summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* Update man pagesHEADv2.3.4masterReuben Thomas2023-02-193-3/+3
|
* Fix potential NULL dereferences (thanks, GCC!)Reuben Thomas2022-12-112-4/+7
| | | | | Also fix some English in the API for enchant_dict_suggest, the implicated function.
* lib.c: fix double-free (thanks, GCC)Reuben Thomas2022-12-081-0/+1
|
* Convert man pages to HTML for online documentation (do not distribute)Reuben Thomas2022-09-255-1/+489
|
* Improve man page markupReuben Thomas2022-09-252-11/+11
| | | | Use curly quotes, and indent code examples.
* Improve the documentationReuben Thomas2022-09-255-48/+97
| | | | | | | | These changes were inspired by issue #311. Move the documentation on personal wordlists from README into enchant.5, and the documentation on ordering files from enchant.1 to enchant.5. Add a cross reference from README to the man page.
* enchant++.h: make is_added() method return a result (fix #304)Reuben Thomas2022-04-161-3/+3
|
* Reduce dependency on gnulib by requiring C99 and POSIX 2008Dimitrij Mijoski2022-03-193-18/+15
| | | | | | | | | - Remove gnulib modules c99, getopt-posix, strdup-posix, ssize_t - Also remove obsolete gnulib module snippet/unused-parameter, use _GL_UNUSED instead of older _GL_UNUSED_PARAMETER, and instead of [[maybe_unused]]. Thanks to @dimztimz for providing the patch on which this commit is based.
* enchant.1: document format of language tags (see issue #295)Reuben Thomas2021-12-031-2/+3
|
* enchant-lsmod.c: show provider error if any when request_dict failsReuben Thomas2021-11-161-1/+5
| | | | | This should prevent confusion such as that which led to https://github.com/AbiWord/enchant/issues/289
* Add debug.h for debuggingReuben Thomas2021-11-163-0/+36
|
* enchant_provider_is_valid: do NULL checks in same order as fields are declaredReuben Thomas2021-11-081-6/+6
|
* Make request_dict member of provider mandatoryReuben Thomas2021-11-081-16/+14
|
* Fix generation of man pagesReuben Thomas2021-11-082-6/+4
| | | | | | | | First, enchant-lsmod man page does not need to be run through the edit script, as it contains no references to datadir. Secondly, datadir needs to be handled differently, to prevent its being prematurely expanded by configure.
* Clean up whitespace in main sourcesReuben Thomas2021-11-085-103/+103
|
* Remove redundant braces; merge initializers with declarationsReuben Thomas2021-11-083-89/+42
| | | | This also makes the style of the code more consistent.
* lib.c: remove redundant NULL checksReuben Thomas2021-11-081-23/+20
| | | | | | A provider cannot be loaded unless it has a ‘list_dicts’ member, which is checked by enchant_provider_is_valid(). Hence, no need to check that ‘list_dicts’ is non-NULL before calling it.
* lib.c, pwl.c: update copyright dates for myselfReuben Thomas2021-11-072-1/+2
| | | | Add myself to pwl.c, which I have worked on for some years!
* Fix license exceptionReuben Thomas2021-08-063-6/+6
| | | | | Update license exception that allows linking with proprietary spellcheckers to include all copyright holders, not just Dom Lachowicz.
* enchant.ordering: do not prefer aspell for “en”Reuben Thomas2021-08-061-1/+0
| | | | | Only for locales it supports. (This was the intent of the previous change, as far as I can tell, in commit c030073.)
* lib.c: ignore hidden files when loading providersReuben Thomas2021-06-181-0/+1
|
* enchant.c: add -p flagReuben Thomas2021-06-142-12/+49
| | | | | Support -p, as ispell and hunspell do, to specify a personal wordlist (in practice, this could be a per-project wordlist).
* pwl.c: allow the PWL APIs to be passed -1 as word lengthReuben Thomas2021-06-142-9/+21
| | | | | | This is strictly an API/ABI change, but bump only the minor version, as in practice it is simply an extension that makes the enchant_pwl_* APIs work like the enchant_dict_* APIs.
* enchant-provider.h: add documentation (fix #239)Reuben Thomas2021-06-031-13/+36
| | | | | | | | | | Document enchant_get_user_config_dir and enchant_get_conf_dirs. Tweak other documentation for consistency. Correct the name of the preprocessor guard token in the closing comment. This commit only changes comments.
* Make Aspell the default backend for English locales it supportsPeter Oliver2021-02-041-0/+5
| | | | | | The data at http://aspell.net/test/common-all-all/ indicates that Aspell is much quicker than Hunspell and achieves slightly better results. Fixes #270.
* enchant.c: report output of --version/-v to stdout, not stderrReuben Thomas2020-11-021-1/+1
| | | | This is the normal output, and may be required on stdout by clients.
* src/lib.c: sort tags in enchant_broker_list_dictsReuben Thomas2020-09-071-27/+26
| | | | | This makes enchant-lsmod produce more readable output. The order of output may as well be useful.
* enchant: ignore -B for ispell compatibility (Emacs passes -B to ispell)Reuben Thomas2020-09-071-2/+3
|
* libenchant: fix use of deprecated GLib macroReuben Thomas2020-09-012-2/+2
| | | | | G_UNICODE_COMBINING_MARK (deprecated since GLib 2.30) becomes G_UNICODE_SPACING_MARK.
* enchant.c: some slight tidy-upReuben Thomas2020-09-011-13/+6
|
* Nuspell provider implementationPanderMusubi2020-02-273-12/+12
|
* src/enchant.c: overhaul command-line parsing and documentationReuben Thomas2020-01-182-110/+108
| | | | | | | | | | | | Use getopt for simpler parsing of command-line arguments and better error handling and reporting. Remove some stream variables, as in many cases we always use stdout. Overhaul the documentation of command-line options and make it consistent between the output of ‘enchant -h’ and enchant(1). Allow -L to be combined with another mode; it is useless otherwise!
* src/pwl.c: fix buffer overflow in trie matchingReuben Thomas2020-01-181-1/+1
| | | | | | Add an extra byte to the ‘word’ buffer, as the check for transposed characters can look two characters (here, NULs, so one character per byte) past the end of the buffer.
* src/enchant.c: remove FILE * arg from print_help(): only stdout is usedReuben Thomas2020-01-181-3/+3
|
* src/pwl.c: treat PWL as text, not binaryReuben Thomas2019-09-011-3/+3
|
* src/pwl.c: cope with different line endings when reading PWLReuben Thomas2019-09-011-5/+3
| | | | | (Fixes a problem with the tests on Windows, which for some reason only manifested recently.)
* src/lib.c: use correct g_free for result of g_build_filename (not free)Reuben Thomas2019-06-301-1/+1
|
* tokenize_line: fix skipping back over non-word characters (fix #212)Reuben Thomas2019-06-171-11/+9
| | | | | | | We were truncating the string a byte at a time rather than a character at a time. Thanks to Juha Jeronen for the bug report.
* Add missing enchant-lsmod.1.in to EXTRA_DIST in srcReuben Thomas2018-01-031-1/+1
|
* Fix FSF addressRobert-André Mauchin2017-12-148-8/+8
|
* Make enchant_relocate public in the provider APIReuben Thomas2017-12-112-1/+11
| | | | As a result, bump version to 2.2.0.
* Fix use of relocatable so that relocatable.c is correctly builtReuben Thomas2017-12-111-1/+0
| | | | See http://lists.gnu.org/archive/html/bug-gnulib/2017-12/msg00012.html
* Make man pages parallel installable (fix #168)Reuben Thomas2017-12-074-13/+22
|
* enchant-lsmod.1 should not be in DISTCLEANFILESReuben Thomas2017-11-261-1/+1
|
* Support parallel installation of different Enchant versionsReuben Thomas2017-11-253-16/+18
| | | | | | Fixes #162. Thanks to Sébastien Wilmet. Followed instructions at https://developer.gnome.org/programming-guidelines/unstable/parallel-installability.html.en
* Overhaul enchant(1)Reuben Thomas2017-08-093-12/+33
| | | | | Generate the system paths in the man page at build time, and rewrite the section FILES AND DIRECTORIES.
* Fixes #146: add a man page for enchant-lsmodReuben Thomas2017-08-073-2/+37
|
* enchant-lsmod: simplify argument parsing, and add -word-charsReuben Thomas2017-08-071-64/+57
| | | | | -word-chars prints the value of enchant_dict_get_extra_word_characters for the given language’s dictionary.
* Deal with strdup failure in enchant_relocateReuben Thomas2017-07-311-6/+26
|
* Allow Enchant to be built without --enable-relocatableReuben Thomas2017-07-311-0/+4
| | | | | | | Thanks to Bruno Haible for pointing out the disadvantages (primarily for GNU system packagers) of forcing --enable-relocatable. Warn that the tests will fail for a non-relocatable build.