390 Commits

Author SHA1 Message Date
Jean-Francois Dockes
2594b71ae8 log 2017-01-16 11:14:54 +01:00
Jean-Francois Dockes
d80531fa62 Fix mimetype filtering (indexedmimetypes/excludedmimetypes) not working for embedded documents 2017-01-13 09:18:18 +01:00
Jean-Francois Dockes
3595109084 detect unicode BOM in text files 2016-11-15 18:31:34 +01:00
Jean-Francois Dockes
93c0001439 pretty 2016-11-08 12:42:46 +01:00
Jean-Francois Dockes
9ce6530e7b execm filters: the change to let filters set arbitrary metadata lost the top doc size, now saved aside 2016-08-12 18:00:52 +02:00
Jean-Francois Dockes
92da4c00cd use std c++11 initializer instead of create_xx hacks 2016-07-16 11:15:31 +02:00
Jean-Francois Dockes
c1fad4afc7 Replaced pthread with std:: thread and mutex 2016-07-12 18:08:21 +02:00
Jean-Francois Dockes
f6a999de84 logging now uses c++ streams 2016-07-12 09:41:04 +02:00
Jean-Francois Dockes
b9e672abda Allow execm input handlers to set arbitrary data fields 2016-07-11 18:13:39 +02:00
Jean-Francois Dockes
1aea57fcb2 defined data access interface for external indexers 2016-06-01 09:46:47 +02:00
Jean-Francois Dockes
627da5a39b Handler timeout should not interrupt the whole indexing pass 2016-04-14 15:48:01 +02:00
Jean-Francois Dockes
f3820471e4 Add cachedir variable allowing to move all data directories by setting a single value. Closes issue #270 2016-04-08 15:09:15 +02:00
Jean-Francois Dockes
a4fd4ee5be moved code around to make smallut and pathut less recoll-specific and reusable. No actual changes 2016-03-21 12:55:31 +01:00
Jean-Francois Dockes
08a810986c Lower log level for xattr op error with errno ENOTSUP 2016-02-23 08:03:17 +01:00
Jean-Francois Dockes
ff15f8fb1c Centralize stat calls to ensure consistency of time fields on windows 2016-01-08 11:23:10 +01:00
Jean-Francois Dockes
a95dcbd4b0 Windows: fix missing O_BINARY 2015-12-02 11:42:44 +01:00
Jean-Francois Dockes
a783ab17dc mh_execm: compute file md5 before activating filter to avoid concurrent open issues on Windows 2015-12-02 10:30:04 +01:00
Jean-Francois Dockes
5ba0be5e58 windows: mh_mbox reverted the test for From lines... 2015-12-01 17:29:44 +01:00
Jean-Francois Dockes
f70c92c629 rcldb::getSubDocs() (called from GUI show subdocs) was returning too many results because the parent/child ipath test was flawed 2015-11-03 08:40:13 +01:00
Jean-Francois Dockes
9a72fa5980 got rid of a number of ifdefs which should not be needed anymore 2015-10-08 14:48:50 +02:00
Jean-Francois Dockes
ffc7d04bdf fix timeouts for mh_execm filters 2015-10-08 10:10:04 +02:00
Jean-Francois Dockes
3b18facc16 Fixed some "unused xxx" warnings + include autoconfig 2015-10-07 08:30:49 +02:00
Jean-Francois Dockes
d942242047 replace all %lld instances 2015-10-03 17:25:17 +02:00
Jean-Francois Dockes
c2e8d451e4 Windows: the qt GUI builds with Qt/MinGW. No link. 2015-10-02 18:30:13 +02:00
Jean-Francois Dockes
031a2a0b4a Small filter fixes
--HG--
branch : WINDOWSPORT
2015-09-14 14:19:23 +02:00
Jean-Francois Dockes
86ef362461 rclimg (tweaks for perl)
--HG--
branch : WINDOWSPORT
2015-09-14 10:33:39 +02:00
Jean-Francois Dockes
24c77d2984 more filter conversion to python: svg and xml. Get rid of rclnull
--HG--
branch : WINDOWSPORT
2015-09-14 09:51:11 +02:00
Jean-Francois Dockes
82bdf21833 windows: arrange so that we can use simple python script names in mimeconf
--HG--
branch : WINDOWSPORT
2015-09-12 16:18:05 +02:00
Jean-Francois Dockes
330c7fc30d Python filters beginning to work, still issues.
--HG--
branch : WINDOWSPORT
2015-09-11 16:16:16 +02:00
Jean-Francois Dockes
06f43c573e 1st successful use of mh_exec (direct antiword exec)
--HG--
branch : WINDOWSPORT
2015-09-11 11:26:53 +02:00
Jean-Francois Dockes
1cbf02f713 Suppressed many integer size warnings by a mix of type adjustments and casts,
none of which should have a real effect.

--HG--
branch : WINDOWSPORT
2015-09-01 19:39:20 +02:00
Jean-Francois Dockes
2fe75dba28 More small windows int types fixes.
--HG--
branch : WINDOWSPORT
2015-09-01 15:03:21 +02:00
Jean-Francois Dockes
94da56aadc Handle windows paths: mostly change tests for relative, and some other tweaks.
--HG--
branch : WINDOWSPORT
2015-08-31 17:20:04 +02:00
Jean-Francois Dockes
3aeffea387 recollindex and recollq build and run but fail because of unac error
--HG--
branch : WINDOWSPORT
2015-08-31 13:08:50 +02:00
Jean-Francois Dockes
14c8e740d6 Windows: fixed a number of int size warnings mostly by casting them away
--HG--
branch : WINDOWSPORT
2015-08-30 17:30:31 +02:00
Jean-Francois Dockes
1d9047df1a fix linux build of windows branch
--HG--
branch : WINDOWSPORT
2015-08-30 15:50:17 +02:00
Jean-Francois Dockes
75517f7497 recollindex builds. Still need to implement quite a lot of ifndefed stuff (pathut, rclconfig)
--HG--
branch : WINDOWSPORT
2015-08-30 15:30:50 +02:00
Jean-Francois Dockes
d4cd1dd91c 1st mods to get a build under windows. Does not build yet, far from it
--HG--
branch : WINDOWSPORT
2015-08-30 11:19:18 +02:00
Jean-Francois Dockes
62f4f7bd81 more safexx posix stuff 2015-08-22 08:37:58 +02:00
Jean-Francois Dockes
c6e228b7c6 Prepared windows port by removing a number of spurious reference to unix-specific interfaces, and using some xapian posix adaptor includes 2015-08-19 14:41:10 +02:00
Jean-Francois Dockes
4713c3e488 Discard data for docs where the alternate transcode-from-8bit trial fails after the transcode from utf-8 has failed 2015-08-13 18:14:47 +02:00
Jean-Francois Dockes
4d1f679eac Use std[::tr1]::shared_ptr instead of local RefCntr by default 2015-08-09 13:54:24 +02:00
Jean-Francois Dockes
72da3ba2a9 merge 2015-08-06 08:26:39 +02:00
Jean-Francois Dockes
0840daf20e Avoid replacing (instead of concatenating) the current author field value with the internal one when the document is a top-level one. This allows metadata from metadatacmds to be used 2015-08-06 08:08:36 +02:00
Jean-Francois Dockes
1e2fdcd21d other ut/test drivers Makefiles 2015-08-05 09:01:11 +02:00
Jean-Francois Dockes
4d35cbabfb Also index non-html files from the web queue and fix the Open operation for them 2015-07-24 16:30:13 +02:00
Jean-Francois Dockes
e59bf51445 Impose memory usage limit on external filters. Fixes issue #259 2015-07-18 19:30:49 +02:00
Jean-Francois Dockes
906312519c Minimum checking that we have enough space before attempting decompression. Change config default compressed file from no limit to 20 MB 2015-04-22 17:08:06 +02:00
Jean-Francois Dockes
3cceffdb9c Use O_NOATIME to avoid disturbing st_atime when possible. Closes issue #230 2015-03-25 13:49:33 +01:00
Jean-Francois Dockes
9ba0b3e8bc Replaced RSA md5 code with public domain OpenBSD/debian dpkg version 2015-03-01 14:28:01 +01:00