ferringb/kdelibs-3.1.4-3.1.5.patch.bz2</a> for those curious.</td></tr>
323     </tbody>
324     </table>
325     <table class="footnote" frame="void" id="id6" rules="none">
326     <colgroup><col class="label" /><col /></colgroup>
327     <tbody valign="top">
328     <tr><td class="label"><a name="id6">[3]</a></td><td>Glep9, 'Gentoo Package Update System'
329     (<a class="reference" href="http://glep.gentoo.org/glep-0009.html">http://glep.gentoo.org/glep-0009.html</a>)</td></tr>
330     </tbody>
331     </table>
332     </div>
333     <div class="section" id="copyright">
334 g2boojum 1.2 <h1><a class="toc-backref" href="#id25" name="copyright">Copyright</a></h1>
335 g2boojum 1.1 <p>This document has been placed in the public domain.</p>
336     </div>
337     </div>
338    
339 g2boojum 1.2 <hr class="footer" />
340 g2boojum 1.1 <div class="footer">
341     <a class="reference" href="glep-0025.txt">View document source</a>.
342 g2boojum 1.2 Generated on: 2004-11-11 21:27 UTC.
343 g2boojum 1.1 Generated by <a class="reference" href="http://docutils.sourceforge.net/">Docutils</a> from <a class="reference" href="http://docutils.sourceforge.net/rst.html">reStructuredText</a> source.
344     </div>
345     </body>
346     </html>

  ViewVC Help
Powered by ViewVC 1.1.20