cparser
11 years agoCheck whether a local variable needs an entity more systematically.
Christoph Mallon [Thu, 18 Oct 2012 07:22:44 +0000 (09:22 +0200)]
Check whether a local variable needs an entity more systematically.

11 years agoRemove unused members from struct compound_member_t and struct variable_t.
Christoph Mallon [Wed, 17 Oct 2012 11:49:21 +0000 (13:49 +0200)]
Remove unused members from struct compound_member_t and struct variable_t.

11 years agoAttach debug info to parameter entities/projs.
Christoph Mallon [Wed, 17 Oct 2012 11:47:13 +0000 (13:47 +0200)]
Attach debug info to parameter entities/projs.

11 years agoadded undefine of __BLOCKS__ macro that is set by cpp under Mac OS 10.7
Sebastian Hack [Thu, 13 Sep 2012 11:58:53 +0000 (13:58 +0200)]
added undefine of __BLOCKS__ macro that is set by cpp under Mac OS 10.7

11 years agoRemove object files instead of whole build dir when cleaning.
Manuel Mohr [Fri, 7 Sep 2012 07:18:10 +0000 (09:18 +0200)]
Remove object files instead of whole build dir when cleaning.

Strangely enough, the old way of running rm -rf $(BUILDDIR)/ did not
work if BUILDDIR was a symbolic link.  In this case, the rm command was
a nop.  rm -rf $(BUILDDIR) would have removed the symbolic link whereas
rm -rf $(BUILDDIR)/* would have removed all directory contents.  We
adopt libfirm's strategy and remove all object files while leaving the
dep files intact.

11 years agoDon't make libfirm when cleaning cparser.
Manuel Mohr [Thu, 6 Sep 2012 14:27:55 +0000 (16:27 +0200)]
Don't make libfirm when cleaning cparser.

11 years agoRemove the AST node for while, use the one for for instead.
Christoph Mallon [Wed, 8 Aug 2012 06:23:05 +0000 (08:23 +0200)]
Remove the AST node for while, use the one for for instead.

This removes quite a bit of code duplication.

11 years agoFixed clang warning.
Sebastian Buchwald [Thu, 30 Aug 2012 08:28:57 +0000 (10:28 +0200)]
Fixed clang warning.

11 years agostrictconv is not necessary anymore in latest firm
Matthias Braun [Tue, 21 Aug 2012 14:50:37 +0000 (16:50 +0200)]
strictconv is not necessary anymore in latest firm

11 years agocparser knowns unroll loops
Matthias Braun [Fri, 10 Aug 2012 12:15:47 +0000 (14:15 +0200)]
cparser knowns unroll loops

11 years agoimproved fix for integer expression printing
Matthias Braun [Thu, 26 Jul 2012 17:21:09 +0000 (19:21 +0200)]
improved fix for integer expression printing

11 years agoonly output digits in integer literals in write_jna
Matthias Braun [Thu, 26 Jul 2012 14:13:02 +0000 (16:13 +0200)]
only output digits in integer literals in write_jna

11 years agoNew output mode to generate size (in bytes) of some compound types
Matthias Braun [Thu, 26 Jul 2012 13:51:30 +0000 (15:51 +0200)]
New output mode to generate size (in bytes) of some compound types

This will be used by the X10 native code integration.

11 years agorememeber attributes on struct types
Matthias Braun [Thu, 26 Jul 2012 13:51:07 +0000 (15:51 +0200)]
rememeber attributes on struct types

11 years agofix memory error
Matthias Braun [Mon, 23 Jul 2012 13:11:19 +0000 (15:11 +0200)]
fix memory error

11 years agoadapt to libFirm ASM changes
Matthias Braun [Mon, 23 Jul 2012 12:56:15 +0000 (14:56 +0200)]
adapt to libFirm ASM changes

11 years agoadd a special case for __attribute__((__const))
Matthias Braun [Wed, 18 Jul 2012 08:50:16 +0000 (10:50 +0200)]
add a special case for __attribute__((__const))

according to gcc documentation only "__XXX__" and "XXX" should be
allowed for attribute names, but it seems gcc incorrectly also accepts
__XXX for keywords with an __XXX variant. And of course some glibc
headers use this feature.

11 years agofix -fsyntax-only
Matthias Braun [Wed, 18 Jul 2012 08:50:05 +0000 (10:50 +0200)]
fix -fsyntax-only

11 years agofix bad include after renaming
Matthias Braun [Mon, 16 Jul 2012 16:13:13 +0000 (18:13 +0200)]
fix bad include after renaming

11 years agoadd missing file to Makefile
Matthias Braun [Mon, 16 Jul 2012 16:10:47 +0000 (18:10 +0200)]
add missing file to Makefile

11 years agoRename hashset.c to hashset.c.inl.
Christoph Mallon [Mon, 16 Jul 2012 15:04:19 +0000 (17:04 +0200)]
Rename hashset.c to hashset.c.inl.

11 years agoprovide our own hashset implementation
Matthias Braun [Mon, 16 Jul 2012 14:40:15 +0000 (16:40 +0200)]
provide our own hashset implementation

We want to be independent of libfirms containertypes (as they may get
deprecated at some point).
The implementation we use is a copy of libfirm :)

11 years agorespect getcwd return value
Matthias Braun [Fri, 13 Jul 2012 21:09:08 +0000 (23:09 +0200)]
respect getcwd return value

this fixes a warning

11 years agoimplement is_universal_char_valid_identifier_c99
Matthias Braun [Fri, 13 Jul 2012 12:16:49 +0000 (14:16 +0200)]
implement is_universal_char_valid_identifier_c99

11 years agoOnly classify several expressions as constant if no error is involved.
Christoph Mallon [Mon, 9 Jul 2012 10:35:53 +0000 (12:35 +0200)]
Only classify several expressions as constant if no error is involved.

11 years agoAdd the option c11 to -std and add the C11 keywords.
Christoph Mallon [Mon, 9 Jul 2012 09:55:23 +0000 (11:55 +0200)]
Add the option c11 to -std and add the C11 keywords.

- Use _Thread_local instead of __thread internally.
- Use _Alignof instead of __alignof__ internally.

11 years agoImplement -idirafter.
Christoph Mallon [Mon, 9 Jul 2012 07:52:49 +0000 (09:52 +0200)]
Implement -idirafter.

11 years agoRecord and handle, whether a search path is a system path.
Christoph Mallon [Mon, 2 Jul 2012 16:38:02 +0000 (18:38 +0200)]
Record and handle, whether a search path is a system path.

11 years agobetter search paths handling; respect -I
Matthias Braun [Thu, 5 Jul 2012 18:13:00 +0000 (20:13 +0200)]
better search paths handling; respect -I

11 years agobugfix for external non-emitting API
Matthias Braun [Thu, 5 Jul 2012 16:33:02 +0000 (18:33 +0200)]
bugfix for external non-emitting API

11 years agowhitespace before tokens affects macro redef
Matthias Braun [Wed, 6 Jun 2012 17:52:19 +0000 (19:52 +0200)]
whitespace before tokens affects macro redef

11 years agoimplement includefile macro expansion
Matthias Braun [Sun, 3 Jun 2012 14:01:38 +0000 (16:01 +0200)]
implement includefile macro expansion

11 years agoshow additional space before macro expanded '#'
Matthias Braun [Sat, 26 May 2012 07:28:19 +0000 (09:28 +0200)]
show additional space before macro expanded '#'

gcc/clang output an additional space before a (macro-expanded) '#' at
line begin. So we replicate this behaviour.

11 years agopreprocessor: implement stringify operation '#'
Matthias Braun [Fri, 25 May 2012 17:30:39 +0000 (19:30 +0200)]
preprocessor: implement stringify operation '#'

11 years agoSimplify peek_expansion().
Christoph Mallon [Sun, 1 Jul 2012 07:28:33 +0000 (09:28 +0200)]
Simplify peek_expansion().

11 years agopreprocessor: output '3' flag for system headers
Matthias Braun [Mon, 18 Jun 2012 23:12:51 +0000 (01:12 +0200)]
preprocessor: output '3' flag for system headers

11 years agoImplement #include_next.
Christoph Mallon [Sun, 1 Jul 2012 06:40:29 +0000 (08:40 +0200)]
Implement #include_next.

11 years agoTest skip_mode in parse_pragma_directive().
Christoph Mallon [Mon, 2 Jul 2012 16:59:29 +0000 (18:59 +0200)]
Test skip_mode in parse_pragma_directive().

11 years agoProperly test for a valid identifer after #define and #undef.
Christoph Mallon [Fri, 29 Jun 2012 18:05:53 +0000 (20:05 +0200)]
Properly test for a valid identifer after #define and #undef.

11 years agoReimplement input_error(), which got lost with the old lexer.
Christoph Mallon [Fri, 22 Jun 2012 20:39:26 +0000 (22:39 +0200)]
Reimplement input_error(), which got lost with the old lexer.

11 years agoImplement #error.
Christoph Mallon [Fri, 22 Jun 2012 09:29:19 +0000 (11:29 +0200)]
Implement #error.

11 years agoCorrect do_include().
Christoph Mallon [Fri, 6 Jul 2012 09:09:59 +0000 (11:09 +0200)]
Correct do_include().

Do not alter headername, which is needed unaltered later on.

11 years agoRemove pointless local variable.
Christoph Mallon [Fri, 6 Jul 2012 09:01:53 +0000 (11:01 +0200)]
Remove pointless local variable.

11 years agoRemove redundante fclose(), which is handles by close_input() a few lines later.
Christoph Mallon [Fri, 6 Jul 2012 08:59:05 +0000 (10:59 +0200)]
Remove redundante fclose(), which is handles by close_input() a few lines later.

11 years agoAlways reset unit->input in close_input().
Christoph Mallon [Fri, 6 Jul 2012 09:00:46 +0000 (11:00 +0200)]
Always reset unit->input in close_input().

11 years agorework compilation loop, use our preproc for -E
Matthias Braun [Thu, 5 Jul 2012 16:19:27 +0000 (18:19 +0200)]
rework compilation loop, use our preproc for -E

also removed --lextest which was broken and not really usefull anymore

11 years agomake pp_token emit API public
Matthias Braun [Thu, 5 Jul 2012 15:57:27 +0000 (17:57 +0200)]
make pp_token emit API public

11 years agoAvoid endless loop when encountering EOF while parsing an initializer.
Christoph Mallon [Thu, 5 Jul 2012 06:57:58 +0000 (08:57 +0200)]
Avoid endless loop when encountering EOF while parsing an initializer.

11 years agoPlug a memory leak (and reduce code duplication) for string initializers with extra...
Christoph Mallon [Thu, 5 Jul 2012 06:33:55 +0000 (08:33 +0200)]
Plug a memory leak (and reduce code duplication) for string initializers with extra curly braces.

11 years agoShow the right type when warning about excessive initializer elements.
Christoph Mallon [Thu, 5 Jul 2012 06:33:29 +0000 (08:33 +0200)]
Show the right type when warning about excessive initializer elements.

11 years agofix line-directive parser
Matthias Braun [Wed, 4 Jul 2012 12:45:30 +0000 (14:45 +0200)]
fix line-directive parser

It was happily eating tokens from the following lines as long as they
were numbers and strings.

11 years agodo NOT use #if for target platform decisions
Matthias Braun [Wed, 4 Jul 2012 09:11:18 +0000 (11:11 +0200)]
do NOT use #if for target platform decisions

(or we will end up like gcc where you have to compile each host/target
 combination again, instead of having 1 compiler which can produce
 everything)

11 years agoremove unnecessary ATOMIC_TYPE_INVALID
Matthias Braun [Tue, 3 Jul 2012 10:47:20 +0000 (12:47 +0200)]
remove unnecessary ATOMIC_TYPE_INVALID

11 years agoslightly simplify wchar_t handling
Matthias Braun [Tue, 3 Jul 2012 10:38:21 +0000 (12:38 +0200)]
slightly simplify wchar_t handling

11 years agoProperly support -ansi.
Christoph Mallon [Tue, 3 Jul 2012 14:55:44 +0000 (16:55 +0200)]
Properly support -ansi.

11 years agoDocument -std=c90 in the manpage.
Christoph Mallon [Tue, 3 Jul 2012 14:54:56 +0000 (16:54 +0200)]
Document -std=c90 in the manpage.

11 years agoAdded documentation for -std=c90.
Sebastian Buchwald [Tue, 3 Jul 2012 14:05:51 +0000 (16:05 +0200)]
Added documentation for -std=c90.

11 years agoAdd -std=c90 as alias for -std=c89.
Christoph Mallon [Tue, 3 Jul 2012 14:32:58 +0000 (16:32 +0200)]
Add -std=c90 as alias for -std=c89.

11 years agoFactorise conversion from lang_standard_t to a string into a function.
Christoph Mallon [Tue, 3 Jul 2012 14:29:53 +0000 (16:29 +0200)]
Factorise conversion from lang_standard_t to a string into a function.

11 years agoFixed warning for -std=iso9899:199409.
Sebastian Buchwald [Tue, 3 Jul 2012 13:54:04 +0000 (15:54 +0200)]
Fixed warning for -std=iso9899:199409.

11 years agoRenamed enum entry.
Sebastian Buchwald [Tue, 3 Jul 2012 13:49:37 +0000 (15:49 +0200)]
Renamed enum entry.

11 years agoUse obstack_copy() instead of obstack_alloc() + memcpy().
Christoph Mallon [Tue, 3 Jul 2012 11:15:44 +0000 (13:15 +0200)]
Use obstack_copy() instead of obstack_alloc() + memcpy().

11 years agoRemove the token registration macro TS.
Christoph Mallon [Tue, 3 Jul 2012 10:52:24 +0000 (12:52 +0200)]
Remove the token registration macro TS.

T is sufficient.

11 years agoConcatenate T_ earlier to the token names.
Christoph Mallon [Tue, 3 Jul 2012 10:43:17 +0000 (12:43 +0200)]
Concatenate T_ earlier to the token names.

This way we get rid of many #undef, which were necessary, because several token names are macros in system headers.

11 years agoRemove _ANSI.
Christoph Mallon [Tue, 3 Jul 2012 10:26:07 +0000 (12:26 +0200)]
Remove _ANSI.

It should be just another word for _C89, but wasn't.

11 years agofix warning
Matthias Braun [Thu, 21 Jun 2012 09:10:21 +0000 (11:10 +0200)]
fix warning

11 years agoReplace the old lexer by the new preprocessor.
Christoph Mallon [Wed, 20 Jun 2012 19:43:35 +0000 (21:43 +0200)]
Replace the old lexer by the new preprocessor.

11 years agocopy dollar in identifier feature from lexer
Matthias Braun [Tue, 19 Jun 2012 00:24:06 +0000 (02:24 +0200)]
copy dollar in identifier feature from lexer

11 years agocopy universal character parsing from lexer
Matthias Braun [Tue, 19 Jun 2012 00:08:09 +0000 (02:08 +0200)]
copy universal character parsing from lexer

11 years agocopy line and pragma directive parsers from lexer
Matthias Braun [Mon, 18 Jun 2012 23:13:15 +0000 (01:13 +0200)]
copy line and pragma directive parsers from lexer

11 years agoimplement macro calling
Matthias Braun [Mon, 21 May 2012 23:58:52 +0000 (01:58 +0200)]
implement macro calling

11 years agomove sym_anonymous to symbol.h
Matthias Braun [Sat, 19 May 2012 14:28:29 +0000 (16:28 +0200)]
move sym_anonymous to symbol.h

11 years agopreprocessor: scanning for :: token was missing
Matthias Braun [Sun, 27 May 2012 22:10:21 +0000 (00:10 +0200)]
preprocessor: scanning for :: token was missing

11 years agoalways set symbol for non-literal tokens
Matthias Braun [Fri, 25 May 2012 16:29:50 +0000 (18:29 +0200)]
always set symbol for non-literal tokens

This now also allows to differentiate digraphs from their replacement in
the preprocessor output

11 years agoSimplify MAYBE macros.
Christoph Mallon [Wed, 20 Jun 2012 13:56:15 +0000 (15:56 +0200)]
Simplify MAYBE macros.

11 years agoCorrect indentation.
Christoph Mallon [Wed, 20 Jun 2012 13:43:28 +0000 (15:43 +0200)]
Correct indentation.

11 years agoTurn T_EOF into a properly registered token.
Christoph Mallon [Wed, 20 Jun 2012 14:58:05 +0000 (16:58 +0200)]
Turn T_EOF into a properly registered token.

11 years agoRewrite print_token().
Christoph Mallon [Wed, 20 Jun 2012 15:37:18 +0000 (17:37 +0200)]
Rewrite print_token().

11 years agopreprocessor: keep unknown chars in preproc mode
Matthias Braun [Fri, 25 May 2012 17:59:03 +0000 (19:59 +0200)]
preprocessor: keep unknown chars in preproc mode

11 years agoRemove the unused return value from parse_include_directive().
Christoph Mallon [Wed, 20 Jun 2012 12:09:40 +0000 (14:09 +0200)]
Remove the unused return value from parse_include_directive().

11 years agoReplace !is_ifndef by is_ifdef.
Christoph Mallon [Wed, 20 Jun 2012 10:16:04 +0000 (12:16 +0200)]
Replace !is_ifndef by is_ifdef.

11 years agoAdd eat_token() for more clarity in the preprocessor.
Christoph Mallon [Wed, 20 Jun 2012 10:14:09 +0000 (12:14 +0200)]
Add eat_token() for more clarity in the preprocessor.

11 years agorework preprocessor token number
Matthias Braun [Mon, 28 May 2012 15:44:21 +0000 (17:44 +0200)]
rework preprocessor token number

We now use the "normal" token numbers where possible, special preprocessor token
numbers are only used to disambiguate between directives. This change was also
necessary because preprocessing directive identifiers are not reserved keywords.

11 years agoMove the skip_mode test into parse_*_directive().
Christoph Mallon [Wed, 20 Jun 2012 12:10:06 +0000 (14:10 +0200)]
Move the skip_mode test into parse_*_directive().

11 years agofix warning
Matthias Braun [Mon, 18 Jun 2012 22:54:32 +0000 (00:54 +0200)]
fix warning

11 years agoReduce code duplication in obstack_grow_utf8().
Christoph Mallon [Sun, 17 Jun 2012 15:18:44 +0000 (17:18 +0200)]
Reduce code duplication in obstack_grow_utf8().

11 years agorename obstack_grow_symbol=>obstack_grow_utf8
Matthias Braun [Sun, 27 May 2012 21:44:25 +0000 (23:44 +0200)]
rename obstack_grow_symbol=>obstack_grow_utf8

This is more logical name since we don't put symbol_t on the obstack but a
unicode character in utf8 encoding.

11 years agofix error output for unknown non-ascii characters
Matthias Braun [Fri, 25 May 2012 17:45:46 +0000 (19:45 +0200)]
fix error output for unknown non-ascii characters

11 years agotypedef token_kind_t as int
Matthias Braun [Fri, 25 May 2012 15:51:08 +0000 (17:51 +0200)]
typedef token_kind_t as int

The enum was renamed to token_kind_tag_t, because extensions like the
ast-grep branch can register additional tokens. This also avoids
warnings, because we have a clearly defined type for the enum now.

11 years agopreprocessor: backslash is not a punctuator token
Matthias Braun [Sun, 3 Jun 2012 14:00:21 +0000 (16:00 +0200)]
preprocessor: backslash is not a punctuator token

11 years agomove docu in front of function
Matthias Braun [Tue, 22 May 2012 14:29:40 +0000 (16:29 +0200)]
move docu in front of function

11 years agofix bad argument to errorf
Matthias Braun [Tue, 29 May 2012 20:54:50 +0000 (22:54 +0200)]
fix bad argument to errorf

11 years agofix warnings
Matthias Braun [Wed, 23 May 2012 19:20:25 +0000 (21:20 +0200)]
fix warnings

11 years agoChange multi-case macros so their usage looks like ordinary case labels.
Christoph Mallon [Sun, 17 Jun 2012 13:00:43 +0000 (15:00 +0200)]
Change multi-case macros so their usage looks like ordinary case labels.

11 years agoLet the lexer produce preprocessor numbers T_NUMBER instead of T_INTEGER and T_FLOATI...
Christoph Mallon [Sun, 17 Jun 2012 11:51:56 +0000 (13:51 +0200)]
Let the lexer produce preprocessor numbers T_NUMBER instead of T_INTEGER and T_FLOATINGPOINT.

Handle the details of parsing a number in the parser.

11 years agoadapt to latest libfirm
Matthias Braun [Fri, 15 Jun 2012 19:59:20 +0000 (21:59 +0200)]
adapt to latest libfirm

11 years agoremove unnecessary include
Matthias Braun [Thu, 17 May 2012 16:53:48 +0000 (18:53 +0200)]
remove unnecessary include

because fluffy is reusing the code and doesn't have the header.

11 years agoSimplify printing string initializers.
Christoph Mallon [Thu, 14 Jun 2012 17:55:54 +0000 (19:55 +0200)]
Simplify printing string initializers.

They use value initializers internally.

11 years agoInclude string encoding in string_t.
Christoph Mallon [Thu, 14 Jun 2012 17:12:01 +0000 (19:12 +0200)]
Include string encoding in string_t.

11 years agoPush finishing the string into make_string()/identify_string() and rename them to...
Christoph Mallon [Thu, 14 Jun 2012 17:00:05 +0000 (19:00 +0200)]
Push finishing the string into make_string()/identify_string() and rename them to sym_make_string().

This resolves some instances where \0 was erroneously counted towards the string length.