39 commits

Author	SHA1	Message	Date
Dale Weiler	489ad486bc	Get it compiling in visual studio again.	2013-04-24 01:43:53 +00:00
Dale Weiler	160e7cf7ee	Remove trailing whitespace	2013-04-21 10:24:55 +00:00
Dale Weiler	bf6eb9432e	Fix memleak and assignment for cached lengths in corrector	2013-02-01 06:22:29 +00:00
Dale Weiler	d7a47eb5f9	Ignore most memory comparisions by validating the lengths are the same first. Becomes O(log(n)) from O(n) for correct_exists.	2013-01-30 23:32:15 +00:00
Dale Weiler	a6ecc3cbc4	Cache lengths in corrector. This speeds up the corrector a little.	2013-01-30 23:22:03 +00:00
Dale Weiler	5f9f6c11b8	Get it compiling with C++ compilers again.	2013-01-30 06:49:50 +00:00
Dale Weiler	1916a6e339	Slightly larger code, but strategic inline for corrector stuff. Makes it a little faster on my RPi (unnoticable on my desktop though).	2013-01-30 06:36:00 +00:00
Wolfgang Bumiller	fa90d2d06e	correction fix: don't try to create edits for an empty string	2013-01-10 21:09:44 +01:00
Wolfgang Bumiller	10c7f4f838	don't call correct_edit anew for each scope	2013-01-09 21:16:22 +01:00
Dale Weiler	ac5cc49840	remove a the	2013-01-06 12:28:27 +00:00
Dale Weiler	9a4e215179	Document the awesome hack	2013-01-06 12:24:05 +00:00
Wolfgang Bumiller	9c8ddb3771	An evil optimization to the trie, now has a fixed amount of branches and uses a char-to-index map to index into the branches...	2013-01-06 13:07:28 +01:00
Dale Weiler	1e17b5f696	Cleaner transformation calls (one less size_t for agruments). We can coalesce it in correct_edit.	2013-01-06 11:02:39 +00:00
Dale Weiler	0f4090402d	No need to +1 for comparision in correct_known_resize	2013-01-06 10:37:09 +00:00
Dale Weiler	cfdeaf4786	Fix comments	2013-01-06 10:35:43 +00:00
Dale Weiler	056779d3b8	Remove trailing shitspace	2013-01-06 10:33:09 +00:00
Dale Weiler	a4c1e63637	Less pointer dereferences	2013-01-06 05:17:30 +00:00
Dale Weiler	cc7e1a3363	Screw you clang :-)	2013-01-06 05:14:57 +00:00
Dale Weiler	ee33b4e5cc	A little faster, plus some more research	2013-01-06 05:13:19 +00:00
Dale Weiler	0c59274c54	This is a work of art, it deserve nice comments :)	2013-01-06 04:09:12 +00:00
Dale Weiler	d98cc564b1	Fixes and more documentation	2013-01-06 04:06:38 +00:00
Dale Weiler	44a7154f58	Add a notice	2013-01-06 03:52:37 +00:00
Dale Weiler	106db76b9d	Use memcmp with strlen for correct_exists (it's a hell of a lot faster)	2013-01-06 03:34:56 +00:00
Dale Weiler	cc8558025b	No more uninitialized	2013-01-06 03:33:21 +00:00
Dale Weiler	9841240aab	Some more nicer loops	2013-01-06 03:29:55 +00:00
Dale Weiler	e2e4907b60	Nicer loops	2013-01-06 03:26:09 +00:00
Dale Weiler	7e87c61d78	Cleanups and documentation	2013-01-06 03:06:56 +00:00
Dale Weiler	294870c5ba	Remove override macros	2013-01-06 02:41:15 +00:00
Dale Weiler	efecd160ca	Remove an illegal mem_d	2013-01-06 02:40:36 +00:00
Dale Weiler	0fb089fbb7	Cleanup the pool system for correct.c. Use blubs correct_resize stuff.	2013-01-06 02:39:07 +00:00
Wolfgang Bumiller	8bc9d4b427	Committing an evil allocator and a trie to speed up the correction stuff	2013-01-05 19:09:36 +01:00
Wolfgang Bumiller	0e3bb0e4f4	Get rid of correct_strndup and correct_concat altogether, reduces each generated edit to a single alloc	2013-01-04 14:50:50 +01:00
Dale Weiler	36d02d010e	Got rid of all the memleaks. We can now merge with master.	2013-01-04 11:53:40 +00:00
Dale Weiler	18b9473cf8	Itegration of corrector. Seems to be some leaks in the score keeping for the probability system.	2013-01-04 11:44:25 +00:00
Dale Weiler	d97e032fcf	Cleanups and add the corrector to the makefile. Starting integration with the parser.	2013-01-04 10:05:41 +00:00
Dale Weiler	7ef051f58a	Fix all the memleaks in the corrector. Holy shit, the amount of hours I wasted trying to find out where I left out ONE little free.	2013-01-04 09:46:22 +00:00
Dale Weiler	ab64706cc3	Fix another two leaks	2013-01-03 21:38:07 +00:00
Dale Weiler	2189dd4b85	Remove some memory leaks in the corrector. There is still some memory leaks that are proving impossible to fix.	2013-01-03 21:25:35 +00:00
Dale Weiler	861987e13e	Work on "did you mean? <ident>" support for errors. Using a three-part Bayes Theorem expression (language model, error model and control mechanisim).	2013-01-03 19:44:13 +00:00