summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
...
* [Minor] Add function to reschedule events at specific timerVsevolod Stakhov2024-04-232-3/+44
|
* [Minor] Deal with descriptors in the wait queueVsevolod Stakhov2024-04-221-2/+39
|
* [Minor] Do not use pointers to local variables, sighVsevolod Stakhov2024-04-221-8/+11
|
* [Minor] Try to fix mess with refcountsVsevolod Stakhov2024-04-221-9/+9
|
* [Rework] Serialize control commandsVsevolod Stakhov2024-04-222-68/+169
| | | | | | | In fact, we cannot send multiple commands and read them through the pipe. It has caused multiple weird issues in the past but I can now see clearly how it should be done. We should send commands and serialize all requests pending to let them being sent one by one, after reply for the previous command has been received.
* Merge pull request #4933 from rspamd/vstakhov-utf8-receivedVsevolod Stakhov2024-04-202-6/+20
|\ | | | | Support UTF8 flag for the received headers
| * [Minor] Support UTF8 in Received flagsVsevolod Stakhov2024-04-192-6/+20
| | | | | | | | Issue: #4919
* | Merge pull request #4932 from rspamd/vstakhov-fix-bayes-disableVsevolod Stakhov2024-04-201-0/+12
|\| | | | | Fix statfile symbols disabling
| * [Fix] If we have one statfile disabled we need to disable allVsevolod Stakhov2024-04-191-0/+12
| |
* | [Fix] Always set the unlearn flag when relearningaduernberger2024-04-181-1/+1
| | | | | | Treat relearning from ham to spam the same as relearning from spam to ham. `res` is only set to true if the mail is already learned.
* | [Enhancement] Exclude false positives of MIME_DOUBLE_BAD_EXTENSIONDmitriy Alekseev2024-04-111-1/+6
| | | | | | If we detected extension of the file and this extension is equal to the real extension then we should not pass second extension to check function as it will trigger false positive for MIME_DOUBLE_BAD_EXTENSION.
* | [Minor] Fix grow_factorAndrew Lewis2024-04-101-1/+1
| |
* | [Project] Adjust some final bitsVsevolod Stakhov2024-04-082-9/+17
| |
* | [Project] Write a function to adjust grow factor properlyVsevolod Stakhov2024-04-081-0/+42
| |
* | [Project] Rework grow factor as it is currently brokenVsevolod Stakhov2024-04-083-44/+10
|/
* [Fix] Honor dynamic thresholds for greylisting moduleVsevolod Stakhov2024-04-052-1/+42
|
* [Minor] Remove redundant metric nameVsevolod Stakhov2024-04-051-1/+1
|
* [Fix] Update greylisting threshold as it can be easily dynamicVsevolod Stakhov2024-04-051-7/+4
| | | | | Issue: #4908 Closes: #4908
* [Fix] Fix shared memory proxying when compression is setVsevolod Stakhov2024-04-031-68/+67
|
* [Feature] Allow custom milter quarantine and tempfail messagesVsevolod Stakhov2024-04-023-20/+73
| | | | | Issue: #4892 Closes: #4892
* [Minor] Fix format of example setting (rspamd.com#736)Andrew Lewis2024-03-251-1/+1
|
* [CritFix] Protect regexp matcher from regexps with empty patternsVsevolod Stakhov2024-03-223-1/+15
| | | | | Issue: #4885 Closes: #4885
* [Minor] Deny zero length in regexp search functionsVsevolod Stakhov2024-03-221-11/+5
|
* [Fix] Fix inconsistent nesting in mime partsVsevolod Stakhov2024-03-201-1/+8
| | | | | | | | | We should do like MUAs do: imlicitly close incomplete mime parts and treat them as ususal. It is weird and utterly stupid but that's what MUA do. Issue: #4740 Closes: #4740
* [Rework] Further types conversion (no functional changes)Vsevolod Stakhov2024-03-18241-8376/+8402
|
* [Rework] Remove some of the GLib types in lieu of standard onesVsevolod Stakhov2024-03-18118-861/+861
| | | | This types have constant conflicts with the system ones especially on OSX.
* [Minor] Another similar `getline` fixVsevolod Stakhov2024-03-151-2/+2
|
* [Minor] Fix irrelevant issue with the `getline` invocationVsevolod Stakhov2024-03-151-1/+1
|
* [Fix] Do not save multipatterns to FS in certain casesVsevolod Stakhov2024-03-157-29/+39
|
* [Fix] Fuzzy add/delete handlers are badly broken for compressionVsevolod Stakhov2024-03-141-2/+7
|
* [Minor] Use zstd compression by default in the clientVsevolod Stakhov2024-03-131-2/+2
| | | | No reasons why not to do it, as there are no supported versions without compression support
* [Feature] Use msgpack in the clientVsevolod Stakhov2024-03-121-4/+11
|
* [Minor] Fix invocationVsevolod Stakhov2024-03-113-3/+3
|
* [Feature] Support reply in message pack formatVsevolod Stakhov2024-03-115-14/+55
| | | | | | Rspamd currently sends reply in JSON format. However, message pack seems to be a better choice for a compatible client. It is faster and does not need escaping or even UTF8 validation. This is a surface for further updates.
* [Rework] Another rework of the dkim signing logicVsevolod Stakhov2024-03-091-21/+18
|
* [Rework] Allow multiple conditions for dkim signing when using vault/redisVsevolod Stakhov2024-03-091-1/+1
|
* [Minor] Add HTTP_MAGIC_SVG keymoisseev2024-03-071-2/+3
|
* [Minor] Serve SVG with the correct CTmoisseev2024-03-071-0/+1
|
* [Feature] Further optimization to the hot pathVsevolod Stakhov2024-03-061-6/+10
| | | | | | | | | | We check userdata very frequently, so the idea here is the following: - Store the address of classname (converted to int) in the metatable at index 1 - When we need to check some udata, we can just compare the static address with the integer stored in metatable - This avoid quite an expensive `lua_rawequal` call for two tables as we know that our classes are quite static
* [Rework] Rework dynamic usageVsevolod Stakhov2024-03-053-6/+112
|
* [Minor] Fix lua_tensor usageVsevolod Stakhov2024-03-052-11/+9
|
* [Rework] Use direct hashing instead of string hashingVsevolod Stakhov2024-03-052-11/+13
|
* [Project] Convert the remainingVsevolod Stakhov2024-03-0543-319/+413
|
* [Minor] Forgotten filesVsevolod Stakhov2024-03-052-0/+46
|
* [Project] Start optimization of userdata hashingVsevolod Stakhov2024-03-0518-57/+65
|
* [Fix] Update `nconns` in proxyVsevolod Stakhov2024-03-021-0/+4
|
* [Minor] Ask for utf8 string from libarchiveVsevolod Stakhov2024-02-281-1/+1
|
* [Minor] Detect encryption when libarchive is usedVsevolod Stakhov2024-02-281-0/+7
|
* [Project] Use libarchive for troublesome archivesVsevolod Stakhov2024-02-281-9/+40
|
* [Fix] Allow different template expansion parametersVsevolod Stakhov2024-02-211-3/+3
| | | | Probably one of the most dirty hacks I've used in the Lua code in Rspamd...