Commit Graph

1282 Commits

Author SHA1 Message Date
aarne 8555c65c36 a first version of PGF.AbsCompute 2008-10-10 15:53:17 +00:00
bjorn 7a5d3da347 Added option to treat some categories as lexical when generating Haskell data types. 2008-10-10 11:55:12 +00:00
aarne e56e6483f6 simplified 1.5 by removing mathematical/ 2008-10-06 12:49:13 +00:00
aarne f376576af0 added a command for tree operations and started a module for defining them 2008-10-05 19:49:26 +00:00
aarne 673bc27755 adapted old AbsCompute for use with source GF terms 2008-10-05 18:01:42 +00:00
aarne 1f8c999f1b prevent pgf from being empty when there are no concrete syntaxes 2008-10-05 15:44:16 +00:00
aarne be9ff0cbca forced checking def definitions; accept pgf with no concretes 2008-10-05 15:21:08 +00:00
aarne 56c2c0e067 print full types instead of lock field heuristics in overload resolution if the heuristic is misleading 2008-10-03 18:18:21 +00:00
aarne 4b707dbf01 return error on undefined opers in instance and resource 2008-10-03 16:16:45 +00:00
bjorn 4b41039a5d Added | syntax for variants. 2008-10-03 07:35:26 +00:00
bjorn f362f6ad01 Regenerate source GF parser from GF.cf. Now, when GF/Source/Makefile is used, no hand-hacking is needed. 2008-10-02 14:11:41 +00:00
bjorn 2fc5c7f579 Added src/GF/Source/Makefile for regenerating the GF source parser from GF.cf. 2008-10-02 13:42:24 +00:00
bjorn c21d0d10b8 Converted GF/Source/*.hs to Unix line endings, to be able to see what happens when I regenerated the files. 2008-10-02 13:23:54 +00:00
bjorn 7c30f07f75 Fail more gracefully on unknown file name extensions. 2008-10-02 12:32:03 +00:00
aarne d53f8e0644 added mode 'gf --run' for running silently a script ; made quizzes handle character encoding correctly ; for this end, collected coding functions in GF.Text.Coding 2008-10-01 16:01:51 +00:00
bjorn cec77015ad Use parantheses for precedence in pgf-pretty lincats. 2008-09-30 12:02:58 +00:00
bjorn 14373feb3a Added --cfg option for specifying which CFG transformations to use. Added startcatonly CFG trasnformation. Removed output formats that are now easily done with --cfg: "regular", "nolr". 2008-09-30 11:52:11 +00:00
bjorn ed7d52843b Fixed CFG start cats for categories with more than one constituent. 2008-09-30 10:53:36 +00:00
bjorn 3ad9d12c81 Use filtering in regular printer. 2008-09-30 09:39:09 +00:00
bjorn e1d90c00b9 SRG conversion: print CFG when a unknown cat is found in renameCats. Do filtering in ebnf printer. 2008-09-30 09:32:16 +00:00
bjorn 6791378846 pgf-pretty: remove single field record lincat hack. handle f "" (S []) correctly. 2008-09-30 09:31:31 +00:00
bjorn 662663ffa3 Some pgf-pretty fixes, including note by Peter about W "apa" (S []). 2008-09-30 07:24:01 +00:00
peb bfcfb513f3 Changes to the prolog output format 2008-09-25 07:32:39 +00:00
bjorn 2884598b7c Added pgf-pretty output-format 2008-09-26 15:21:32 +00:00
bjorn 0e869a3574 Added isLiteralCat function to PGF.Macros. 2008-09-26 15:04:13 +00:00
bjorn 8fe2a1cc59 Added semantic interpretation tag printing to the *bnf grammar printers. 2008-09-26 12:57:20 +00:00
bjorn b64d25b3be Use ebnf format for "regular" output-format. 2008-09-26 10:20:18 +00:00
bjorn a84849bee2 Changed nolr printer to use EBNF format. 2008-09-26 10:03:42 +00:00
bjorn 9010afef50 ebnf printer: print external cats first. 2008-09-26 09:58:57 +00:00
bjorn d54ba6f112 bnf and ebnf printer: align ::= symbols. 2008-09-26 09:54:09 +00:00
bjorn 93e9d60765 SRG category renaming: don't count the external cats when calculating indices. 2008-09-26 09:44:02 +00:00
bjorn d80c71f688 Changed SRG category renaming to use GF cat + index among CFG cats for that GF cat, instead of the old Concrete syntax name + index among all CFG cats. 2008-09-26 09:39:48 +00:00
bjorn c588e8ca2b Oops, I accidentally disabled all CFG simplification and LR-removal with the ENBF printing patch. 2008-09-26 09:39:04 +00:00
bjorn 8c348f963f Change category naming in CFG generation: the number now starts from 1 for each GF category, instead of being the FCFG category number. 2008-09-26 09:28:03 +00:00
bjorn 824d72bf16 Added --output-format=ebnf. 2008-09-26 09:20:39 +00:00
bjorn 84dec7aaa2 In CFG generation: don't add linrow number if the lincat has only a single constituent. 2008-09-26 09:06:52 +00:00
bjorn 367cc0acb9 Added --output-format=nolr. Minor makeRegular refactoring. 2008-09-25 13:41:07 +00:00
bjorn 8001ea8ea0 Added --output-format=regular. 2008-09-25 11:49:00 +00:00
bjorn 4d96eb7e9a Fixe copy-paste error: srgs_abnf_nonrec produced xml, fixed now. 2008-09-25 11:36:04 +00:00
bjorn 716250bcf1 Somewhat nicer FCFG output format. 2008-09-25 08:44:23 +00:00
bjorn 90b2441c34 Added an FCFG output format (--output-format=fcfg). This now lives in GF.Speech.PGFToCFG, but should probably move somewhere else. 2008-09-25 08:41:49 +00:00
bjorn 0d1b2c01b9 Got SRGS ABNF generation to compile. 2008-09-23 12:56:25 +00:00
bjorn 66d4909142 Added srgs_abnf_nonrec grammar printer to GF.Infra.Options. 2008-09-23 12:44:17 +00:00
bjorn c01af6168b Re-added old PrSRGS_ABNF (does not compile yet). 2008-09-23 12:41:44 +00:00
aarne 961261c46e use hex in transliteration table 2008-09-23 09:04:03 +00:00
aarne 2269a5ba41 Telugu transliteration 2008-09-23 08:49:58 +00:00
bjorn 3857f4a6a6 A somewhat better solution to the words/UTF-8 problem: do encoding last, but only on what appears to be string literals. 2008-09-15 13:16:04 +00:00
bjorn a39f8cc5da Temporary fix for the grave accent a encoding problem: change compatPrint to id.
The problem is that lower case a with a grave accent is coded in UTF-8 as \195\160. 
Unicode character \160 is non-breaking space, so Haskell's words function
will break a UTF-8 encoded string at this character.
String literals in the .gfo file are UTF-8 encoded in generateModuleCode,
just before the call to prGrammar (which uses compactPrint, which used words).
The real solution would be to pretty-print the grammar to Unicode, and then
encode as UTF-8. The problem with that is Latin-1 identifers. They are now
kept in Latin-1 in the .gfo file, since Alex can't handle Unicode.
The real solution to that would be to fix Alex to handle Unicode, but 
that is non-trivial. GHC interally uses a very hacky .x file to be
able to lex UTF-8 source files. 

An alternative solution that doesn't address the weirdness of using two different
encodings in the same .gfo as we do now, is to incorporate compactPrint
into the grammar printer, to avoid having to do any postprocessing.
2008-09-15 12:38:37 +00:00
krasimir 2d6e7ad625 print the search path when we can't find a module 2008-09-10 08:55:16 +00:00
aarne 279c0367d8 better error msg for missing record labels 2008-09-09 19:43:33 +00:00