summaryrefslogtreecommitdiff
path: root/src/runtime/c/pgf
AgeCommit message (Collapse)Author
2015-05-11added all orthographic primitiveskrasimir
2015-05-08implemented CAPIT in parsingkrasimir
2015-05-08fix in the dictionary lookupkrasimir
2015-05-08a new nice and elegant algorithm for dictionary lookup in the parserkrasimir
2015-05-08the parser is now fully Unicode compatiblekrasimir
2015-05-07fix the NERC in C to work for PN as literalkrasimir
2015-04-20added getpagesize() definition for Windowskrasimir
2015-03-06remove the meta_item variable from the parse statekrasimir
2015-02-26fix some out-of-date code for the PGF_RESULT_DEBUG mode in the parserkrasimir
2015-02-26use the same trick for comparison of doubles in cmp_expr_statekrasimir
2015-02-26bugfix in the parser which caused some items to be poped in the wrong order ↵krasimir
if their probability differ with less that 0.5
2015-02-24remove three more uses of meta productions that were left in debugging modekrasimir
2015-02-24implemented escape characters when reading/printing an abstract expression ↵krasimir
with string literals
2015-02-19silence some warnings in GNU Lightning for i386_64krasimir
2015-02-19parsing with meta rules is now removed since we don't use them anymore and I ↵krasimir
would gladly remove some code.
2015-02-18fix in the parser for callbacks in the middle of a wordkrasimir
2015-02-11fix in the JIT compiler that would hopefully make it work on iOSkr.angelov
2014-12-29API for word alignment in the C runtime and in the Haskell bindingkr.angelov
2014-12-19added mkStr to the Haskell bindingkr.angelov
2014-12-19added mkApp to the Haskell bindingkr.angelov
2014-12-16change in the API for literalskr.angelov
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.
2014-12-01fix the last patchkr.angelov
2014-12-01a small optimization in the PGF parser which gives me ~5% speed upkr.angelov
2014-12-01fix in GNU lightning to enable the compilation of two of the the floating ↵kr.angelov
point operations.
2014-12-01fix in linearizer.h to make jpgf compilablekr.angelov
2014-11-27fix in linearizer.h which resolves a warningkr.angelov
2014-11-27remove pgf/parser.hkr.angelov
2014-11-26fix the handling of the heuristic factor. it was broken when I added the ↵kr.angelov
BIND token which means that the App grammar was running about 2-3 times slower.
2014-11-07expand the JIT window for x64kr.angelov
2014-10-31a skeletal module for higher-order pattern unification in the C runtimekr.angelov
2014-10-31now (+) in the abstract syntax works, i.e. it knows how to deal with partial ↵kr.angelov
sums
2014-10-30a more efficient tail call by using the new TUCK instructionkr.angelov
2014-10-26an explicit SAVE instructionkr.angelov
2014-10-23a simple optimization in cmp_item_prob with great effect on the parser's speedkr.angelov
2014-10-23fix the debug mode in the reasonerkr.angelov
2014-10-22CAPIT is now naturally supported in the C runtimekr.angelov
2014-10-22an explicit PUSH_FRAME instructionkr.angelov
2014-10-17the C runtime now supports reading and writing expressions with indexed meta ↵kr.angelov
variables. The type checker generates fresh indices
2014-10-16bugfix in the parser for abstract expressions in the C runtimekr.angelov
2014-10-16finally proper stack unwind in the evaluatorkr.angelov
2014-10-14the compiler now allows + to be used as a floating point addition in the ↵kr.angelov
abstract syntax
2014-10-14bugfix in the bytecode generatorkr.angelov
2014-10-14simplify jit_bare_retkr.angelov
2014-10-09Prelude.CAPIT is now a built-in primitive. It still generates &| in the ↵kr.angelov
Haskell runtime but will be intepreted in the C runtime
2014-10-09throw away the long obsolete runtime type information in the C runtimekr.angelov
2014-10-09fix for lightning for ARMkr.angelov
2014-10-09replace the hash maps in the abstract syntax with binary search tableskr.angelov
2014-10-08now we statically allocate closures for all top-level functions and all ↵kr.angelov
nullary constructors. closures are dynamically allocated only for CAFs. this reduces memory use and time to allocate dynamic closures
2014-10-07a missing case in instruction SETkr.angelov
2014-10-07added a missing case for PUSH instruction with a global closurekr.angelov