d22c07dcfb
refine the GNU lightning patch to avoid generating unnecessary code
2015-07-16 15:29:14 +00:00
fac48d43e8
bugfixes in the JIT compiler for x86_64. One of the fixes is actually in GNU lightning
2015-07-16 15:21:09 +00:00
192b1cc018
the reasoner's states are now closures just like the code for the def rules. This means more compact JIT compiled code and uniformity with the code for def rules
2015-07-04 15:06:34 +00:00
b3680b6025
merge the states for reasoning and evaluation into a single structure
2015-07-03 11:53:52 +00:00
6f2afdd53e
bugfix in the parser for lexical lookup
2015-06-30 12:54:19 +00:00
c7f1803410
bugfix in the bottom-up prediction for the parser
2015-06-25 09:46:43 +00:00
c6cdb66cf2
runtime/c/pf/jit.c: silence warning about alloca on MINGW32
2015-06-23 15:49:31 +00:00
e352742b8c
the orthographic tokens were broken when they appear inside a dictionary entry. Now this is fixed by treating those as syntactic rules
2015-06-12 13:03:19 +00:00
9b0edaacdb
print error message for missing abstract categories in the C runtime
2015-06-05 07:41:59 +00:00
786da46a99
added callback in C for unknown words
2015-05-19 09:15:48 +00:00
616a24cba4
the NERC in C is now fully up-to-date with the other NERCs. In addition the NERC in Haskell and Java now refuses to recognize names that match words in the lexicon. The latest is useful for German
2015-05-19 07:26:56 +00:00
09a57b34a8
bugfix in the NERC in C
2015-05-18 15:02:44 +00:00
ce05465365
the callbacks in C now take the concrete syntax as a parameter
2015-05-18 14:41:33 +00:00
773a33979b
pgf_lookup_morpho and pgf_lookup_word_prefix are now also case insensitive if case_sensitive=off
2015-05-18 14:20:35 +00:00
b4e41cabdd
"flags case_sensitive=off" makes the parser case insensitive
2015-05-11 14:11:42 +00:00
1e0d7be4f4
added all orthographic primitives
2015-05-11 13:01:39 +00:00
13998e3287
implemented CAPIT in parsing
2015-05-08 18:54:50 +00:00
483df35954
fix in the dictionary lookup
2015-05-08 18:12:24 +00:00
ea0f74d455
a new nice and elegant algorithm for dictionary lookup in the parser
2015-05-08 14:38:24 +00:00
365c7bb1d8
the parser is now fully Unicode compatible
2015-05-08 09:23:29 +00:00
bf3b6f2794
fix the NERC in C to work for PN as literal
2015-05-07 15:16:55 +00:00
5d7633fdb4
added getpagesize() definition for Windows
2015-04-20 09:57:28 +00:00
8f0624bc30
remove the meta_item variable from the parse state
2015-03-06 15:18:50 +00:00
a5817ff492
fix some out-of-date code for the PGF_RESULT_DEBUG mode in the parser
2015-02-26 13:31:26 +00:00
9c83b2e27d
use the same trick for comparison of doubles in cmp_expr_state
2015-02-26 13:10:01 +00:00
532647c978
bugfix in the parser which caused some items to be poped in the wrong order if their probability differ with less that 0.5
2015-02-26 12:17:59 +00:00
715ac1ca62
remove three more uses of meta productions that were left in debugging mode
2015-02-24 10:59:18 +00:00
da11a7127f
implemented escape characters when reading/printing an abstract expression with string literals
2015-02-24 09:35:12 +00:00
e893d41393
silence some warnings in GNU Lightning for i386_64
2015-02-19 12:53:52 +00:00
b92bbcfc37
parsing with meta rules is now removed since we don't use them anymore and I would gladly remove some code.
2015-02-19 11:59:51 +00:00
e3a87c657b
fix in the parser for callbacks in the middle of a word
2015-02-18 16:14:29 +00:00
d361b7f5d2
fix in the JIT compiler that would hopefully make it work on iOS
2015-02-11 13:07:06 +00:00
3bd40dbab6
API for word alignment in the C runtime and in the Haskell binding
2014-12-29 10:59:20 +00:00
f44f9a1b00
added mkStr to the Haskell binding
2014-12-19 08:58:02 +00:00
bb1c1dd27c
added mkApp to the Haskell binding
2014-12-19 08:47:00 +00:00
9bc5349e62
change in the API for literals
...
The API in the C runtime as well as in the Haskell, Python and Java binding
is changed. Now instead of adding the literal callbacks to the concrete syntax
you need to supply them every time when you need to parse. The main reason is:
- referentially transparent API for Haskell
- when we start using memory mapped files we will not be allowed to change
anything in the grammar data structures. At that point the old API would
be impossible to use.
2014-12-16 10:21:26 +00:00
64feca59c8
fix the last patch
2014-12-01 11:03:50 +00:00
960ba43149
a small optimization in the PGF parser which gives me ~5% speed up
2014-12-01 10:48:15 +00:00
14a2821fd8
fix in GNU lightning to enable the compilation of two of the the floating point operations.
2014-12-01 09:42:21 +00:00
35e76a5fe8
fix in linearizer.h to make jpgf compilable
2014-12-01 09:41:48 +00:00
6f2195f793
fix in linearizer.h which resolves a warning
2014-11-27 15:20:46 +00:00
f63e8faf00
remove pgf/parser.h
2014-11-27 11:12:14 +00:00
41e217a641
fix the handling of the heuristic factor. it was broken when I added the BIND token which means that the App grammar was running about 2-3 times slower.
2014-11-26 15:56:36 +00:00
cabafd68a6
expand the JIT window for x64
2014-11-07 15:43:12 +00:00
70c70fa791
a skeletal module for higher-order pattern unification in the C runtime
2014-10-31 19:57:56 +00:00
b1e6a21c56
now (+) in the abstract syntax works, i.e. it knows how to deal with partial sums
2014-10-31 14:16:11 +00:00
9b0f354c7c
a more efficient tail call by using the new TUCK instruction
2014-10-30 13:09:50 +00:00
efedec187c
an explicit SAVE instruction
2014-10-26 11:40:12 +00:00
e169f225cf
a simple optimization in cmp_item_prob with great effect on the parser's speed
2014-10-23 21:22:22 +00:00
fc6658c2e7
fix the debug mode in the reasoner
2014-10-23 08:15:12 +00:00
krasimir