Commit Graph

30 Commits

Author SHA1 Message Date
hallgren
1048a89ca7 ModuleName and Ident are now distinct types
This makes the documentation clearer, and can potentially catch more
programming mistakes.
2014-10-21 19:20:31 +00:00
hallgren
f27d509075 Introducing GF.Text.Pretty for more concise pretty printers and GF.Infra.Location for modularity
GF.Text.Pretty provides the class Pretty and overloaded versions of the pretty
printing combinators in Text.PrettyPrint, allowing pretty printable values to
be used directly instead of first having to convert them to Doc with functions
like text, int, char and ppIdent. Some modules have been converted to use
GF.Text.Pretty, but not all. Precedences could be added to simplify the pretty
printers for terms and patterns.

GF.Infra.Location contains the types Location and L, factored out from
GF.Grammar.Grammar, and the class HasSourcePath. This allowed the import
of GF.Grammar.Grammar to be removed from GF.Infra.CheckM, making it more
like a pure library module.
2014-07-27 22:06:23 +00:00
aarne
bcabc34527 disabled ad hoc overloading, as it still causes mysterious errors. It is just a few lines in Rename.hs 2014-02-09 19:55:17 +00:00
aarne
2552b859fd ad hoc overloading: a new way to deal with name resolution conflicts. Previously the renamer warned if there was e.g. an unqualified reference to mkAdv, which could come from either Syntax or Paradigms. The renamer picked randomly one of the alternatives, which then often failed in type checking. Now, all candidates are collected into a new structure AdHocOverload [Term], which is accessed by the type checker to make the choice based on the type of the constant. This eliminates some of the warnings and some of the error due to wrong choices. In some rare cases, the inherited constants have the same type, which cannot be resolved by overloading. In such cases, the type checker does the same as the renamer did before: pick the "first" option (i.e. the one that happens to be the first in the list returned by the renamer) and issues a warning. In this patch, only a couple of lines are changed. The typechecker (RConcrete) has more substantial changes, and will be recorded as the next patch. 2014-02-01 13:10:36 +00:00
hallgren
41827b1aab Show relative file paths in error messages
This is to avoid one trivial reason for failures in the test suite.
2013-12-06 15:43:34 +00:00
hallgren
c8cbd4477f Reduced clutter in monadic code
+ Eliminated vairous ad-hoc coersion functions between specific monads 
  (IO, Err, IOE, Check) in favor of more general lifting functions
  (liftIO, liftErr).
+ Generalized many basic monadic operations from specific monads to
  arbitrary monads in the appropriate class (MonadIO and/or ErrorMonad),
  thereby completely eliminating the need for lifting functions in lots
  of places.

This can be considered a small step forward towards a cleaner
compiler API and more malleable compiler code in general.
2013-11-20 00:45:33 +00:00
hallgren
decd7122de Eliminate mutual dependencies between the GF compiler and the PGF library
+ References to modules under src/compiler have been eliminated from the PGF
  library (under src/runtime/haskell). Only two functions had to be moved (from
  GF.Data.Utilities to PGF.Utilities) to make this possible, other apparent
  dependencies turned out to be vacuous.

+ In gf.cabal, the GF executable no longer directly depends on the PGF library
  source directory, but only on the exposed library modules. This means that
  there is less duplication in gf.cabal and that the 30 modules in the
  PGF library will no longer be compiled twice while building GF.

  To make this possible, additional PGF library modules have been exposed, even
  though they should probably be considered for internal use only. They could
  be collected in a PGF.Internal module, or marked as "unstable", to make
  this explicit.

+ Also, by using the -fwarn-unused-imports flag, ~220 redundant imports were
  found and removed, reducing the total number of imports by ~15%.
2013-11-05 13:11:10 +00:00
kr.angelov
43fb9b3b7a added the linref construction in GF. The PGF version number is now bumped 2013-10-30 12:53:36 +00:00
kr.angelov
e6dbd74dc9 GF.Grammar.Grammar.allExtends now returns a list of source modules instead of just the module names. This saves extra lookups later 2013-09-23 09:18:27 +00:00
hallgren
1846e6bf65 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
885aaca6de Consistenly use SourceGrammar instead of [SourceModule] when calling compiler passes 2012-10-19 19:56:00 +00:00
hallgren
abb1aedd3d GF.Compile.Rename: report many errors instead of stopping after the first one
Using accumulated errors in the Check monad.

TODO: some errors are still not accumulated, but thanks to checkMapRecover
at least one error per judgement is reported.
2012-06-25 14:15:02 +00:00
aarne
7f047a0ab5 checkMapRecover: find undefined idents in all jments in Rename 2012-06-25 11:14:52 +00:00
kr.angelov
48558197a8 more structured format for errors and warnings from the compiler 2011-11-15 13:33:44 +00:00
kr.angelov
e161f93f4d the new design for -tags 2011-11-14 16:08:56 +00:00
kr.angelov
bbe42d1e90 Now PMCFG is compiled per module and at the end we only link it. The new compilation schema is few times faster. 2011-11-10 14:09:41 +00:00
kr.angelov
54019c75ae merge GF.Infra.Modules and GF.Grammar.Grammar. This is a preparation for the separate PGF building 2011-11-02 13:57:11 +00:00
kr.angelov
bc35626489 Now the compiler maintains more precise information for the source locations of the different definitions. There is a --tags option which generates a list of all identifiers with their source locations. 2011-11-02 11:44:59 +00:00
aarne
f7dc299c87 qualification with real module name accepted, as in GF refman 2011-10-25 14:03:27 +00:00
hallgren
38db834a65 GF.Infra.Modules: minor tweaks
Still keeping the modules both in a list and in a finite map. The overhead is
smaller than I initially thought.
2011-08-31 11:18:16 +00:00
hallgren
0325f7264d GF.Infra.Modules: keep the modules of a grammar in a finite map instead of a list
This speeds up the compilation of PhrasebookFin.pgf by 12%, mosly by speeding
up calls to lookupModule in calls from lookupParamValues, in calls
from allParamValues.

The invariant "modules are stored in dependency order" is no longer respected!
But the type MGrammar is now abstract, making it easier to maintain this or
other invariants in the future.
2011-08-30 18:54:50 +00:00
aarne
9023382c31 GenIP, GenRP in Extra and any_Quant in ExtraEng 2011-07-21 08:25:04 +00:00
krasimir
fe1205059b operations in the abstract syntax 2010-11-12 19:37:19 +00:00
krasimir
e92151caf8 redesign the open-literals API 2010-07-01 08:51:59 +00:00
krasimir
b9b19b6334 rename GF.Grammar.Predef.isPredefCat to isLiteralCat 2010-06-17 15:08:12 +00:00
krasimir
4f1f8b6308 refactoring in GF.Grammar.Grammar 2010-05-28 14:15:15 +00:00
krasimir
d3a84f994b store and propagate the exact source location for all judgements in the grammar. It may not be used accurately in the error messages yet 2010-03-22 21:15:29 +00:00
krasimir
899ed163a1 no need to keep the list of constructors per category in .gfo 2010-02-16 09:34:02 +00:00
krasimir
4485e97181 refactor GF.Infra.Modules for better error messages 2010-01-31 15:54:25 +00:00
krasimir
c92f9d1c0c reorganize the directories under src, and rescue the JavaScript interpreter from deprecated 2009-12-13 18:50:29 +00:00