summaryrefslogtreecommitdiffstats
path: root/src/libserver/re_cache.c
Commit message (Collapse)AuthorAgeFilesLines
* [Fix] Fix extra hits in PCRE mode for regular expressionsVsevolod Stakhov2018-01-241-3/+2
| | | | | MFH: rspamd-1.6 Issue: #1880
* [Minor] Fix addition in pcre fallbackVsevolod Stakhov2017-05-201-1/+2
|
* [Fix] Allow to follow symlinks when safeVsevolod Stakhov2017-05-091-1/+1
| | | | Issue: #1625
* [Fix] Further memory leaks fixesVsevolod Stakhov2017-01-181-0/+5
|
* [Rework|Minor] More updates for mime parsingVsevolod Stakhov2016-12-161-5/+5
|
* [Rework] Add preliminary implementation of the mime parserVsevolod Stakhov2016-12-131-1/+1
|
* [Fix] Fix re cache initialisationVsevolod Stakhov2016-11-251-1/+1
|
* [Minor] Fix hyperscan-less buildVsevolod Stakhov2016-11-171-1/+1
|
* [Minor] Improve debugging for re_cacheVsevolod Stakhov2016-11-171-26/+35
|
* [CritFix] Introduce raw content to text partsVsevolod Stakhov2016-10-311-2/+2
| | | | | | | | | | | | | | Previously we had the following types of text content: - `orig`: MIME *decoded* content - `content`: same as `orig` but converted to utf8 and without HTML tags - `stripped`: same as `content` but without newlines Now we add `raw`: - `raw`: raw mime content as it was in an original message This also fixes R_BAD_CTE_7BIT rule
* [Feature] Stop using of GLists for headers, improve performanceVsevolod Stakhov2016-09-281-4/+3
|
* [Feature] Add ffi friendly version of process_regexp functionVsevolod Stakhov2016-09-221-2/+23
|
* [CritFix] Fix hyperscan compilation on regexp changeVsevolod Stakhov2016-09-091-0/+10
| | | | | | | | | | | | | | Rspamd could use an incorrect hyperscan database in the case when a single regexp has been changed. In this case, Rspamd did not recalculate the cached files causing shifting of regexp IDs. Subsequently, that caused random regexp to match whilst completely different patterns were expected. With this change, Rspamd also takes care about the order of regexps (by including it in the crypto hash). This change eliminates the issue and Rspamd can deal with regexps changes correctly. MFH: true
* [CritFix] Check hyperscan cache sanity before loadingVsevolod Stakhov2016-08-091-2/+40
| | | | | | | | | | | | | | | | When regexp database is changed it is possible that hyperscan cache will contain some invalid or shifted regexps IDs. In this situation, Rspamd could not correctly detect that such an event occurred and loaded an incorrect database. This commit adds strict crc checking that includes: 1. HS ids 2. HS flags 3. HS serialized blob If this crc fails to match then Rspamd recompiles the corrupted cache file.
* [Fix] Forgot to commit leftover changesVsevolod Stakhov2016-07-291-1/+16
|
* [Fix] Ensure that hyperscan cache written is written properlyVsevolod Stakhov2016-07-291-1/+10
|
* [Fix] Don't mix hyperscan and pcre processing within a same taskVsevolod Stakhov2016-07-191-2/+7
|
* [Feature] Reduce number of timers queriesVsevolod Stakhov2016-07-191-6/+12
|
* [Rework] Rename mime parts structuresVsevolod Stakhov2016-07-041-1/+1
|
* [Minor] Add guards for header regexpsVsevolod Stakhov2016-05-241-2/+2
|
* [Minor] Use xxhash indirectlyVsevolod Stakhov2016-05-231-1/+2
|
* [Feature] New abstract hashing API in cryptoboxVsevolod Stakhov2016-05-101-7/+7
|
* [Minor] Use more portable constants for uint64_tVsevolod Stakhov2016-04-291-11/+12
|
* [CritFix] Properly support SA body regexpsVsevolod Stakhov2016-04-281-52/+75
|
* [CritFix] Fix SA rawbody processing - exclude top partVsevolod Stakhov2016-04-281-0/+9
|
* [Feature] Optimize rspamd_re_cache_type_from_stringVsevolod Stakhov2016-04-121-18/+34
|
* [Minor] Properly place timers calculationsVsevolod Stakhov2016-04-121-7/+7
|
* [Fix] Fix sabody rules in regexp cacheVsevolod Stakhov2016-04-111-14/+35
|
* [CritFix] Fix regexps results combinationVsevolod Stakhov2016-04-111-2/+7
| | | | | | When converting to vectored mode we need to remember results between consequent calls of regexp match engine. Prior to this patch this behaviour was broken and caused regexp rules to be matched incorrectly.
* [Feature] Implement SA compatible body regexpsVsevolod Stakhov2016-04-111-0/+46
|
* [Feature] Add support of mime headers expressionsVsevolod Stakhov2016-04-091-0/+46
|
* [Feature] Implement vectored mode for hyperscanVsevolod Stakhov2016-03-291-102/+202
| | | | | Rspamd can now use the fastest possible mode of hyperscan to allow further speed up in regexps matching.
* [Feature] Add more information about regexpsVsevolod Stakhov2016-03-111-0/+2
|
* Add more sanity checks when loading hyperscan cacheVsevolod Stakhov2016-02-181-8/+50
|
* Pcre2 compilation supportVsevolod Stakhov2016-02-091-5/+16
|
* Switch the rest to apache 2Vsevolod Stakhov2016-02-041-21/+12
|
* Add warning for slow regexpsVsevolod Stakhov2016-01-271-15/+30
|
* Rework hyperscan notifies to reduce noise levelVsevolod Stakhov2016-01-181-14/+34
|
* Fix warnings.Vsevolod Stakhov2016-01-151-4/+6
|
* Do not free HS database after deserializationVsevolod Stakhov2016-01-151-12/+14
|
* Fix a huge memory leak in hyperscanVsevolod Stakhov2016-01-151-0/+15
|
* Improve hyperscan compiler helper loggingVsevolod Stakhov2016-01-041-2/+2
|
* Fix multiple regexp in the cache for lua APIVsevolod Stakhov2015-12-261-2/+0
|
* Fix issue with non-unique regexps in the cacheVsevolod Stakhov2015-12-261-13/+23
|
* Fix issue with changing number of regexps in a cacheVsevolod Stakhov2015-12-241-4/+12
|
* Log on info level about HS prefiltersVsevolod Stakhov2015-12-241-1/+1
|
* Grr, really fix the loggingVsevolod Stakhov2015-12-241-12/+32
|
* Fix logging as headers now have \0 at the endVsevolod Stakhov2015-12-241-2/+2
|
* Minor fixesVsevolod Stakhov2015-12-221-0/+2
|
* Assume empty input as failed scanVsevolod Stakhov2015-12-141-5/+8
|
generated by cgit v1.2.3 (git 2.39.1) at 2025-03-31 00:20:17 +0000