Contents of /xml/htdocs/proj/en/glep/glep-0044.html

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1.10 - (show annotations) (download) (as text)
Sun Oct 14 17:00:15 2007 UTC (10 years, 11 months ago) by antarus
Branch: MAIN
Changes since 1.9: +6 -253 lines
File MIME type: text/html
the canary on 53 went well, changing the rest

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">
5 <head>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
7 <meta name="generator" content="Docutils 0.4: http://docutils.sourceforge.net/" />
8 <title>GLEP 44 -- Manifest2 format</title>
9 <link rel="stylesheet" href="tools/glep.css" type="text/css" />
10 </head>
11 <body bgcolor="white">
12 <table class="navigation" cellpadding="0" cellspacing="0"
13 width="100%" border="0">
14 <tr><td class="navicon" width="150" height="35">
15 <a href="http://www.gentoo.org/" title="Gentoo Linux Home Page">
16 <img src="http://www.gentoo.org/images/gentoo-new.gif" alt="[Gentoo]"
17 border="0" width="150" height="35" /></a></td>
18 <td class="textlinks" align="left">
19 [<b><a href="http://www.gentoo.org/">Gentoo Linux Home</a></b>]
20 [<b><a href="http://www.gentoo.org/proj/en/glep">GLEP Index</a></b>]
21 [<b><a href="http://www.gentoo.org/proj/en/glep/glep-0044.txt">GLEP Source</a></b>]
22 </td></tr></table>
23 <table class="rfc2822 docutils field-list" frame="void" rules="none">
24 <col class="field-name" />
25 <col class="field-body" />
26 <tbody valign="top">
27 <tr class="field"><th class="field-name">GLEP:</th><td class="field-body">44</td>
28 </tr>
29 <tr class="field"><th class="field-name">Title:</th><td class="field-body">Manifest2 format</td>
30 </tr>
31 <tr class="field"><th class="field-name">Version:</th><td class="field-body">1.7</td>
32 </tr>
33 <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.cgi/xml/htdocs/proj/en/glep/glep-0044.txt?cvsroot=gentoo">2006/10/14 02:55:39</a></td>
34 </tr>
35 <tr class="field"><th class="field-name">Author:</th><td class="field-body">Marius Mauch &lt;genone&#32;&#97;t&#32;gentoo.org&gt;,</td>
36 </tr>
37 <tr class="field"><th class="field-name">Status:</th><td class="field-body">Accepted</td>
38 </tr>
39 <tr class="field"><th class="field-name">Type:</th><td class="field-body">Standards Track</td>
40 </tr>
41 <tr class="field"><th class="field-name">Content-Type:</th><td class="field-body"><a class="reference" href="glep-0002.html">text/x-rst</a></td>
42 </tr>
43 <tr class="field"><th class="field-name">Created:</th><td class="field-body">04-Dec-2005</td>
44 </tr>
45 <tr class="field"><th class="field-name">Post-History:</th><td class="field-body">06-Dec-2005, 23-Jan-2006, 3-Sep-2006</td>
46 </tr>
47 </tbody>
48 </table>
49 <hr />
50 <div class="contents topic">
51 <p class="topic-title first"><a id="contents" name="contents">Contents</a></p>
52 <ul class="simple">
53 <li><a class="reference" href="#abstract" id="id9" name="id9">Abstract</a></li>
54 <li><a class="reference" href="#motivation" id="id10" name="id10">Motivation</a></li>
55 <li><a class="reference" href="#specification" id="id11" name="id11">Specification</a><ul>
56 <li><a class="reference" href="#compability-entries" id="id12" name="id12">Compability Entries</a></li>
57 <li><a class="reference" href="#scope" id="id13" name="id13">Scope</a></li>
58 <li><a class="reference" href="#number-of-hashes" id="id14" name="id14">Number of hashes</a></li>
59 </ul>
60 </li>
61 <li><a class="reference" href="#rationale" id="id15" name="id15">Rationale</a><ul>
62 <li><a class="reference" href="#removal-of-digest-files" id="id16" name="id16">Removal of digest files</a></li>
63 <li><a class="reference" href="#reducing-redundancy" id="id17" name="id17">Reducing redundancy</a></li>
64 <li><a class="reference" href="#removal-of-checksum-collisions" id="id18" name="id18">Removal of checksum collisions</a></li>
65 <li><a class="reference" href="#flexible-verification-system" id="id19" name="id19">Flexible verification system</a></li>
66 </ul>
67 </li>
68 <li><a class="reference" href="#backwards-compatibility" id="id20" name="id20">Backwards Compatibility</a></li>
69 <li><a class="reference" href="#other-problems" id="id21" name="id21">Other problems</a><ul>
70 <li><a class="reference" href="#impacts-on-infrastructure" id="id22" name="id22">Impacts on infrastructure</a></li>
71 </ul>
72 </li>
73 <li><a class="reference" href="#reference-implementation" id="id23" name="id23">Reference Implementation</a></li>
74 <li><a class="reference" href="#options" id="id24" name="id24">Options</a></li>
75 <li><a class="reference" href="#credits" id="id25" name="id25">Credits</a></li>
76 <li><a class="reference" href="#references" id="id26" name="id26">References</a></li>
77 <li><a class="reference" href="#copyright" id="id27" name="id27">Copyright</a></li>
78 </ul>
79 </div>
80 <div class="section">
81 <h1><a class="toc-backref" href="#id9" id="abstract" name="abstract">Abstract</a></h1>
82 <p>This GLEP proposes a new format for the Portage Manifest and digest file system
83 by unifying both filetypes into one to improve functional and non-functional
84 aspects of the Portage Tree.</p>
85 </div>
86 <div class="section">
87 <h1><a class="toc-backref" href="#id10" id="motivation" name="motivation">Motivation</a></h1>
88 <p>Please see <a class="footnote-reference" href="#reorg-thread" id="id1" name="id1">[1]</a> for a general overview.
89 The main long term goals of this proposal are to:</p>
90 <ul class="simple">
91 <li>Remove the tiny digest files from the tree. They are a major annoyance as on a
92 typical configuration they waste a lot of disk space and the simple transmission
93 of the names for all digest files during a <tt class="docutils literal"><span class="pre">emerge</span> <span class="pre">--sync</span></tt> needs a substantial
94 amount of bandwidth.</li>
95 <li>Reduce redundancy when multiple hash functions are used</li>
96 <li>Remove potential for checksum collisions if a file is recorded in more than one
97 digest file</li>
98 <li>Difference between filetypes for a more flexible verification system</li>
99 </ul>
100 </div>
101 <div class="section">
102 <h1><a class="toc-backref" href="#id11" id="specification" name="specification">Specification</a></h1>
103 <p>The new Manifest format would change the existing format in the following ways:</p>
104 <ul>
105 <li><p class="first">Addition of a filetype specifier, currently planned are</p>
106 <ul class="simple">
107 <li><tt class="docutils literal"><span class="pre">AUX</span></tt> for files directly used by ebuilds (e.g. patches or initscripts),
108 located in the <tt class="docutils literal"><span class="pre">files/</span></tt> subdirectory</li>
109 <li><tt class="docutils literal"><span class="pre">EBUILD</span></tt> for all ebuilds</li>
110 <li><tt class="docutils literal"><span class="pre">MISC</span></tt> for files not directly used by ebuilds like <tt class="docutils literal"><span class="pre">ChangeLog</span></tt> or
111 <tt class="docutils literal"><span class="pre">metadata.xml</span></tt> files</li>
112 <li><tt class="docutils literal"><span class="pre">DIST</span></tt> for release tarballs recorded in the <tt class="docutils literal"><span class="pre">SRC_URI</span></tt> variable of an ebuild,
113 these were previously recorded in the digest files</li>
114 </ul>
115 <p>Future portage improvements might extend this list (for example with types
116 relevant for eclasses or profiles)</p>
117 </li>
118 <li><p class="first">Only have one line per file listing all information instead of one line per
119 file and checksum type</p>
120 </li>
121 <li><p class="first">Remove the separated digest-* files in the <tt class="docutils literal"><span class="pre">files/</span></tt> subdirectory</p>
122 </li>
123 </ul>
124 <p>Each line in the new format has the following format:</p>
125 <pre class="literal-block">
126 &lt;filetype&gt; &lt;filename&gt; &lt;filesize&gt; &lt;chksumtype1&gt; &lt;chksum1&gt; ... &lt;chksumtypen&gt; &lt;chksumn&gt;
127 </pre>
128 <p>However theses entries will be stored in the existing Manifest files.</p>
129 <p>An <a class="reference" href="glep-0044-extras/manifest2-example.txt">actual example</a> <a class="footnote-reference" href="#id7" id="id8" name="id8">[6]</a> for a (pure) Manifest2 file..</p>
130 <div class="section">
131 <h2><a class="toc-backref" href="#id12" id="compability-entries" name="compability-entries">Compability Entries</a></h2>
132 <p>To maintain compability with existing portage versions a transition period after
133 is the introduction of the Manifest2 format is required during which portage
134 will not only have to be capable of using existing Manifest and digest files but
135 also generate them in addition to the new entries.
136 Fortunately this can be accomplished by simply mixing old and new style entries
137 in one file for the Manifest files, existing portage versions will simply ignore
138 the new style entries. For the digest files there are no new entries to care
139 about.</p>
140 </div>
141 <div class="section">
142 <h2><a class="toc-backref" href="#id13" id="scope" name="scope">Scope</a></h2>
143 <p>It is important to note that this proposal only deals with a change of the
144 format of the digest and Manifest system.</p>
145 <p>It does not expand the scope of it to cover eclasses, profiles or anything
146 else not already covered by the Manifest system, it also doesn't affect
147 the Manifest signing efforts in any way (though the implementations of both
148 might be coupled).</p>
149 <p>Also while multiple hash functions will become standard with the proposed
150 implementation they are not a specific feature of this format <a class="footnote-reference" href="#multi-hash-thread" id="id3" name="id3">[2]</a>.</p>
151 </div>
152 <div class="section">
153 <h2><a class="toc-backref" href="#id14" id="number-of-hashes" name="number-of-hashes">Number of hashes</a></h2>
154 <p>While using multiple hashes for each file is a major feature of this proposal
155 we have to make sure that the number of hashes listed is limited to avoid
156 an explosion of the Manifest size that would revert the main benefit of this proposal
157 (reduzing tree size). Therefore the number of hashes that will be generated
158 will be limited to three different hash functions. For compability though we
159 have to rely on at least one hash function to always be present, this proposal
160 suggest to use SHA1 for this purpose (as it is supposed to be more secure than MD5
161 and currently only SHA1 and MD5 are directly available in python, also MD5 doesn't
162 have any benefit in terms of compability).</p>
163 </div>
164 </div>
165 <div class="section">
166 <h1><a class="toc-backref" href="#id15" id="rationale" name="rationale">Rationale</a></h1>
167 <p>The main goals of the proposal have been listed in the <a class="reference" href="#motivation">Motivation</a>, here now
168 the explanation why they are improvements and how the proposed format will
169 accomplish them.</p>
170 <div class="section">
171 <h2><a class="toc-backref" href="#id16" id="removal-of-digest-files" name="removal-of-digest-files">Removal of digest files</a></h2>
172 <p>Normal users that don't use a &quot;tuned&quot; filesystem for the portage tree are
173 wasting several dozen to a few hundred megabytes of disk space with the current
174 system, largely caused by the digest files.
175 This is due to the filesystem overhead present in most filesystem that
176 have a standard blocksize of four kilobytes while most digest files are under
177 one kilobyte in size, so this results in approximately a waste of three kilobytes
178 per digest file (likely even more). At the time of this writing the tree contains
179 roughly 22.000 digest files, so the overall waste caused by digest files is
180 estimated at about 70-100 megabytes.
181 Furthermore it is assumed that this will also reduce the disk space wasted by
182 the Manifest files as they now contain more content, but this hasn't been
183 verified yet.</p>
184 <p>By unifying the digest files with the Manifest these tiny files are eliminated
185 (in the long run), reducing the apparent tree size by about 20%, benefitting
186 both users and the Gentoo infrastructure.</p>
187 </div>
188 <div class="section">
189 <h2><a class="toc-backref" href="#id17" id="reducing-redundancy" name="reducing-redundancy">Reducing redundancy</a></h2>
190 <p>When multiple hashes are used with the current system
191 both the filename and filesize are repeated for every checksum type used as each
192 checksum is standalone. However this doesn't add any functionality and is
193 therefore useless, so the new format removes this redundancy.
194 This is a theoretical improvement at this moment as only one hash function is in
195 use, but expected to change soon (see <a class="footnote-reference" href="#multi-hash-thread" id="id4" name="id4">[2]</a>).</p>
196 </div>
197 <div class="section">
198 <h2><a class="toc-backref" href="#id18" id="removal-of-checksum-collisions" name="removal-of-checksum-collisions">Removal of checksum collisions</a></h2>
199 <p>The current system theoretically allows for a <tt class="docutils literal"><span class="pre">DIST</span></tt> type file to be recorded
200 in multiple digest files with different sizes and/or checksums. In such a case
201 one version of a package would report a checksum violation while another one
202 would not. This could create confusion and uncertainity among users.
203 So far this case hasn't been observed, but it can't be ruled out with the
204 existing system.
205 As the new format lists each file exactly once this would be no longer possible.</p>
206 </div>
207 <div class="section">
208 <h2><a class="toc-backref" href="#id19" id="flexible-verification-system" name="flexible-verification-system">Flexible verification system</a></h2>
209 <p>Right now portage verifies the checksum of every file listed in the Manifest
210 before using any file of the package and all <tt class="docutils literal"><span class="pre">DIST</span></tt> files of an ebuild
211 before using that ebuild. This is unnecessary in many cases:</p>
212 <ul class="simple">
213 <li>During the &quot;depend&quot; phase (when the ebuild metadata is generated) only
214 files of type <tt class="docutils literal"><span class="pre">EBUILD</span></tt> are used, so verifying the other types isn't
215 necessary. Theoretically it is possible for an ebuild to include other
216 files like those of type <tt class="docutils literal"><span class="pre">AUX</span></tt> at this phase, but that would be a
217 major QA violation and should never occur, so it can be ignored here.
218 It is also not a security concern as the ebuild is verified before parsing
219 it, so each manipulation would show up.</li>
220 <li>Generally files of type <tt class="docutils literal"><span class="pre">MISC</span></tt> don't need to be verified as they are
221 only used in very specific situations, aren't executed (just parsed at most)
222 and don't affect the package build process.</li>
223 <li>Files of type <tt class="docutils literal"><span class="pre">DIST</span></tt> only need to be verified directly after fetching and
224 before unpacking them (which often will be one step), not every time their
225 associated ebuild is used.</li>
226 </ul>
227 </div>
228 </div>
229 <div class="section">
230 <h1><a class="toc-backref" href="#id20" id="backwards-compatibility" name="backwards-compatibility">Backwards Compatibility</a></h1>
231 <p>Switching the Manifest system is a task that will need a long transition period
232 like most changes affecting both portage and the tree. In this case the
233 implementation will be rolled out in several phases:</p>
234 <ol class="arabic simple">
235 <li>Add support for verification of Manifest2 entries in portage</li>
236 <li>Enable generation of Manifest2 entries in addition to the current system</li>
237 <li>Ignore digests during <tt class="docutils literal"><span class="pre">emerge</span> <span class="pre">--sync</span></tt> to get the size-benefit clientside.
238 This step may be ommitted if the following steps are expected to follow soon.</li>
239 <li>Disable generation of entries for the current system</li>
240 <li>Remove all traces of the current system from the tree (serverside)</li>
241 </ol>
242 <p>Each step has its own issues. While 1) and 2) can be implemented without any
243 compability problems all later steps have a major impact:</p>
244 <ul class="simple">
245 <li>Step 3) can only be implemented when the whole tree is Manifest2 ready
246 (ideally speaking, practically the requirement will be more like 95% coverage
247 with the expectation that for the remaining 5% either bugs will be filed after
248 step 3) is completed or they'll be updated at step 5).</li>
249 <li>Steps 4) and 5) will render all portage versions without Manifest2 support
250 basically useless (users would have to regenerate the digest and Manifest
251 for each package before being able to merge it), so this requires a almost
252 100% coverage of the userbase with Manifest2 capabale portage versions
253 (with step 1) completely implemented).</li>
254 </ul>
255 <p>Another problem is that some steps affect different targets:</p>
256 <ul class="simple">
257 <li>Steps 1) and 3) target portage versions used by users</li>
258 <li>Steps 2) and 4) target portage versions used by devs</li>
259 <li>Step 5) targets the portage tree on the cvs server</li>
260 </ul>
261 <p>While it is relatively easy to get all devs to use a new portage version this is
262 practically impossible with users as some don't update their systems regulary.
263 While six months are probably sufficient to reach a 95% coverage one year is
264 estimated to reach an almost-complete coverage. All times are relative to the
265 stable-marking of a compatible portage version.</p>
266 <p>No timeframe for implementation is presented here as it is highly dependent
267 on the completion of each step.</p>
268 <p>In summary it can be said that while a full conversion will take over a year
269 to be completed due to compability issues mentioned above some benefits of the
270 system can selectively be used as soon as step 2) is completed.</p>
271 </div>
272 <div class="section">
273 <h1><a class="toc-backref" href="#id21" id="other-problems" name="other-problems">Other problems</a></h1>
274 <div class="section">
275 <h2><a class="toc-backref" href="#id22" id="impacts-on-infrastructure" name="impacts-on-infrastructure">Impacts on infrastructure</a></h2>
276 <p>While one long term goal of this proposal is to reduce the size of the tree
277 and therefore make life for the Gentoo Infrastructure easier this will only
278 take effect once the implementation is rolled out completely. In the meantime
279 however it will increase the tree size due to keeping checksums in both formats.
280 It's not possible to give a usable estimate on the degree of the increase as
281 it depends on many variables such as the exact implementation timeframe,
282 propagation of Manifest2 capable portage versions among devs or the update
283 rate of the tree. It has been suggested that Manifest files that are not gpg
284 signed could be mass converted in one step, this could certainly help but only
285 to some degree (according to a recent research <a class="footnote-reference" href="#gpg-numbers" id="id5" name="id5">[3]</a> about 40% of
286 all Manifests in the tree are signed, but this number hasn't been verified).</p>
287 </div>
288 </div>
289 <div class="section">
290 <h1><a class="toc-backref" href="#id23" id="reference-implementation" name="reference-implementation">Reference Implementation</a></h1>
291 <p>A patch for a prototype implementation of Manifest2 verification and partial
292 generation has been posted at <a class="footnote-reference" href="#manifest2-patch" id="id6" name="id6">[4]</a>, it will be reworked before
293 being considered for inclusion in portage. However it shows that adding support
294 for verification is quite simple, but generation is a bit tricky and will
295 therefore be implemented later.</p>
296 </div>
297 <div class="section">
298 <h1><a class="toc-backref" href="#id24" id="options" name="options">Options</a></h1>
299 <p>Some things have been considered for this GLEP but aren't part of the proposal
300 yet for various reasons:</p>
301 <ul class="simple">
302 <li>timestamp field: the author has considered adding a timestamp field for
303 each entry to list the time the entry was created. However so far no practical
304 use for such a feature has been found.</li>
305 <li>convert size field into checksum: Another idea was to treat the size field
306 like any other checksum. But so far no real benefit (other than a slightly
307 more modular implementation) for this has been seen while it has several
308 drawbacks: For once, unlike checksums, the size field is definitely required
309 for all <tt class="docutils literal"><span class="pre">DIST</span></tt> files, also it would slightly increase the length of
310 each entry by adding a <tt class="docutils literal"><span class="pre">SIZE</span></tt> keyword.</li>
311 <li>removal of the <tt class="docutils literal"><span class="pre">MISC</span></tt> type: It has been suggested to completely drop
312 entries of type <tt class="docutils literal"><span class="pre">MISC</span></tt>. This would result in a minor space reduction
313 (its rather unlikely to free any blocks) but completely remove the ability
314 to check these files for integrity. While they don't influence portage
315 or packages directly they can contain viable information for users, so
316 the author has the opinion that at least the option for integrity checks
317 should be kept.</li>
318 </ul>
319 </div>
320 <div class="section">
321 <h1><a class="toc-backref" href="#id25" id="credits" name="credits">Credits</a></h1>
322 <p>Thanks to the following persons for their input on or related to this GLEP
323 (even though they might not have known it):
324 Ned Ludd (solar), Brian Harring (ferringb), Jason Stubbs (jstubbs),
325 Robin H. Johnson (robbat2), Aron Griffis (agriffis)</p>
326 <p>Also thanks to Nicholas Jones (carpaski) to make the current Manifest system
327 resistent enough to be able to handle this change without too many transition
328 problems.</p>
329 </div>
330 <div class="section">
331 <h1><a class="toc-backref" href="#id26" id="references" name="references">References</a></h1>
332 <table class="docutils footnote" frame="void" id="reorg-thread" rules="none">
333 <colgroup><col class="label" /><col /></colgroup>
334 <tbody valign="top">
335 <tr><td class="label"><a class="fn-backref" href="#id1" name="reorg-thread">[1]</a></td><td><a class="reference" href="http://thread.gmane.org/gmane.linux.gentoo.devel/21920">http://thread.gmane.org/gmane.linux.gentoo.devel/21920</a></td></tr>
336 </tbody>
337 </table>
338 <table class="docutils footnote" frame="void" id="multi-hash-thread" rules="none">
339 <colgroup><col class="label" /><col /></colgroup>
340 <tbody valign="top">
341 <tr><td class="label"><a name="multi-hash-thread">[2]</a></td><td><em>(<a class="fn-backref" href="#id3">1</a>, <a class="fn-backref" href="#id4">2</a>)</em> <a class="reference" href="http://thread.gmane.org/gmane.linux.gentoo.devel/33434">http://thread.gmane.org/gmane.linux.gentoo.devel/33434</a></td></tr>
342 </tbody>
343 </table>
344 <table class="docutils footnote" frame="void" id="gpg-numbers" rules="none">
345 <colgroup><col class="label" /><col /></colgroup>
346 <tbody valign="top">
347 <tr><td class="label"><a class="fn-backref" href="#id5" name="gpg-numbers">[3]</a></td><td>gentoo-core mailing list, topic &quot;Gentoo key signing practices
348 and official Gentoo keyring&quot;, Message-ID &lt;<a class="reference" href="mailto:20051117075838.GB15734&#64;curie-int.vc.shawcable.net">20051117075838.GB15734&#64;curie-int.vc.shawcable.net</a>&gt;</td></tr>
349 </tbody>
350 </table>
351 <table class="docutils footnote" frame="void" id="manifest2-patch" rules="none">
352 <colgroup><col class="label" /><col /></colgroup>
353 <tbody valign="top">
354 <tr><td class="label"><a class="fn-backref" href="#id6" name="manifest2-patch">[4]</a></td><td><a class="reference" href="http://thread.gmane.org/gmane.linux.gentoo.portage.devel/1374">http://thread.gmane.org/gmane.linux.gentoo.portage.devel/1374</a></td></tr>
355 </tbody>
356 </table>
357 <table class="docutils footnote" frame="void" id="manifest2-example" rules="none">
358 <colgroup><col class="label" /><col /></colgroup>
359 <tbody valign="top">
360 <tr><td class="label"><a name="manifest2-example">[5]</a></td><td><a class="reference" href="http://www.gentoo.org/proj/en/glep/glep-0044-extras/manifest2-example">http://www.gentoo.org/proj/en/glep/glep-0044-extras/manifest2-example</a></td></tr>
361 </tbody>
362 </table>
363 <table class="docutils footnote" frame="void" id="id7" rules="none">
364 <colgroup><col class="label" /><col /></colgroup>
365 <tbody valign="top">
366 <tr><td class="label"><a class="fn-backref" href="#id8" name="id7">[6]</a></td><td><a class="reference" href="glep-0044-extras/manifest2-example.txt">glep-0044-extras/manifest2-example.txt</a></td></tr>
367 </tbody>
368 </table>
369 </div>
370 <div class="section">
371 <h1><a class="toc-backref" href="#id27" id="copyright" name="copyright">Copyright</a></h1>
372 <p>This document has been placed in the public domain.</p>
373 </div>
375 </div>
376 <div class="footer">
377 <hr class="footer" />
378 <a class="reference" href="glep-0044.txt">View document source</a>.
379 Generated on: 2007-10-13 13:39 UTC.
380 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.
382 </div>
383 </body>
384 </html>

  ViewVC Help
Powered by ViewVC 1.1.20