With no changes to the build system affecting Happy, I am getting
errors from compiler/parser/Parser.hs when using the Windows binary
of Happy 1.16. First, there is a parse error for the CPP #include
statement on line 6 (ghc is given the option -cpp, of course). After
moving the statement out of Parser.hs and onto the command line with -
#include there are parse errors such as a do statement without a
terminating expression (indentation error) and a parse error before a
where clause.
I am using:
ghc-6.6.1 (Windows binary, Neil's install)
happy-1.16
gcc-3.5
Note: oddly enough, gcc-3.5 seems to emit numbered lines with gcc -E -
u -P still. They are of the form:
# [number]
They are not the typical line pragmas. Anyway, I had to change the
preprocessing statements in mk/suffix.mk and mk/package.mk from:
...
grep -v '^#pragma GCC' ...
to
...
grep -v '^#' ...
This should not have affected the Happy-generated Parser.hs or ghc -cpp.
On a related note, I have run out of time to work on this for the
moment. What is done? Practically everything except a full test and
proper warnings for deprecated options. For TargetPlatform i386-
unknown-windows GHC there are:
* options and command lines using cl, link and yasm (these should work)
* changes to the makefiles and libraries (conditional compilation,
mostly--literally hundreds of #if (defined(mingw32_HOST_OS) || defined
(windows_HOST_OS)) and the like)
* changes to the RTS (some inline assembler), except that rts/
StgCRun.c works off of USE_MINIINTERPRETER because I haven't gotten
as far as examining the assembler output from CL to redo StgRun.
If anyone is interested in the Windows-native build thus far I will
send a patch with directions or post the directions I have so far the
wiki. (I didn't before because they are subject to change as things
evolve.)
Cheers,
Pete
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
