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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.3 - (show annotations) (download)
Fri Apr 1 01:32:19 2005 UTC (9 years, 9 months ago) by g2boojum
Branch: MAIN
CVS Tags: HEAD
Changes since 1.2: +4 -4 lines
File MIME type: text/plain
spelling fix

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

  ViewVC Help
Powered by ViewVC 1.1.20