summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Fix for libevent.0.4.0Vsevolod Stakhov2011-07-221-0/+4
|
* Write error about too few tokens during learning.Vsevolod Stakhov2011-07-221-0/+5
|
* Increase buffer for output.Vsevolod Stakhov2011-07-221-1/+1
|
* Added tag 0.4.0 for changeset c52f190b0592Vsevolod Stakhov2011-07-221-0/+1
|
* Add workaround for clang under linux.Vsevolod Stakhov2011-07-2117-37/+45
| | | | Fix problems found by static analyzing.
* Another fix to avoid race with settings - add reference counter.Vsevolod Stakhov2011-07-212-22/+50
|
* * Add start script for red hat compatible systemsVsevolod Stakhov2011-07-2025-57/+656
| | | | | | | Add descriptions for some rspamd API functions (no functional changes). --HG-- rename : linux/rspamd => linux/rspamd_debian.in
* Rework http chunked encoding parsing.Vsevolod Stakhov2011-07-205-97/+115
|
* Do not try to use information about dispatcher after callback fails (found ↵Vsevolod Stakhov2011-07-191-1/+0
| | | | by valgrind).
* * Add classifiers pre-selection scriptVsevolod Stakhov2011-07-192-0/+52
|
* Fix coredumps on some specific messages with specific urls.Vsevolod Stakhov2011-07-193-11/+24
| | | | | Fix coredumps while closing log file. Fix parsing of chunked HTTP replies.
* Ref hash table at settings loading.Vsevolod Stakhov2011-07-181-1/+1
|
* Fix statfiles class determination euristic.Vsevolod Stakhov2011-07-182-12/+12
| | | | Fix call of classifier pre-callback.
* Fix textpart:get_language lua function.Vsevolod Stakhov2011-07-181-0/+1
|
* Create statfiles using learn_spam function for bayes classifier.Vsevolod Stakhov2011-07-182-1/+19
| | | | Fix call of pre callbacks for a classifier.
* Use event_set correctly after event_del.Vsevolod Stakhov2011-07-181-0/+1
|
* Fix bug with data corruption during settings application.Vsevolod Stakhov2011-07-181-1/+8
|
* Begin to write normal and updated default configuration.Vsevolod Stakhov2011-07-152-20/+30
| | | | | --HG-- rename : rspamd.xml.sample => conf/rspamd-basic.xml.in
* Another fix for comparing parts without content - two empty parts are equal.Vsevolod Stakhov2011-07-141-1/+6
|
* Fix coredump on messages with one url only.Vsevolod Stakhov2011-07-141-1/+1
|
* Fix core dumps when no symbols are found in a message.Vsevolod Stakhov2011-07-141-1/+1
|
* * Add learn_spam/learn_ham interface to librspamdclient and to rspamcVsevolod Stakhov2011-07-149-47/+332
| | | | | * Improve logic of io dispatcher restoration Remove correction factor from bayes as it leads to classify errors.
* Adjust interval.Vsevolod Stakhov2011-07-141-2/+2
|
* * Remove completion logic from controller: it is hardly used but breaks new ↵Vsevolod Stakhov2011-07-141-35/+26
| | | | commands logic
* Fix multiply compare_parts_distance calls.Vsevolod Stakhov2011-07-141-2/+12
|
* Ignore arguments order in compare_parts_distance function.Vsevolod Stakhov2011-07-141-2/+2
|
* Change logic of params inside compare parts distance.Vsevolod Stakhov2011-07-142-11/+51
| | | | | During learning and classifying compare parts using new algorithm. Raise similarity factor.
* * Add new algorithm based on diff algorithm to compare relatively short text ↵Vsevolod Stakhov2011-07-1314-11/+498
| | | | parts
* Add validity detector for statfiles inside classifier.Vsevolod Stakhov2011-07-133-2/+59
| | | | Add euristic to detect spam/ham classes based on statfile symbol.
* * Add second argument to compare_parts_distance function so it can be used ↵Vsevolod Stakhov2011-07-131-21/+37
| | | | as interval: arg2 <= distance <= arg1
* * Add ability to get difference between two parts from lua codeVsevolod Stakhov2011-07-122-0/+50
|
* * First commit to implement multi-statfile filter system with new learning ↵Vsevolod Stakhov2011-07-1217-82/+649
| | | | mechanizm (untested yet)
* Cache data of parts distance function to speed up multiply rules with such ↵Vsevolod Stakhov2011-07-122-1/+24
| | | | function.
* * Make fuzzy hashes utf8 compatible.Vsevolod Stakhov2011-07-122-35/+75
|
* * Add a simple logic of language detection for text parts (unicode script based)Vsevolod Stakhov2011-07-114-2/+118
|
* Fix phishing detection with img flag.Vsevolod Stakhov2011-07-114-32/+61
| | | | | | Handle unclosed HTML tags properly. Remove warnings for types on 32 bit archs. Do not touch grow factor many times when one shot mode is turned on.
* * Improve performance of settings lookupVsevolod Stakhov2011-06-306-106/+122
|
* * Add correcting factor to statistics.Vsevolod Stakhov2011-06-286-36/+105
| | | | | | Now learning increments version of a statfile. Avoid learning and classifying of similar text parts if a message has 2 text parts. Several fixes to statistics.
* * Add ability to specify noip option for uribl suffix to avoid checking urls ↵Vsevolod Stakhov2011-06-282-3/+29
| | | | with ip addresses on such lists.
* Fix statshow utility.Vsevolod Stakhov2011-06-272-4/+5
|
* Remove debug.Vsevolod Stakhov2011-06-271-1/+0
|
* Fix incorrect calculating of token length.Vsevolod Stakhov2011-06-272-2/+3
|
* * Welcome 0.4.0Vsevolod Stakhov2011-06-2421-184/+276
| | | | | | | | | | | | | | | | | | Uncompatible changes: - Statistics is uncompatible in utf8 mode Major changes: - Improved utf8 mode - Convert all characters to lowercase in statistics - Skip URL's in statistics - Improve speed of bayes classifier by using integer arithmetics - Fixed statfiles synchronization that was broken for a long time - Synchronization is now configurable Minor changes: - Bugfixes - Removed some of legacy code - Types polishing
* Oops, remove debug.Vsevolod Stakhov2011-06-231-7/+0
|
* * Fixes to fuzzy hashing logic, skip urls while estimating fuzzy hashVsevolod Stakhov2011-06-2312-69/+213
| | | | | Fix tags stripping. Fix phishing checks (ignore img tags).
* Another fix with reload command.Vsevolod Stakhov2011-06-201-3/+4
|
* Fix reload command.Vsevolod Stakhov2011-06-202-1/+2
|
* Handle files with zero lenght properly.Vsevolod Stakhov2011-06-172-0/+22
| | | | Reported by: Andrej Zverev
* Fix rspamc client to handle multiply files properly.Vsevolod Stakhov2011-06-171-8/+9
|
* Actually all times are in GMT already, so avoid conversion to prevent dst ↵Vsevolod Stakhov2011-06-152-3/+1
| | | | loosing.