| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2014-10-09 | replace the hash maps in the abstract syntax with binary search tables | kr.angelov | |
| 2014-10-09 | add the const modifier to the key in the two procedures for binary search in ↵ | kr.angelov | |
| libgu | |||
| 2014-10-08 | use MAP_FIXED in gu_mmap_pool | kr.angelov | |
| 2014-10-08 | an experimental memory mapped pool. still not in use | kr.angelov | |
| 2014-10-08 | now 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 | |||
| 2014-10-07 | a missing case in instruction SET | kr.angelov | |
| 2014-10-07 | added a missing case for PUSH instruction with a global closure | kr.angelov | |
| 2014-10-07 | the indirection should be done in the evaluate_expr_thunk gate instead of in ↵ | kr.angelov | |
| the pgf_evaluate_expr_thunk function. this ensures lazyness | |||
| 2014-10-06 | using instruction RET was wrong; now use EVAL | kr.angelov | |
| 2014-10-06 | fix the order in which arguments are stored in the mk_const gate | kr.angelov | |
| 2014-10-06 | push a stack frame around recursive calls to guarantee that a nested call to ↵ | kr.angelov | |
| a failing function will not crash | |||
| 2014-10-06 | partial implementation for the FAIL instruction | kr.angelov | |
| 2014-09-30 | bugfix in pgf_evaluate_expr_thunk | kr.angelov | |
| 2014-09-30 | fix pgf_value2expr for partial applications | kr.angelov | |
| 2014-09-30 | enough fixes to get the JIT compiler compile for ARM. It is still broken if ↵ | kr.angelov | |
| you try to use it | |||
| 2014-09-30 | fix in reader.h for compiling jpgf.c | kr.angelov | |
| 2014-09-30 | fix in pgf_print_expr for lambda abstractions with more than one argument | kr.angelov | |
| 2014-09-30 | bugfix in the gate evaluate_value_lambda | kr.angelov | |
| 2014-09-29 | bugfix in the pattern matching compiler and a number of other fixes that I ↵ | kr.angelov | |
| somehow did not push before | |||
| 2014-09-29 | bugfix in the RET instruction | kr.angelov | |
| 2014-09-29 | enlarge the code window in pgf_jit_gates to ensure enough space on 64-bit ↵ | kr.angelov | |
| machines | |||
| 2014-09-25 | the literals API is now exposed in Python | kr.angelov | |
| 2014-09-25 | switch off the debugging of the JIT compiler which I had turned on ↵ | kr.angelov | |
| accidentally. silence two harmless warnings | |||
| 2014-09-25 | now a complete JIT compiler and ByteCode compiler for the def rules in the ↵ | kr.angelov | |
| abstract syntax. there might be some bugs yet to be found, meta variables and computation under lambda is only partially supported | |||
| 2014-09-22 | haskell-bind/PGF2.hsc: unexport functions that break referential transparency | hallgren | |
| loadConcr, unloadConcr and addLiteral modify the Concr structure as a side effect. This means that other functions with a Concr argument (e.g. parse and linearize) are no longer pure. Possible solutions: 1. Don't try to hide the imperative nature of the C run-time system: remove all uses of unsafePerformIO and let all functions operate in the IO monad. 2. Don't export functions with side effects. Perhaps the desired functionality of loadConcr, unloadConcr and addLiteral can be folded into readPGF. The Concr structures can then treaded as immutable after after the readPGF function returns... | |||
| 2014-09-17 | fix a portability problem with the Python binding (found by Prasanth) | kr.angelov | |
| 2014-09-16 | dummy jit_base_tail_finishr for x86_64. | kr.angelov | |
| 2014-09-11 | a major revision of the bytecode generator and JIT compiler. the effect is ↵ | kr.angelov | |
| that now we can compute with lambda functions and with true tail recursion | |||
| 2014-09-10 | now release the FunPtr:s that are allocated for each literal callback | kr.angelov | |
| 2014-09-10 | added an API for custom literals in the Haskell binding | kr.angelov | |
| 2014-09-10 | added loadConcr/unloadConcr to the Haskell binding. This exposes an API for ↵ | kr.angelov | |
| loading grammars compiled with -split-pgf | |||
| 2014-09-05 | the code for def rules now uses proper graph update to preserve lazyness | kr.angelov | |
| 2014-09-05 | full support for recursive def rules in the C runtime | kr.angelov | |
| 2014-09-01 | partial implementation for recursive def rules | kr.angelov | |
| 2014-09-01 | bug fixes in the JIT compiler | kr.angelov | |
| 2014-08-29 | added evaluator.h which I had forgoten | kr.angelov | |
| 2014-08-13 | Fix warnings in 16 modules, mostly forward compatibility warnings from GHC 7.8 | hallgren | |
| 2014-08-11 | pattern matching in def rules is now supported | kr.angelov | |
| 2014-08-11 | a partial support for def rules in the C runtime | kr.angelov | |
| The def rules are now compiled to byte code by the compiler and then to native code by the JIT compiler in the runtime. Not all constructions are implemented yet. The partial implementation is now in the repository but it is not activated by default since this requires changes in the PGF format. I will enable it only after it is complete. | |||
| 2014-07-15 | PGF Web Service: include entire completion in full mode | john.j.camilleri | |
| When using full=yes in the web service 'complete' command, you now get an additional field 'seq' with the longest possible completion. So, given: lin f1 = ss "the" ; f2 = ss ("the red house" | "the real deal") ; and trying to complete on input "th", you get: [ { "from": "TestCnc", "brackets": { "cat": "_", "fid": 0, "index": 0, "fun": "_", "children": [] }, "text": "th", "completions": [ { "token": "the", "funs": [ { "fun": "f1", "hyps": [], "cat": "C", "seq": "the" }, { "fun": "f2", "hyps": [], "cat": "C", "seq": "the red house" }, { "fun": "f2", "hyps": [], "cat": "C", "seq": "the real deal" } ] } ] } ] | |||
| 2014-07-11 | PGF web service: Return additional completion info with 'full' flag | john | |
| BETA! The 'complete' command now has a new flag 'full' which when set returns additional info about completions. Without 'full' flag (default): [ { "from": "PhrasebookEng", "brackets": { "cat": "_", "fid": 0, "index": 0, "fun": "_", "children": [ { "token": "the" } ] }, "text": "su", "completions": [ "supermarket", "suspect" ] } ] With full=true or full=yes: [ { "from": "PhrasebookEng", "brackets": { "cat": "_", "fid": 0, "index": 0, "fun": "_", "children": [ { "token": "the" } ] }, "text": "su", "completions": [ { "token": "supermarket", "funs": [ { "fid": 421, "fun": "Supermarket", "hyps": [], "cat": "PlaceKind" } ] }, { "token": "suspect", "funs": [ { "fid": 445, "fun": "Suspect", "hyps": [], "cat": "Property" } ] } ] } ] | |||
| 2014-07-08 | implemented computing with abstract syntax trees. It passes all test cases ↵ | kr.angelov | |
| except those that require def rules. The design is consistent with the STG virtual machine | |||
| 2014-07-08 | bugfix for the typechecker in the C runtime | kr.angelov | |
| 2014-07-06 | an API for visitor patterns on abstract syntax trees from Python. This makes ↵ | kr.angelov | |
| the embedded grammars especially pleasing | |||
| 2014-07-06 | small fix for the embedded mode in Python | kr.angelov | |
| 2014-07-06 | embedded mode for GF grammar in Python. Only generation for now | kr.angelov | |
| 2014-07-02 | added bracketedLinearize in the Java binding | kr.angelov | |
| 2014-07-02 | fixed rare memory leak in the Java bindings | kr.angelov | |
| 2014-07-01 | a few changes in the Nerc that I have forgoten to push | kr.angelov | |
| 2014-07-01 | fix warnings in the compilation of the Java binding | kr.angelov | |
 |
index : gf-core.git | |
| Unnamed repository; edit this file 'description' to name the repository. | Adrian |
| summaryrefslogtreecommitdiff |