Граф коммитов

100 Коммитов

Автор SHA1 Сообщение Дата
ajschult%verizon.net d082190de3 Bug 215701: Handle corrupt training.dat files better by not bailing if a token claims to extend beyond the end of the file. r=mscott, sr=bienvenu 2006-09-30 02:10:17 +00:00
bienvenu%nventure.com e024302e83 add hooks between mime parser, content handlers, and message display, sr=mscott 351610 2006-09-27 20:39:35 +00:00
gavin%gavinsharp.com c1553fb22f Bug 240788: Marking message as junk sometimes hangs Thunderbird, patch by Jeongkyu Kim <jeongkyu.kim@gmail.com>, r=mscott, sr=bienvenu 2006-09-12 22:15:49 +00:00
timeless%mozdev.org 97d86b0df2 Bug 106386 Correct misspellings in source code
r=bernd rs=brendan
2006-08-28 20:12:37 +00:00
gavin%gavinsharp.com 7af2b70e56 Bug 283493: flushing of training data during session practically never happens due to too high "# of changes" treshold, patch by Tuukka Tolvanen <tuukka.tolvanen@gmail.com>, r=eyalroz, sr=mscott 2006-07-19 13:38:19 +00:00
benjamin%smedbergs.us d8f81df51d Bug 78081 - don't export intermediate libraries - initial makefile changes to switch from $(DIST)/lib to in-tree library locations - blanket-r=mento 2006-03-21 18:19:14 +00:00
bienvenu%nventure.com 887b3aa97d fix assertion when analyzing spam with japanese characters, for example, sr=mscott 251882 2006-02-24 23:26:15 +00:00
cbiesinger%web.de c2d981f764 bug 183156 remove *UCS2* functions, replacing them with *UTF16* ones
r+sr=darin
2006-02-03 14:18:39 +00:00
mrbkap%gmail.com c4f20f9306 Remove the obsolete concept of "DTD verification". bug 323958, r+sr=jst 2006-01-30 23:45:19 +00:00
mrbkap%gmail.com 5e3b39b84e bug 324738: Remove some unnecessary complexity from the parser and hardcode DTDs. r=sicking sr=jst 2006-01-26 19:07:41 +00:00
bugzilla%standard8.demon.co.uk db22b0e21d Bug 63083 mailnews makefiles use EXPORTS instead of INCLUDES. Part 4 - move *CID.h files to public directories if they are exported. r/sr=dmose 2005-12-28 19:52:40 +00:00
bugzilla%standard8.demon.co.uk 08ca64963b Fix burning from bug checkin of 33451 attempt 2, put linking against xpcom_obsolete back in to make mac and windows happy. 2005-10-05 17:11:18 +00:00
bugzilla%standard8.demon.co.uk ebd2318596 Part of Bug 33451 convert mailnews to nsIFile. Remove some redundant xpcom_obsolete dependencies from mailnews (follow up to bug 132180). r=dmose,sr=bienvenu 2005-10-05 15:52:30 +00:00
bienvenu%nventure.com 554b7d4aeb extend support for opening .eml files in a stand-alone msg window, sr=mscott, a=chofmann 241212 2005-06-01 19:04:21 +00:00
bsmedberg%covad.net 659b20504a Adding MOZILLA_INTERNAL_API to internal makefiles, bug 288626, rs=darin 2005-04-05 13:09:19 +00:00
scott%scott-macgregor.org 099da7ba6d Bug #283080 --> bayesian training set is not getting flushed to disk on exit because
the profile directory service has already gone away.

sr=bienvenu
2005-02-22 18:11:55 +00:00
scott%scott-macgregor.org 1e5b836983 Bug #282174 --> Podcasting support for Thunderbird Add a custom mime part header that specifies a URL for the attachment.
This mime header is used by Thunderbird's RSS reader and eventually by the code to handle storing attachments separately
from the message. Removes an
obsolete attachment attribute called notDownloaded and replaces it with isExternalAttachment. sr=bienvenu a=sspitzer
2005-02-15 00:34:38 +00:00
scott%scott-macgregor.org 842ff87d68 Bug #277354 --> Make the bayesian filter handle japanese tokens
Patch by Noriyuki INOUE <norinoue@mbox.kyoto-inet.or.jp>

sr=me
2005-01-17 18:53:44 +00:00
timeless%mozdev.org 5a770b0df2 Bug 245499 Need to alter junk plugin training data write policy and/or cancel Start/EndBatch calls
patch by eyalroz@technion.ac.il r=bienvenu sr=mscott
2005-01-13 20:41:55 +00:00
gerv%gerv.net 670fb87360 Bug 236613: change to MPL/LGPL/GPL tri-license. 2004-12-01 22:27:03 +00:00
bienvenu%nventure.com c12b3d3c63 set has attachment flag at hdr download time by looking for content type multipart mixed, 241203 2004-10-25 18:48:22 +00:00
scott%scott-macgregor.org aa62f1a883 Bug #244357 --> Fix several underflow situations in the chi2 algorithm used for junk mail detection. Simplify the math to make it less computensive.
Patch origianlly by tenthumbs@cybernex.net

sr=mscott
2004-10-19 21:25:00 +00:00
scott%scott-macgregor.org fc12f7e08c Bug #240819 --> Crash in mail.dll when checking mail - TB073 [@ nsTransform2D::SetToIdentity ]
Fix a potential divide by zero floating point operation in the bayesian algorithm to hopefully prevent
a top crash.
2004-08-26 03:32:11 +00:00
scott%scott-macgregor.org d9839dd2c9 Bug #216133 --> port thunderbird changes for the new content policy manager from the branch to the trunk. 2004-07-01 02:43:02 +00:00
scott%scott-macgregor.org 70fa757a05 Bug #244325 --> broadcast a notification to the UI when a message is finished being rendered.
sr=bienvenu
2004-06-29 07:08:18 +00:00
scott%scott-macgregor.org 8d71634ad7 Bug #245176 --> several unsigned integers in the core bayesian algorithm should be signed integers because
they can be negative. Slightly improves our false negative rate on the benchmark.

Patch by Lorenzo Colitti
2004-06-01 03:26:28 +00:00
scott%scott-macgregor.org 024a5acba0 First part of Bug #243680 --> Marking mails as "not junk" does not update training.dat / spam filter does not create that file at all.
No one was ever calling shutdown on the plugin so we never forced ourselves to save the training data
if it was dirty.

sr=bienvenu
2004-05-18 23:41:54 +00:00
scott%scott-macgregor.org 5f44b200a9 Try to fix ports bustage by using PR_ABS instead of manually defining abs
thanks to bz for the suggestion
2004-05-13 04:03:17 +00:00
scott%scott-macgregor.org ea0646f211 Bug #230093, Bug #181534, Bug #237095 --> Port thunderbird junk mail improvements to the trunk.
Replace the core bayesian junk mail algorithm with a chi-squared probability distribution
modeled after spam bayes and Gary Robinson's work.

Change the model for how we count tokens across messages.

Token counts get out of alignment when re-training against already classified messages.

Revamp the junk mail tokenizer. Make it a hdr sink listener and add custom tokens for attachment
information. Ignore tokens larger than 13 characters. Tokenize purely off of white space.
Ignore tokens less than 3 bytes in length. There is still a lot more work to be done to the tokenizer.


Many thanks to Miguel Varga for working out the initial core algorithm improvement and to all
of the folks at spam bayes and of course Gary Robinson for helping to make this happen.
2004-05-12 18:16:32 +00:00
gerv%gerv.net dfb29e9236 Bug 236613: change to MPL/LGPL/GPL tri-license. 2004-04-17 18:33:16 +00:00
scott%scott-macgregor.org 89564fdd30 Bug #237151 --> Make the junk mail filter support the ability to reset the training data. This code is not called by anyone yet but is part of the build.
sr=bienvenu
a=asa
2004-03-12 19:14:50 +00:00
darin%meer.net 8d65e087ae landing string defragmentation patch for bug 231995, r/sr=dbaron,jst,dougt 2004-02-19 02:44:03 +00:00
bsmedberg%covad.net d0f309943a Continuing to land the PACKAGING_20030906_BRANCH for bug 20640. Not part of the build, yet. 2004-01-07 13:37:00 +00:00
bsmedberg%covad.net 274ef7cd49 Beginning to land the PACKAGING_20030906_BRANCH for bug 20640. Not part of the build, yet. 2004-01-07 01:22:31 +00:00
bienvenu%nventure.com 438ed1651a make bayesian filter plugin not rely on the profile manager, r/sr=mscott 219722 2003-09-19 19:02:00 +00:00
bryner%brianryner.com 5cd0bcdb7a Fix build on gcc 3.4 by removing extra semicolons (bug 218551). r/sr=dbaron, a=brendan. 2003-09-07 22:56:05 +00:00
cls%seawood.org 1b51ba858c Set MODULE in makefiles at the top of a heirarchy so that module-deps lists are more precise and builds will have the proper order if some subdirs contain other modules. 2003-08-16 00:42:35 +00:00
brendan%mozilla.org b7cdb7debb Add shared DHashTableOps for [const] char *key use-cases, clean up dhash API abusages (214839, r=dougt, sr=dbaron). 2003-08-05 20:09:21 +00:00
jaggernaut%netscape.com 2d7ed90a69 Bug 73353: clean up MODULE/REQUIRES story. r=cls, sr=bryner 2003-07-13 22:29:00 +00:00
seawood%netscape.com b28ce0a530 Removing old cfm build files. Use the CFM_LAST_RITES tag to resurrect. r=macdev 2003-06-10 20:12:33 +00:00
sspitzer%netscape.com 13da6b1dd4 fix for bug #206339
add PR_LOGGING for the bayesian filter.
people will have problems, and this will help dev and QA track them down.
it's *really* verbose, but that can be addressed later.
r/sr=bienvenu, a=sspitzer for 1.4 final
2003-05-23 07:34:19 +00:00
scott%scott-macgregor.org eb5faba111 Part of Bug #179533 --> remove obsolete CID file (has no effect on the current build) since this file has
been moved to ..\src.
2003-05-18 19:25:59 +00:00
shliang%netscape.com fa7043c160 bug 205054 - need descriptive text dialog to come up for user's first junk mail action. r/sr/a=sspitzer 2003-05-16 01:10:33 +00:00
scott%scott-macgregor.org 8b5be13a9d Bug #179533 --> add the option to make a static build of mailnews which is disabled by default.
r/sr=sspiter
2003-04-20 20:56:21 +00:00
brendan%mozilla.org bdeb25acd4 Add const to PLDHashTableOps static; fix warning and else-after-return. 2003-04-20 03:36:53 +00:00
scott%scott-macgregor.org 7bf4126720 Part of Bug #179533 --> need to move mail CID files from the build directory to the src directory.
This will make aggregating mail dlls easier.
r/sr=sspitzer
2003-04-19 05:29:35 +00:00
sspitzer%netscape.com adce3a35d1 fix for bug #194238
the junk mail controls won't detect a message as junk until you mark a message as good.

the algorithm in "A Plan For Spam" assumes that you have a large good
corpus and a large junk corpus.

that won't be the case with users who first use the junk mail feature
so, we do certain things to encourage them to train.

r/sr=bienvenu
2003-03-16 07:09:41 +00:00
bryner%netscape.com 1d7a118270 attempting to fix mac bustage 2003-03-15 05:30:48 +00:00
dbaron%dbaron.org acdd4e46c0 Bug 178643: Remove uses of NS_INIT_ISUPPORTS, since it is no longer needed. r=timeless sr=jag 2003-01-08 22:05:52 +00:00
bienvenu%netscape.com c42faf989e move link of unicharutil to build from src to try to fix build bustage 2002-12-07 01:37:09 +00:00