2003-02-20 13:12 jmason * lib/Mail/SpamAssassin.pm: getting ready for 2.50 release 2003-02-20 13:10 jmason * rules/20_uri_tests.cf: Dan's fix for bug 1501: WEIRD_PORT misfiring on port 443 for https, which is the https port 2003-02-20 13:05 jmason * BUGS, t/razor.t: updated with 2 new known bugs 2003-02-20 13:01 jmason * BUGS, MANIFEST, MANIFEST.SKIP, MailAudit.patch, masses/RUNME.TWOSTEP, masses/craig-evolve.c.ALTIVEC, rules/40_spam_phrases.cf, t/reportheader.t, t/spamd_report.t, t/data/nice/001, wordfreqs/.cvsignore, wordfreqs/README, wordfreqs/RUNME, wordfreqs/RUNME.DOGMA, wordfreqs/RUNME.SONIC, wordfreqs/join-and-settle-phrases, wordfreqs/mass-phrase-freq, wordfreqs/mass-phrase-freq-for-jm, wordfreqs/settle-phrases: removed more obsolete code from CVS and MANIFEST 2003-02-20 12:23 jmason * sql/: README, spamassassin.sql: changed suggested username length limit to 100 chars in SQL setup, so that virtual domains can work easily 2003-02-20 11:22 quinlan * MANIFEST: sort 2003-02-20 11:14 quinlan * MANIFEST: adding stuff to MANIFEST 2003-02-20 10:55 msquadrat * .cvsignore, masses/.cvsignore, qmail/.cvsignore, t/.cvsignore: Updating cvsignore 2003-02-20 05:04 felicity * .cvsignore: we merged the sa-learn* scripts... 2003-02-20 04:50 quinlan * MANIFEST: remove duplicated triplets.txt file 2003-02-19 21:48 msquadrat * Makefile.PL: remove the spamassassin.conf manpage again: it's a nasty hack violating the FHS 2003-02-19 19:27 jmason * rules/: 20_phrases.cf, 20_porn.cf, 25_body_tests_es.cf, 25_body_tests_pl.cf, 30_text_de.cf, 30_text_es.cf, 30_text_fr.cf, 30_text_it.cf, 30_text_pl.cf: old removed rule descriptions removed from translation files 2003-02-19 19:01 jmason * rules/20_uri_tests.cf: bug 1497: PORN_4 still fping, hopeuflly fixed a little 2003-02-19 18:48 jmason * INSTALL: noting that RAzor tests do not work on windows 2003-02-19 18:22 jmason * rules/: STATISTICS-set1.txt, STATISTICS-set2.txt, STATISTICS-set3.txt, STATISTICS.txt: removed command lines ;) 2003-02-19 18:05 felicity * masses/mk-baseline-results: Updated slightly to deal with scoresets better 2003-02-19 18:03 jmason * README: patch from John Peacock to README 2003-02-19 17:58 felicity * rules/: STATISTICS-set1.txt, STATISTICS-set2.txt, STATISTICS-set3.txt, STATISTICS.txt: Fixed the stats files. Oops. 2003-02-19 15:17 felicity * sa-learn.raw: fixed typo in docs 2003-02-19 13:06 climent * rules/25_body_tests_es.cf: New tweaks, additions and corrections. 2003-02-19 06:22 felicity * rules/: STATISTICS-set1.txt, STATISTICS-set2.txt, STATISTICS-set3.txt, STATISTICS.txt: Statistics files for the 2.50 GA'ed scores, all score sets. :) 2003-02-19 01:14 msquadrat * Makefile.PL: implement bug 1496: a (kludgy) spamassassin.conf man page 2003-02-19 00:49 msquadrat * Makefile.PL, t/razor.t: Make RUN_RAZOR[12]_TESTS less strict and put it all into a nice clean block 2003-02-18 22:36 msquadrat * Makefile.PL, build/preprocessor.pm: Bugger about ExtUtils::MakeMaker < 5.45 at the right place 2003-02-18 22:33 quinlan * MANIFEST: add t/rule_tests.t 2003-02-18 22:19 jmason * rules/70_cvs_rules_under_test.cf: some 419 tests for later 2003-02-18 22:02 jmason * rules/70_cvs_rules_under_test.cf: new nigerian scam phrase test 2003-02-18 21:35 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: We should only add autolearn= if we actual did learning. If the message was already seen, etc, let's not add the token. 2003-02-18 21:19 msquadrat * build/preprocessor.pm: fix a problem with missing inheritance from ExtUtils::Install 2003-02-18 20:16 jmason * USAGE: added site-wide user-notification request 2003-02-18 17:54 felicity * lib/Mail/SpamAssassin/CmdLearn.pm: Now that rebuild is an option, not a parameter to the function, it's not in $opt{} anymore. 2003-02-18 17:47 felicity * lib/Mail/SpamAssassin/BayesStore.pm: All of the bayes_* files ought to have the same file permissions, so we need to do a chmod after the open for msgcount and journal. 2003-02-18 15:55 jmason * Makefile.PL: oops, broke Makefile.PL there 2003-02-18 15:39 jmason * Makefile.PL: Mike Bell: suggested fix for win32 errors when installing local.cf 2003-02-18 15:37 jmason * lib/Mail/SpamAssassin/: PerMsgStatus.pm, Util.pm: another way to do the getpwuid wrapper trick; should be safe on win32 for sure 2003-02-18 12:02 jmason * masses/mk-baseline-results: allow caller to specify output using redirection, so we can generate stats for each scoreset 2003-02-17 21:16 felicity * sa-learn.raw: Changed POD docs a little, s/nonspam/ham/, added --no-rebuild option to sample command. 2003-02-17 18:18 felicity * INSTALL: Added note about requiring Sys::Syslog if you're using spamd 2003-02-17 18:14 felicity * lib/Mail/SpamAssassin/NoMailAudit.pm: Fix for bug 1487. With the Date header at the end of the message, the RE was grabbing the blank lines as "folded"... 2003-02-17 10:54 msquadrat * lib/Mail/SpamAssassin/ArchiveIterator.pm: fix bug 1491: glob only leading tilde; escape white spaces 2003-02-17 10:20 felicity * lib/Mail/SpamAssassin/Reporter.pm: I think this fixes bug 1492. 2003-02-17 08:15 felicity * lib/Mail/SpamAssassin/Util.pm: $AM_TAINTED is always defined, we care what the value is 2003-02-17 00:42 felicity * sa-forget.raw, sa-learn-nonspam.raw, sa-learn-rebuild.raw, sa-learn-spam.raw: All of these got merged into sa-learn.raw. :) 2003-02-17 00:23 felicity * rules/: 20_html_tests.cf, 50_scores.cf: bug 1429: HTML_IMAGE_AREA_09 was autodisabled by GA, and _10 gets very good results, so rename _10 and lower the range to cover 09 as well. 2003-02-16 23:24 felicity * MANIFEST: updated for the merged sa-learn script 2003-02-16 23:03 felicity * rules/70_broken_rules.cf: added regression tests for rules that were disabled 2003-02-16 23:02 felicity * rules/regression_tests.cf: removed tests for rules that no longer exist 2003-02-16 22:33 felicity * t/rule_tests.t: Since we can test subrules now, we can't count on hits but have to look at what rules got hit. 2003-02-16 22:11 msquadrat * rules/20_ratware.cf: fix bug 1488: __OUTLOOK_MSGID too restrictive 2003-02-16 22:10 felicity * sa-learn.raw: This script is the now single tool to modify the learning systems in SpamAssassin. Used to be sa-learn-nonspam, the documentation was updated to cover all of the new functionality. 2003-02-16 22:07 felicity * Makefile.PL, README: added support for merge of sa-learn-* and sa-forget 2003-02-16 22:01 felicity * lib/Mail/SpamAssassin/CmdLearn.pm: added support for merge of sa-learn-* and sa-forget 2003-02-16 16:32 jmason * lib/Mail/SpamAssassin/BayesStore.pm, rules/20_meta_tests.cf, rules/regression_tests.cf: fixed some wierd warnings; OBFUSCATING_COMMENT regression test still failing, needs to test the __ version of the rule instead of the whole meta rule. 2003-02-16 07:39 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Headers that should be added (X-Pyzor, etc.) weren't being added to the report_safe header. 2003-02-16 07:31 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: If we call delete_fulltext_tmpfile() we ought to forget the previous filename that we used. 2003-02-16 05:46 felicity * spamd/libspamc.c: debug statement outside of an ifdef. oops. 2003-02-16 03:58 felicity * rules/: 20_html_tests.cf, 20_meta_tests.cf: fix for bug 1423. OBFUSCATING_COMMENT looks for larger comments and is now a meta with MIME_HTML_ONLY 2003-02-16 02:19 felicity * rules/: 50_scores.cf, 70_broken_rules.cf: re-enabled 3 rules which are needed for meta rules. left USER_AGENT_* rules disabled since USER_AGENT is the meta rule which has the dependencies, and it's disabled too. 2003-02-16 01:58 felicity * rules/: 50_scores.cf, 70_broken_rules.cf: Lower any abs(score) < 0.01 to 0. Added entries to broken_rules to set those rules to score 0.001 so that mass-check will still use them, but not the released version. 2003-02-15 23:24 felicity * rules/20_head_tests.cf: Changed description for NO_DNS_FOR_FROM 2003-02-15 23:20 felicity * lib/Mail/SpamAssassin/EvalTests.pm, rules/20_head_tests.cf, rules/30_text_de.cf, rules/30_text_es.cf, rules/30_text_fr.cf, rules/30_text_it.cf, rules/30_text_pl.cf, rules/50_scores.cf: bug 1381, renamed NO_MX_FOR_FROM to NO_DNS_FOR_FROM since the code checks MX and A records... 2003-02-15 15:02 msquadrat * t/spamc.t: Port 9 (discard) is the correct one to use in this test 2003-02-15 14:28 felicity * rules/50_scores.cf: GA set 3 scores 2003-02-15 11:07 climent * rules/25_body_tests_es.cf: Tweaks and new entries in old rules 2003-02-15 02:23 felicity * rules/50_scores.cf: small mistakes in a few net test scores for set 3 causing ga issues ... 2003-02-15 02:06 felicity * rules/50_scores.cf: Set 2 scores put in 2003-02-14 22:20 felicity * lib/Mail/SpamAssassin/: Dns.pm, Reporter.pm: Pyzor and DCC options weren't being set correctly due to non-match of the RE to untaint them. 2003-02-14 21:57 felicity * lib/Mail/SpamAssassin/Util.pm: Fixed debug output for "Are we tainted?" 2003-02-14 21:57 jmason * tools/check_bayes_db: oops 2003-02-14 21:50 jmason * lib/Mail/SpamAssassin/BayesStore.pm, lib/Mail/SpamAssassin/Conf.pm, lib/Mail/SpamAssassin/HTML.pm, tools/check_bayes_db: bug 1460: bayes_expiry_use_scan_count is now the default, using a method that should not be much slower than using the time. This now means that expiry can take place during a mass-check, which is very important. Implementation uses a separate file, bayes_msgcount, to track scan-count. This will not require rescanning the bayes dbs. 2003-02-14 21:10 felicity * lib/Mail/: SpamAssassin.pm, SpamAssassin/Dns.pm: Fixed up the dcc/pyzor debugging statements 2003-02-14 21:06 felicity * lib/Mail/SpamAssassin/Dns.pm: Added a little more debugging info around pyzor and dcc 2003-02-14 20:59 felicity * README: Added some documentation about default file locations 2003-02-14 20:22 jmason * spamassassin.raw, lib/Mail/SpamAssassin/Conf.pm: oops, using 'blib' meant that rules were not being found when run from build dir. fixed 2003-02-14 19:48 felicity * lib/Mail/SpamAssassin.pm: need to untaint user prefs file before trying to write to it 2003-02-14 19:05 jmason * t/spamc_c.t: oops, too many tests 2003-02-14 19:04 jmason * lib/Mail/: SpamAssassin.pm, SpamAssassin/Dns.pm, SpamAssassin/Util.pm: bug 1479: getpwuid on Win32 does really wierd things 2003-02-14 18:58 jmason * lib/Mail/SpamAssassin/: Conf.pm, PerMsgStatus.pm: bug 1462: tag rule failures with rule names where possible 2003-02-14 18:52 felicity * lib/Mail/SpamAssassin/Util.pm: Added more debug output 2003-02-14 18:43 jmason * lib/Mail/SpamAssassin/AutoWhitelist.pm, lib/Mail/SpamAssassin/ConfSourceSQL.pm, spamd/spamd.raw: bug 1457: added skeleton POD doc to private classes 2003-02-14 18:38 jmason * MANIFEST, spamd/libspamc.c, t/spamc_c.t: bug 1481: spamc -c broken, fixed by Eugene Miretsky (eugene .AT. invision .DOT. net); added test too 2003-02-14 11:59 climent * rules/25_body_tests_es.cf: Small addition 2003-02-13 16:16 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Added a section to X-Spam-Status specifying that autolearning has taken place, and whether it was ham or spam. 2003-02-13 13:21 jmason * spamassassin.raw, lib/Mail/SpamAssassin.pm, lib/Mail/SpamAssassin/Reporter.pm: did I forget to commit the revoke patch? if so, here it is. 2003-02-12 23:05 jmason * lib/Mail/SpamAssassin/PerMsgStatus.pm: bug 1414: unicode chars still creeping in via HTML::Parser 2003-02-12 22:57 jmason * spamd/libspamc.c, spamd/libspamc.h, spamd/spamc.c, spamd/spamd.raw, t/SATest.pm: bug 1452: libspamc: message is always reported as EX_TOOBIG in m.is_spam. This requires spamc protocol version 1.3, which can now support multiple headers in reply from spamd. 2003-02-12 21:48 felicity * spamassassin.raw, t/SATest.pm: fix for bug 1464. makes spamassassin look for a "blib/lib", and if it exists, use it exclusively. 2003-02-12 21:11 felicity * rules/20_head_tests.cf: missing end periods on 3 tests. 2003-02-12 18:10 jmason * build/: update_devel, update_stable: don't add development doco to website every night 2003-02-12 17:41 felicity * rules/50_scores.cf: Added some comments to the scores file. Organized the network/etc area a little bit more. 2003-02-12 15:57 felicity * lib/Mail/SpamAssassin/Dns.pm: chomp the response to dcc and pyzor. don't need the newline and makes the debug statements look better. :) 2003-02-11 19:44 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Modified the debug statements around auto-learning. Show the thresholds, etc. 2003-02-11 16:59 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Fix for bug 1470, modified version for bug 1417. Basically HTML parsing was broken due to forcing byte mode during the parse. pack was called with bad parameters which nullified the text. doh. 2003-02-11 16:29 felicity * rules/50_scores.cf: 1 rule got disabled in set1 that's a meta dependency. re-enabled it... 2003-02-11 16:26 felicity * rules/50_scores.cf: A few more minor score cleanups 2003-02-11 16:16 felicity * rules/50_scores.cf: old version was fscked up (only 607 scores!?!) this one has all 844. 2003-02-11 13:13 jmason * t/SATest.pm: bug 1464: use 'perl -I../blib' to run test commands 2003-02-11 13:09 jmason * rules/: 70_cvs_rules_under_test.cf, 70_cvs_rules_under_test.cf: checked in possible replacement T_PENIS_ENLARGE rules for test after release 2003-02-11 13:07 jmason * rules/10_misc.cf, rules/20_anti_ratware.cf, rules/20_body_tests.cf, rules/20_compensate.cf, rules/20_head_tests.cf, rules/20_html_tests.cf, rules/20_meta_tests.cf, rules/20_phrases.cf, rules/20_porn.cf, rules/20_ratware.cf, rules/20_uri_tests.cf, rules/23_bayes.cf, rules/25_body_tests_es.cf, rules/25_body_tests_pl.cf, rules/25_head_tests_es.cf, rules/25_head_tests_pl.cf, rules/30_text_de.cf, rules/30_text_es.cf, rules/30_text_fr.cf, rules/30_text_it.cf, rules/30_text_pl.cf, rules/50_scores.cf, rules/60_whitelist.cf, rules/70_broken_rules.cf, rules/70_cvs_rules_under_test.cf, rules/regression_tests.cf, t/data/spam/009: added more spammy phrases to utf8 source message. But all those rules have disappeared too. ;) something's up with the scores file for sure... 2003-02-11 10:59 jmason * sa-learn-nonspam.raw: oops, documentation was out of date on auto_learn 2003-02-10 23:17 msquadrat * Makefile.PL: Don't require ExtUtils::MakeMaker >= 5.45; we have compatibility code in build/preprocessor.pm 2003-02-10 23:01 msquadrat * lib/Mail/SpamAssassin/Conf.pm: Backported fix for Bug 1120: dcc_path docu mixup (patch 1.115 -> 1.116) 2003-02-10 21:50 msquadrat * t/razor2.t: Backport fix for Bug 1117: Razor tests ignore enabled/disabled flag 2003-02-10 21:49 msquadrat * t/razor.t: Backport fix for Bug 1117: Razor tests ignore enabled/disabled flag (patch 1.5 -> 1.6) 2003-02-10 21:26 msquadrat * lib/Mail/SpamAssassin/Reporter.pm: Backport workaround for Razor2 disabling the timeout (patch 1.39 -> 1.40) 2003-02-10 18:04 jmason * lib/Mail/SpamAssassin/PerMsgStatus.pm: test_logs were persisting from one line to the next. fixed 2003-02-10 11:20 jmason * lib/Mail/SpamAssassin/Conf.pm, rules/70_cvs_rules_under_test.cf, t/data/nice/cjk/gb2312.2: documented always-add-report 'this mail is spam' message when the mail is not spam. Also added another good gb2312 mail to t/data/nice/cjk 2003-02-10 00:18 felicity * lib/Mail/SpamAssassin/Dns.pm: If we get a response, we're not skipping it, are we? 2003-02-10 00:09 felicity * lib/Mail/SpamAssassin/Dns.pm: Another stupid RAZOR2_CF_* bug ... If the first part didn't get a response (like when it's blank and Razor says "skipme=1"), the other parts aren't checked for a response. Fixed now. Doh. :) 2003-02-09 15:03 jmason * lib/Mail/SpamAssassin/Conf.pm: bayes_use_chi2_combining POD fixed to reflect default of 0 2003-02-09 14:46 jmason * lib/Mail/SpamAssassin/Reporter.pm: tidying up reporter undef value fix to match Dns.pm 2003-02-09 01:29 felicity * lib/Mail/SpamAssassin/Reporter.pm: possible undefined warning? dunno, it was added to dns ... 2003-02-09 01:17 felicity * lib/Mail/: SpamAssassin.pm, SpamAssassin/CmdLearn.pm: fixes for 1456. makes bayes_path work properly. basically the scoreset was chosen before the configs were read, so if the bayes stuff doesn't live in the user's directory it'll only choose sets 0 or 1, then bayes never runs. 2003-02-08 16:38 jmason * lib/Mail/SpamAssassin/Dns.pm: poss fix for that undefined warning in Dns.pm line 693 2003-02-08 16:21 jmason * lib/Mail/SpamAssassin/PerMsgStatus.pm: always de-unicode strings going into HTML::Parser, since recent versions do their own Unicode stuff 2003-02-08 15:44 jmason * USAGE: added UPPERCASE rules to set disabled for CJK users. added more doco on CJK rules and spamtraps 2003-02-08 15:31 jmason * rules/user_prefs.template: added UPPERCASE rules to set disabled for CJK users 2003-02-08 15:30 jmason * t/data/nice/cjk/gb2312.1: the easy way around the 8-bit-FP problem, for now, without redoing GA or mass-check runs; tell the users to disable the FPing rules. Also a collection of known-ham in CJK charsets 2003-02-08 15:26 jmason * INSTALL, MANIFEST.SKIP, rules/user_prefs.template, t/data/nice/cjk/big5-base64.1, t/data/nice/cjk/gb2312.1, t/data/nice/cjk/iso-2022-jp.1, t/data/nice/cjk/iso-2022-jp.2: the easy way around the 8-bit-FP problem, for now, without redoing GA or mass-check runs; tell the users to disable the FPing rules 2003-02-07 20:25 jmason * masses/: CORPUS_SUBMIT, CORPUS_SUBMIT_NIGHTLY: removed -u (update only, don't overwrite newer files) from rsync flags in corpus-submit doco, since it's just caused me some confusion at least ;) 2003-02-07 19:20 felicity * lib/Mail/SpamAssassin/Reporter.pm: Need to clean up after the tmp file ... 2003-02-07 19:15 felicity * lib/Mail/SpamAssassin/: Dns.pm, Reporter.pm: cleaned up a few things in dns. cleaned up a _lot_ of stuff in reporter. while working on a fix for bug 1447 I found that it was just easier to clean things up and fix it at the same time. reporter now calls some code directly in dns to avoid the "one gets updated without the other one" problem we keep having... Report and check are still seperate, but the *_available code is now the same. Reporter also does helper_mode now (which fixed the tainted path issue). I've tested against razor and dcc and they seem to work fine. I don't have pyzor installed, but since dcc works I don't see why pyzor wouldn't ... 2003-02-06 21:35 jmason * rules/70_cvs_rules_under_test.cf: some more new rules for when CVS reopens 2003-02-06 18:44 felicity * masses/mass-check: nullified --bayes option (still accepted, just doesn't do anything) 2003-02-06 16:43 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Doh. Like an idiot I put a good fix in, tested it, then changed the code and submitted (decided to do $set&2 instead of $set<2) and didn't pay attention. So now I reverted to the code that did work, tested it, and am now committing it. Somebody please slap me around for doing this ... 2003-02-06 16:25 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Fixes bug 1451. If Bayes isn't enabled, PerMsgStatus::check() will now look to see if we can enable it (ie: change from scoreset {0,1} => {2,3}) and do it if possible. 2003-02-05 19:21 jmason * rules/70_cvs_rules_under_test.cf: some rules noted for uncommenting after release 2003-02-04 20:23 climent * rules/25_body_tests_es.cf: vim tweak 2003-02-04 11:43 jmason * rules/60_whitelist.cf: bug 1348: some default whitelists 2003-02-04 10:43 msquadrat * lib/Mail/SpamAssassin/Dns.pm: Backport workaround for Razor2 disabling the timeout (patch 1.67 -> 1.68) 2003-02-04 10:21 msquadrat * spamd/libspamc.c: Fixed Bug 1442: Segfault when connection to spamd fails 2003-02-04 06:23 felicity * masses/craig-evolve.c: craig-evolve code used to generate sets 0 and 1 for 2.50 applies allen's patch to give score limits 2003-02-04 05:42 felicity * rules/50_scores.cf: USER_AGENT_MOZILLA_XM was autodisabled by the GA but is a dependency of USER_AGENT meta rule re-enabled with -0.001 score 2003-02-04 03:05 felicity * rules/50_scores.cf: GA generated scores for sets 0 and 1 2003-02-04 02:09 felicity * rules/20_phrases.cf: Oops, some of my tflags made it in here somehow ... :( 2003-02-04 02:05 felicity * rules/: 20_head_tests.cf, 20_phrases.cf, 20_ratware.cf: Updated to remove rules that went in broken rules 2003-02-03 19:50 felicity * lib/Mail/SpamAssassin/NoMailAudit.pm: fixed bug 1439: blank header would cause rewrite_as_spam to see next header in message... 2003-02-03 16:59 duncf * debian/: changelog.in, control, rules, spamassassin.config: Bump version number. Misc fixes 2003-02-03 14:22 felicity * rules/50_scores.cf: New set0 scores w/out BODY_8BITS 2003-02-03 14:15 jmason * Makefile.PL: fixed 'spamc not being rebuilt after changes' dependency problem for DuncF 2003-02-03 12:54 jmason * lib/Mail/SpamAssassin/PerMsgStatus.pm: always add X-Spam-Checker-Version header for all mails (bug 1342) 2003-02-03 12:42 jmason * lib/Mail/SpamAssassin.pm: bug 1392: languages_filename cannot be set if using config_text. now fixed 2003-02-03 12:37 jmason * lib/Mail/SpamAssassin.pm: added main->{username} attribute so SA code can usefully figure out the current username 2003-02-03 07:23 quinlan * lib/Mail/SpamAssassin/: UnixLocker.pm, Util.pm, Win32Locker.pm: replace Win32 mkdir-based locking with sysopen-based locking add some debugging, comments, and warnings to locking code 2003-02-03 07:22 quinlan * rules/20_body_tests.cf: fix bug 1428: BODY_8BITS: will hit lots of Asian ham 2003-02-03 07:22 quinlan * lib/Mail/SpamAssassin/PerMsgStatus.pm: fix bug 1371: some tests don't work on CRLF line ends 2003-02-03 07:21 quinlan * lib/Mail/SpamAssassin/EvalTests.pm: fix bug 1371: some tests don't work on CRLF line ends fix bug 1428: BODY_8BITS: will hit lots of Asian ham 2003-02-03 07:15 quinlan * lib/Mail/SpamAssassin/Dns.pm: make a dbg() statement a bit clearer 2003-02-03 04:35 felicity * spamassassin.spec, spamd/redhat-rc-script.sh: add support for "condrestart" for redhat installs 2003-02-02 19:29 jmason * masses/corpora/mass-find-nonspam: bug 1432: mass-find-nonspam was broken. fixed 2003-02-02 19:21 jmason * sa-forget.raw, sa-learn-nonspam.raw, sa-learn-spam.raw, lib/Mail/SpamAssassin/CmdLearn.pm: bug 1433: added --local switch to learner UIs, for future use 2003-02-02 19:07 felicity * masses/score-ranges-from-freqs, rules/50_scores.cf: Set 0 scores from GA run My version of score-ranges for said scores 2003-02-02 15:30 jmason * spamd/libspamc.c: removed connect() debugging from libspamc, bug seems fixed now 2003-02-02 00:09 msquadrat * lib/Mail/SpamAssassin/Conf.pm: A more sophisticated language handling 2003-02-01 20:01 felicity * masses/logs-to-c: fixed bug where range_lo == range_hi means score==1 ... doh! 2003-02-01 17:06 jmason * sa-learn-nonspam.raw, lib/Mail/SpamAssassin/ArchiveIterator.pm, masses/mass-check: bug 1391: globbing in mass-check (and sa-learn-whatever) 2003-02-01 16:57 jmason * spamd/spamd.raw: bug 773: more flexible --virtual-config switch, --virtual-config-dir. bit more extension 2003-02-01 16:52 jmason * MANIFEST, spamd/spamd.raw: bug 773: more flexible --virtual-config switch, --virtual-config-dir 2003-02-01 16:21 jmason * spamd/libspamc.c, spamd/libspamc.h, spamd/spamc.c, spamd/spamc.pod, spamd/spamd.raw, t/SATest.pm, t/spamd_report.t, t/spamd_report_ifspam.t, t/spamd_symbols.t: added support for REPORT, REPORT_IFSPAM, and SYMBOLS methods in spamc: bug 1393, bug 191. Thanks to Eugene Miretsky (eugene at invision dot net) for the patch. 2003-02-01 15:45 felicity * lib/Mail/SpamAssassin/Conf.pm: Updated docs for auto-training a little bit 2003-01-31 22:57 jmason * rules/70_cvs_rules_under_test.cf: added commented T_MSGID_SPAMSIGN_99X9XX99 for later use 2003-01-31 22:40 jmason * lib/Mail/SpamAssassin/BayesStore.pm, lib/Mail/SpamAssassin/Conf.pm, spamd/spamd.raw: verified that per-user bayes dbs work with spamd, fixed some bugs and some doco bugs as a result. 2003-01-31 22:11 jmason * lib/Mail/SpamAssassin/Conf.pm: updated 'meta' doco 2003-01-31 22:07 jmason * lib/Mail/SpamAssassin/Conf.pm: update bayes_path doco 2003-01-31 15:35 jmason * lib/Mail/SpamAssassin.pm: 2.44 RELEASED. starting hopefully-not-required 2.45 cycle 2003-01-31 15:14 jmason * Changes, build/update_devel, build/update_stable: preparing build scripts for 2.44 release 2003-01-30 17:45 felicity * lib/Mail/SpamAssassin.pm: Now we can either comment out or set to 0 for release. Hopefully stop the confusion that keeps happening. :) 2003-01-30 17:40 felicity * masses/logs-to-c: [no log message] 2003-01-30 16:15 msquadrat * lib/Mail/SpamAssassin.pm: A more obvious way to disable the DEVEL flag ;-) 2003-01-30 15:41 felicity * rules/50_scores.cf: Reset rules to pre-2.50 GA run. 2003-01-30 15:36 felicity * rules/24_auto_meta.cf: Per discussion with other developers, auto_meta rules will wait for 2.60. 2003-01-30 13:09 msquadrat * lib/Mail/SpamAssassin.pm: Prepare for release 2003-01-30 13:02 msquadrat * spamd/spamd.raw: Backport workaround from Bug 526: Failed sanity check because of clobbered STDOUT (patch 1.147 -> 1.148) 2003-01-30 12:55 msquadrat * spamd/libspamc.c: Backport fix for Bug 1306: Possible buffer overflow in libspamc when running in BSMTP mode (patch 1.15 -> 1.18) 2003-01-30 12:29 msquadrat * spamd/libspamc.c: Backport fix for Debian Bug 160206: Insufficient buffer in libspamc (patch 1.8 -> 1.9) 2003-01-30 12:23 msquadrat * lib/Mail/SpamAssassin.pm: Backport fix for warinings in sed_path (patch 1.141 -> 1.142) 2003-01-30 12:20 msquadrat * lib/Mail/SpamAssassin/NoMailAudit.pm: Backport fix for Bug 1127: Existing lowercase x-spam-status header kills SpamAssassin (patch 1.40 -> 1.41) 2003-01-29 21:08 felicity * rules/50_scores.cf: Made sure that rules used by meta rules are not disabled 2003-01-29 20:08 felicity * rules/: 24_auto_meta.cf, 50_scores.cf: Updated score sets 0 and 1 per Allen's results. Added 24_auto_meta for the auto generated meta rules. 2003-01-29 17:56 jmason * lib/Mail/SpamAssassin/: BayesStore.pm, CmdLearn.pm: bug 1326: sa-learn-rebuild now shows dots 2003-01-29 12:12 jmason * lib/Mail/SpamAssassin/EvalTests.pm: bug 1427: SUBJ_FULL_OF_8BITS should use raw, non-decoded header 2003-01-28 23:46 jmason * lib/Mail/SpamAssassin/: Conf.pm, ConfSourceSQL.pm: added patch from Michael 'Moose' Dinn, dinn at twistedpair.ca, to make SQL preferences work as part of Horde's SQL preferences. 2003-01-28 23:37 jmason * lib/Mail/SpamAssassin/Conf.pm: removed obsolete unnamed-rule support, which was never used 2003-01-27 21:32 jmason * wordfreqs/README: note for wordfreqs README 2003-01-27 21:31 jmason * MANIFEST: Malte, you forgot the spamc test's data file ;) 2003-01-27 21:12 jmason * MANIFEST, lib/Mail/SpamAssassin.pm, lib/Mail/SpamAssassin/BayesStore.pm, lib/Mail/SpamAssassin/DBBasedAddrList.pm, lib/Mail/SpamAssassin/Locker.pm, lib/Mail/SpamAssassin/UnixLocker.pm, lib/Mail/SpamAssassin/Util.pm, lib/Mail/SpamAssassin/Win32Locker.pm: abstracted UNIX-specific locking code into new UnixLocker class; implemented Dan's win32 code as Win32Locker, loaded on win32 only: bug 1246 2003-01-27 20:19 jmason * spamd/libspamc.h: bug 1330: spamc license fixed 2003-01-27 18:46 jmason * lib/Mail/SpamAssassin/Conf.pm: finally did away with all that [-_] crap 2003-01-27 18:30 jmason * lib/Mail/SpamAssassin/: Conf.pm, Dns.pm: pyzor_options conf item added, thanks to sabat 2003-01-27 17:56 jmason * lib/Mail/SpamAssassin/Bayes.pm: oops, spotted typo in IGNORED_HDRS 2003-01-27 10:21 msquadrat * t/: SATest.pm, spamc.t: Fix some warnings in spamc test 2003-01-26 17:02 jmason * lib/Mail/SpamAssassin/PerMsgStatus.pm: bug 1398: cannot use $-[0] on perl 5.005 2003-01-26 16:54 jmason * spamd/: libspamc.c, spamc.c, utils.c: bug 1330: spamc license fixup, spamc and libspamc can now legally be used outside SpamAssassin 2003-01-26 16:49 msquadrat * t/: spamc.t, data/etc/hello.txt: Added spamc test to differentiate between spamc and spamd errors 2003-01-26 16:45 jmason * spamd/: libspamc.c, spamc.c, spamc.pod: added debug instrumentation to spamc to track down spamc coredumps 2003-01-25 23:28 msquadrat * lib/Mail/SpamAssassin/Dns.pm: Remove a spurious dot from the DCC command line 2003-01-25 22:55 msquadrat * lib/Mail/SpamAssassin.pm: Fix some _really_ weird and buggy code 2003-01-25 22:16 msquadrat * lib/Mail/SpamAssassin/Dns.pm: Some more pipe fixing (bug 1410) 2003-01-25 04:15 felicity * rules/50_scores.cf: reset scores to first set0 run 2003-01-25 02:43 felicity * rules/70_broken_rules.cf: these rules have both low hits and more ham hits than spam hits ... 2003-01-24 16:38 felicity * spamassassin.raw: warn about -S in spamassassin too 2003-01-24 16:37 felicity * spamd/spamd.raw: we should only warn people that -S was removed, don't die yet 2003-01-24 15:57 msquadrat * INSTALL: Added lock file to procmail "tutorial" 2003-01-24 15:01 felicity * lib/Mail/SpamAssassin/Conf.pm: corrected output statement 2003-01-24 12:16 msquadrat * Makefile.PL: Just empty (don't remove and recreate) PKG_LOCAL_RULES_DIR on every install 2003-01-24 12:01 msquadrat * lib/Mail/SpamAssassin/Dns.pm: Do what I mean not what I think. 2003-01-24 11:01 msquadrat * lib/Mail/SpamAssassin/Dns.pm: Fix some unchecked broken pipes for Pyzor & DCC 2003-01-23 13:17 jmason * Makefile.PL, masses/logs-to-c: spamd now not built on Windows, since it doesn't. ;) suggestion from Roland Bauer, roland.bauer AT fff.at 2003-01-23 10:56 jmason * lib/Mail/SpamAssassin/BayesStore.pm: avoid 'argument '' isn't numeric in addition' warning 2003-01-23 04:12 felicity * rules/50_scores.cf: new set1 scores 2003-01-23 00:58 jmason * spamd/: libspamc.c, libspamc.h, spamc.c, utils.c, utils.h: bug 1212: spamc -t timeout code from eugene AT invision.net 2003-01-22 15:06 jmason * spamd/libspamc.c: removed debugs, left 1 in at LOG_DEBUG level 2003-01-22 14:50 felicity * rules/50_scores.cf: New set1 GA run completed 2003-01-22 14:35 jmason * spamd/libspamc.c: reworked the failover-fix code a little to avoid coredump reported by Malte 2003-01-21 22:15 felicity * lib/Mail/SpamAssassin/: Conf.pm, PerMsgStatus.pm: Fix for bug 1400. If a rule was defined with no score, only a single scoreset would get a score set for it. We now set a score for all scoresets unless one is currently defined. Changed the autolearn code back to not checking for existence. 2003-01-21 21:48 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: For some stupid reason, scoresets may have different rules from each other, so we need to check for existence. :( 2003-01-21 21:02 felicity * lib/Mail/SpamAssassin/Bayes.pm: give debug info if we can't use a bayes db due to insufficient messages 2003-01-21 19:38 felicity * lib/Mail/SpamAssassin/Conf.pm: always_add_report should be 0, not 1 by default 2003-01-21 19:36 felicity * rules/23_bayes.cf: a bunch of BAYES rules didn't have tflags learn ... 2003-01-21 18:44 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: let autolearn ignore userconf too 2003-01-21 18:08 felicity * lib/Mail/SpamAssassin/PerMsgStatus.pm: Code skips 0 score sets in autolearning We now ignore tflag "learn" rules in autolearning instead of /BAYES/ 2003-01-21 17:57 jmason * lib/Mail/SpamAssassin/ArchiveIterator.pm: leftover crud. sorry about that 2003-01-21 17:33 jmason * sa-learn-nonspam.raw, lib/Mail/SpamAssassin/ArchiveIterator.pm, lib/Mail/SpamAssassin/Conf.pm, lib/Mail/SpamAssassin/PerMsgStatus.pm, masses/corpora/mass-find-nonspam: added always_add_report param, patch from sabat (sabat /at/ eloan /dot/ com) 2003-01-21 16:09 felicity * lib/Mail/SpamAssassin/: Conf.pm, PerMsgStatus.pm: autolearning will now only take place with scoresets 0 and 1. if we're using 2 or 3, the autolearn code will revert to 0 or 1 and recompute the score, then check about autolearning. this gives us consistency in autolearning on the same scoresets all the time. 2003-01-21 15:13 felicity * rules/50_scores.cf: Updated scores after set1 run 2003-01-21 01:55 felicity * rules/20_head_tests.cf: whitelists are supposed to be nice! 2003-01-21 01:37 felicity * rules/50_scores.cf: setup for set1 GA run 2003-01-21 01:36 felicity * masses/corpora/remove-tests-from-logs, rules/20_head_tests.cf: fixing some stuff to remove tests, X_OSIRU_DUL_FH is supposed to be nice 2003-01-20 19:54 felicity * masses/: hit-frequencies, lint-rules-from-freqs: More work to make net tests mutable if we're doing a net test (scoreset%2 == 1) 2003-01-20 19:40 felicity * rules/50_scores.cf: GA set0 scores > 4.8 brought down to 4.0 (except those that are very unlikely to FP) 2003-01-20 19:28 felicity * rules/50_scores.cf: GA scored set 0 2003-01-20 19:05 jmason * spamd/libspamc.c: struct hostent was overwritten in syslog call so failover was not working; now fixed, I think 2003-01-20 18:52 felicity * masses/Makefile: Makefile now knows about scoresets. Do a "make SCORESET=#" or it'll default to 0. 2003-01-20 18:46 felicity * masses/logs-to-c: put back includes, lets evolve compile on macosx. :( 2003-01-20 16:50 felicity * masses/score-ranges-from-freqs: make net tests mutable if scoreset indicates a net run 2003-01-20 15:49 felicity * lib/Mail/SpamAssassin/EvalTests.pm, masses/hit-frequencies, masses/logs-to-c, masses/parse-rules-for-masses, masses/score-ranges-from-freqs: Comment out CARRIAGE_RETURN eval code. masses code now understands scoresets, at least enough. :) 2003-01-20 01:47 felicity * rules/: 20_body_tests.cf, 30_text_fr.cf, 50_scores.cf, 70_broken_rules.cf: got rid of CARRIAGE_RETURNS 2003-01-19 20:23 felicity * masses/corpora/remove-tests-from-logs: fixed a few more bugs 2003-01-19 19:50 felicity * masses/corpora/remove-tests-from-logs: fixed the code some more to deal with -t option 2003-01-19 17:28 felicity * rules/50_scores.cf: boo boo with HABEAS_SWE 2003-01-19 05:59 felicity * masses/corpora/remove-tests-from-logs: take a -t option to specify a timestamp. rule removal will only happen up to that timestamp. this is useful for net tests where you only want to keep X months. 2003-01-19 05:20 felicity * masses/mass-check: total_check wasn't being incremented before progress, so at 100% done, we were one off. :( 2003-01-19 04:25 felicity * lib/Mail/SpamAssassin/: BayesStore.pm, Conf.pm, DBBasedAddrList.pm, NoMailAudit.pm, PerMsgStatus.pm, Util.pm: Put umask around any open or tie commands. This will 1) let the *_mode options work as expected, and 2) keep some of our temp files more secure. 2003-01-19 00:56 felicity * masses/score-ranges-from-freqs: it didn't know how to deal with multiple scoresets, so I let it default to set0 2003-01-19 00:41 felicity * rules/: 20_compensate.cf, 20_head_tests.cf, 20_phrases.cf, 20_ratware.cf, 20_uri_tests.cf, 50_scores.cf, 70_broken_rules.cf: pushed rules to 70_broken from the first 2.50 mass-check with bad results. 2003-01-18 22:40 felicity * rules/20_compensate.cf: [no log message] 2003-01-18 20:58 felicity * rules/: 20_meta_tests.cf, 20_phrases.cf: NMS_CGI_NOT_BUGGY wasn't hitting anything, so I made it part of a meta rule with BUGGY_CGI per suggestion from eallen. 2003-01-18 16:51 felicity * masses/: craig-evolve.c, logs-to-c: Applied Allen's GA patches from bug 1088. 2003-01-18 11:56 msquadrat * spamd/spamd.raw: Removed small logging glitch (more code cleanup coming for 2.60) 2003-01-18 06:19 felicity * masses/hit-frequencies: hit-frequencies doesn't know what to do with multiple scoresets, so I let it default to scoreset 0... 2003-01-18 06:08 felicity * rules/50_scores.cf: setting pre-GA scores for bayes and net tests 2003-01-17 01:45 felicity * masses/mass-check: $date is passed undef to progress, so I added in a "set to 0 if undefined" to at least stop the warning. $total_count is set to 0 in the parent, then each child increases their copy, then the parent calls progress which still sees $total_count==0. So I moved the increment into the parent. 2003-01-16 20:28 felicity * lib/Mail/SpamAssassin/Dns.pm: added a debug statement in case the object format changes incompatibly again 2003-01-16 20:22 felicity * lib/Mail/SpamAssassin/: Dns.pm, EvalTests.pm: Fixed up Razor2 check code some more. We now support both razor object layouts (<2.14 and >=2.14). Got rid of extra variable in Razor2 code. Fixed comment for SEMIFORGED_HOTMAIL_RCVD. 2003-01-16 16:10 felicity * masses/mass-check: mass-check now shows status every $updates messages and at 100% 2003-01-16 14:42 felicity * masses/mass-check: fixed status displays to not give "illegal modulus 0" errors on small checks. 2003-01-16 12:06 msquadrat * masses/mass-check: Some tweaks to the --mid code (again): Make the mid meta RE safe and remove the time part (we don't need to create unique MIDs, this is just for logging) 2003-01-16 08:07 quinlan * lib/Mail/SpamAssassin/ArchiveIterator.pm: remove comment 2003-01-16 07:55 quinlan * lib/Mail/SpamAssassin/ArchiveIterator.pm, masses/mass-check: fix mass-check progress indicator fix mid so it will be consistent every run and use "." for whitespace so regexp search for the Message-ID: will work 2003-01-16 07:04 climent * rules/25_body_tests_es.cf: New entry 2003-01-15 20:43 jmason * masses/mass-check: mass-check eye candy ;) -- now reports progress and timestamps to STDERR if --showdots is off. 2003-01-15 19:59 jmason * lib/Mail/SpamAssassin/Conf.pm: [no log message] 2003-01-15 19:56 jmason * lib/Mail/SpamAssassin/ArchiveIterator.pm, masses/corpora/mass-find-nonspam: minor tweak: differentiate between MH and cyrus maildir formats 2003-01-15 19:47 msquadrat * spamd/spamd.raw: Fixing some possible undef errors related to logging 2003-01-15 19:31 quinlan * masses/hit-frequencies: add -t option for looking at certain tflags 2003-01-15 19:09 felicity * lib/Mail/SpamAssassin/Dns.pm: Fixed 1386, if people use engines other than the default (1,3,4), the response array is different than expected. 2003-01-15 16:56 felicity * masses/rewrite-cf-with-new-scores: rewrite-cf-with-new-scores now supports score lines with multiple scoresets. bug 1384. 2003-01-15 15:48 msquadrat * spamd/spamd.raw: Save a variable 2003-01-15 15:36 felicity * rules/50_scores.cf: fixed small typo 2003-01-15 15:30 msquadrat * spamd/spamd.raw: Fixed logging issues 2003-01-15 14:42 msquadrat * masses/mass-check: Fix a cute typo while I'm here 2003-01-15 14:34 msquadrat * masses/mass-check: Remove path from generated MIDs in --mid mode 2003-01-15 07:20 quinlan * masses/freqdiff: update for new mass-check format 2003-01-14 23:03 felicity * lib/Mail/SpamAssassin.pm: Added documentation about options to Reporter. 2003-01-14 22:55 jmason * lib/Mail/: SpamAssassin.pm, SpamAssassin/PerMsgStatus.pm: fixed bug 1356: spamd tests all failed if run as root 2003-01-14 22:45 felicity * lib/Mail/SpamAssassin/Dns.pm: use_* was looking at the wrong place ($self->{main}, not $self->{conf}) 2003-01-14 20:48 jmason * t/SATest.pm: dump spamd log in case it fails to start 2003-01-14 20:43 jmason * lib/Mail/SpamAssassin/Conf.pm: more config item rearranging to allow more settings to be set in user_prefs 2003-01-14 20:38 jmason * sa-learn-nonspam.raw, lib/Mail/SpamAssassin/PerMsgStatus.pm: fixed 2 pod bugs 2003-01-14 20:36 jmason * masses/mass-check: re-added remove_spamassassin_markup() code, working this time 2003-01-14 19:43 felicity * lib/Mail/SpamAssassin/Reporter.pm: don't need temp copy of fulltext in report, it's already a copy 2003-01-14 19:36 jmason * lib/Mail/SpamAssassin/AutoWhitelist.pm, lib/Mail/SpamAssassin/Bayes.pm, lib/Mail/SpamAssassin/Conf.pm, lib/Mail/SpamAssassin/DBBasedAddrList.pm, lib/Mail/SpamAssassin/Dns.pm, lib/Mail/SpamAssassin/PerMsgStatus.pm, t/razor2.t: bug 1354: add-addr-to-foolist broke after we removed support for IPless entries in the AWL, now fixed again. 2003-01-14 17:49 felicity * lib/Mail/SpamAssassin/: Dns.pm, Reporter.pm: standardize our razor text, use a copy of fulltext to avoid munging 2003-01-14 16:49 felicity * lib/Mail/SpamAssassin/Dns.pm: Fix what we pass to Razor, and don't fsck up the $$fulltext string 2003-01-14 15:56 felicity * lib/Mail/SpamAssassin/EvalTests.pm: fixes a problem where parsing an IP out with '[' (or any other odd RE-ism) causes the check to die. 2003-01-14 14:19 felicity * lib/Mail/SpamAssassin/Dns.pm: fixed typo in comment 2003-01-14 14:17 jmason * masses/hit-frequencies: oops 2003-01-14 14:16 jmason * masses/: Makefile, hit-frequencies: allow hit-freqs to display net test results. surely someone did this before? can't find it... 2003-01-13 21:26 quinlan * masses/mass-check: chomp $host while we're at it 2003-01-13 21:13 quinlan * masses/mass-check: privacy concern: add --mid option so Message-ID is not reported unless requested 2003-01-13 21:06 quinlan * masses/mass-check: add message-id to last field, make people happy again 2003-01-13 16:48 quinlan * lib/Mail/SpamAssassin/BayesStore.pm: changing "breaking lock" debug message to "unlocking lock" as it should be 2003-01-13 16:09 quinlan * masses/: logs-to-c, post-ga-analysis.pl: allow for time=