Commit Graph

176 Commits

Author SHA1 Message Date
hallgren
aa9280b287 Workaround for a problem with lock fields
This problem showed up after delaing eta expansion until just before partial
evaluation
2013-09-10 13:43:59 +00:00
hallgren
b93f817058 new-comp: delay eta expansion until just before partial evaluation
This seems to work for the most part, but a problem showed up in WordsCat.gf in
the phrasebook.
2013-05-30 16:01:12 +00:00
hallgren
bf5dfb2293 Nonlinear patterns in concrete syntax are now detected and reported as errors
Before, they were silently converted to linear patterns.
Nonlinear patterns in MorphoCat.gf, ParadigmsGre.gf and ParadigmsFin.gf have
been make linear by renaming pattern variables.
2013-09-09 19:52:08 +00:00
hallgren
3543cb7a16 partial evaluator: prettier complaint about nonlinears patterns 2013-09-09 16:23:01 +00:00
hallgren
990789eb91 partial evaluator: complain about nonlinear patterns 2013-09-09 15:53:14 +00:00
hallgren
83283b0d6a Fix an old name shadowing bug in concrete syntax by removing the refresh pass
The refresh pass does not correctly keep track of the scope of local variables
and can convert things like \x->(\x->x) x into \x1->(\x2->x2) x2. Fortunately,
it appears that the refresh pass is not needed anymore, so it has been removed.
2013-09-09 14:29:57 +00:00
kr.angelov
df26b134fc fix in the GF compiler and runtime which let us to define pre construct detecting whether this is the last token. 2013-09-03 07:51:25 +00:00
kr.angelov
a20cd77d25 nonExist now does the expected thing 2013-08-23 13:17:45 +00:00
aarne
850e02cb6e hs datatype generation for empty abstract types added 2013-07-24 08:11:16 +00:00
kr.angelov
4394b8a3bc a bit more informative error message in GrammarToPGF 2013-04-22 12:14:39 +00:00
hallgren
5bce1d7a16 gf -s/-q now silences "no linearization of" warnings 2013-04-10 23:00:01 +00:00
hallgren
80fe693546 Fix a problem with pattern macros in pre { } expressions
The old partial evaluator has special rules to convert pattern macros in
pre { } expressions. These rules were missing in the new partial evaluator.
2013-03-16 13:36:23 +00:00
hallgren
e5ce4d4b1e partial evaluator: push predefined functions inside variants
This should prevent errors like

Internal error in Compute.ConcreteNew:
    Applying Predef.drop: Expected a value of type String, got VFV [VString "gewandt",VString "gewendet"]
2013-03-12 16:36:58 +00:00
Sergei Trofimovich
0173b5f4ff ghc-7.6: allow directory-1.2
Get rid of old-time depend (and ClockTime in favour of UTCTime).
time-compat helps to retain backward compatibility with directory-1.1
and lower.
2013-03-09 21:38:43 +00:00
hallgren
0feb386691 Faster regular expression pattern matching in the grammar compiler.
The sequence operator (x+y) was implemented by splitting the string to be
matched at all positions and trying to match the parts against the two
subpatterns. To reduce the number of splits, we now estimate the minimum and
maximum length of the string that the subpatterns could match. For common
cases, where one of the subpatterns is a string of known length, like
in (x+"y") or (x + ("a"|"o"|"u"|"e")+"y"), only one split will be tried.
2013-02-27 20:59:43 +00:00
kr.angelov
4922ab6cc4 now the beam size for the statistical parser can be configured by using the flag beam_size in the top-level concrete module 2013-02-12 10:53:13 +00:00
hallgren
d4b6d1b6fe Better error message for unsupported token gluing
Instead of "Internal error in ...", you now get a proper error message with
a source location and a function name.
2013-01-29 16:25:03 +00:00
hallgren
14c8da214c Fix a bug with record extension
Add a conversion rule for ({ l1 = e } ** x).l2 in PMCFG generation. (A rule
for the symmetric case (x ** { l1 = e }).l2 was added some time ago.)
2013-01-29 14:59:16 +00:00
hallgren
713e883ad7 Better error message for Predef.error
+ Instead of "Internal error in ...", you now get a proper error message with
  a source location and a function name.
+ Also added some missing error value propagation in the partial evaluator.
+ Also some other minor cleanup and error handling fixes.
2013-01-28 16:12:56 +00:00
aarne
1eb45d955b improved error message for overloading in case the given signature looks the same as one of the expected ones: it shows full records rather than just lock fields. 2013-01-28 14:00:23 +00:00
hallgren
a8f281491c partial evaluator: fix token glueing bug
"a"+("b"++"c") was simplified to "bb"++"c" instead of "ab"++c.
2013-01-11 15:14:42 +00:00
hallgren
12dec2503d bug fix in the new partial evaluator
It can leave wildcard tables in their origial form, but it easy to handle
them in the unfactor function in GeneratePMCFG.
2012-12-20 16:41:43 +00:00
hallgren
ece3e33c52 partial evaluator bug fix
It failed to delay table selection when the selector contains a run-time
variable, causing "gf: Prelude.(!!): index too large" instead.

Also:
  + Show better source locations on unexpected errors, to aid bug hunting.
  + Removed unused SourceGrammar argument to value2term.
2012-12-19 23:12:37 +00:00
hallgren
b4207d1b00 GF.Grammar.Lookup: new function lookupResDefLoc
It's like lookupResDef but it includes a source location in the output.
2012-12-19 23:08:56 +00:00
hallgren
c3beefba47 partial evaluator bug fix
Int was missing from the list of predefined canonical constants.
2012-12-18 13:03:20 +00:00
kr.angelov
20aaa4a989 The first prototype for exhaustive generation in the C runtime. The trees are always listed in decreasing probability order. There is also an API for generation from Python 2012-12-14 15:32:49 +00:00
hallgren
9ab3a6034d Add language extension for ghc<7.4
FlexibleInstances does not imply TypeSynonymInstances, apparently.
2012-12-14 14:21:46 +00:00
hallgren
d7e3c869c2 More work on the new partial evaluator
The work done by the partial evaluator is now divied in two stages:
 - A static "term traversal" stage that happens only once per term and uses
   only statically known information. In particular, the values of lambda bound
   variables are unknown during this stage. Some tables are transformed to
   reduce the cost of pattern matching.
 - A dynamic "function application" stage, where function bodies can be
   evaluated repeatedly with different arguments, without the term traversal
   overhead and without recomputing statically known information.

Also the treatment of predefined functions has been reworked to take advantage
of the staging and better handle partial applications.
2012-12-14 14:00:21 +00:00
hallgren
5e091d2e3d partial evaluator work
* Evaluate operators once, not every time they are looked up
* Remember the list of parameter values instead of recomputing it from the
  pattern type every time a table selection is made.
* Quick fix for partial application of some predefined functions.
2012-12-11 15:37:41 +00:00
hallgren
621291a575 Compute.ConcreteNew: add missing case for variant functions
Also adding a test case in the test suite for this.
2012-12-10 13:25:32 +00:00
hallgren
81f843ac06 Compute.ConcreteNew: bug fix for indirectly defined pattern macros
More changes are probably needed to make pattern macros first class values.
Also includes minor changes related to variants and error messages.
2012-12-06 16:44:03 +00:00
aarne
03c3c57290 produce error message instead of failure of irrefutable pattern Ok ty_C in GrammarToPGF, to help find compilation errors; the ones I've found are because an inherited abstract excludes something that the inherited concrete does not exclude. 2012-12-02 19:40:45 +00:00
hallgren
7d616c01b4 new-comp: rewrite f (x|y) into (f x|f y)
With this change, all languages in molto/mgl/mixture except German and Polish
can be compiled.
2012-11-16 13:47:10 +00:00
hallgren
27e675910a Adding a new experimental partial evalutator
GF.Compile.Compute.ConcreteNew + two new modules contain a new
partial evaluator intended to solve some performance problems with the old
partial evalutator in GF.Compile.Compute.ConcreteLazy. It has been around for
a while, but is now complete enough to compile the RGL and the Phrasebook.

The old partial evaluator is still used by default. The new one can be activated
in two ways:

  - by using the command line option -new-comp when invoking GF.
  - by using cabal configure -fnew-comp to make -new-comp the default. In this
    case you can also use the command line option -old-comp to revert to the old
    partial evaluator.

In the GF shell, the cc command uses the old evaluator regardless of -new-comp
for now, but you can use "cc -new ..." to invoke the new evaluator.

With -new-comp, computations happen in GF.Compile.GeneratePMCFG instead of
GF.Compile.Optimize. This is implemented by testing the flag optNewComp in
both modules, to omit calls to the old partial evaluator from GF.Compile.Optimize
and add calls to the new partial evaluator in GF.Compile.GeneratePMCFG.
This also means that -new-comp effectively implies -noexpand.

In GF.Compile.CheckGrammar, there is a check that restricted inheritance is used
correctly. However, when -noexpand is used, this check causes unexpected errors,
so it has been converted to generate warnings, for now.

-new-comp no longer enables the new type checker in
GF.Compile.Typeckeck.ConcreteNew.

The GF version number has been bumped to 3.3.10-darcs
2012-11-13 14:09:15 +00:00
hallgren
63093c32f3 Eliminate warnings about deprecated use of catch and try
This is also needed for compatibility with GHC 7.6.
2012-11-08 15:53:46 +00:00
hallgren
9f8c0f86f1 Compute.ConcreteNew: support variants
Also add a missing check for Predef values in apply.
2012-10-24 17:49:20 +00:00
hallgren
6fcd435cd9 GeneratePMCFG: prefix messages about "impossible" errors with 'Internal error:'
Just to make them easier to spot when wading through thousands of lines of
warnings...
2012-10-24 17:08:52 +00:00
hallgren
be75546965 Refactor compileSourceModule
There was 55 lines of rather repetitive code with calls to 6 compiler passes.
They have been replaced with 19 lines that call the 6 compiler passes
plus 26 lines of helper functions.
2012-10-19 20:14:11 +00:00
hallgren
1195db1da3 Consistenly use SourceGrammar instead of [SourceModule] when calling compiler passes 2012-10-19 19:56:00 +00:00
hallgren
584b6ba86d Use NOINLINE for build info and darcs version info
... to avoid unnecessary recompilation of other modules.
2012-10-18 20:01:22 +00:00
kr.angelov
91ca7c9a1b catch all case in GenerateBC.genFun 2012-09-18 08:32:21 +00:00
hallgren
997734c8ba Add type info to "Warning: ignoring lock fields in resolving..." 2012-09-06 13:15:48 +00:00
kr.angelov
545e48e881 another fix for teyjus 2012-08-30 08:09:30 +00:00
peter.ljunglof
b416f5bbf7 Use nub' instead of nub in some places, remove some unused nub imports 2012-08-29 21:48:34 +00:00
peter.ljunglof
a7de16c34b Added an O(n log n) version of nub
The new nub is called nub', and it replaces the old sortNub which was 
not lazy and did not retain the order between the elements.
2012-08-29 21:45:10 +00:00
peter.ljunglof
e2ecdfed1f Emacs only recognizes utf-8, not UTF-8, in file headers 2012-08-29 21:40:49 +00:00
kr.angelov
f8fe23fda7 A basic infrastructure for generating Teyjus bytecode from the GF abstract syntax 2012-08-29 11:43:02 +00:00
kr.angelov
9fc7cac46f the first draft of GF.Compile.Instructions 2012-08-01 09:21:59 +00:00
peter.ljunglof
94e2999ac5 Minor changes in Python export 2012-07-02 21:11:20 +00:00
peter.ljunglof
a162488b05 Fixed minor bug in prolog export 2012-06-28 07:01:29 +00:00