- Cvs-hugs - Haskell.org

cvs commit: hugs98/tests/static mod113.hs hugs98/tests testScript
by Sigbjorn Finne 16 Jul '02

16 Jul '02

sof 2002/07/15 21:56:48 PDT Modified files: tests testScript Added files: tests/static mod113.hs Log: added empty decls/stmts/alts regr. test Revision Changes Path 1.10 +1 -0 hugs98/tests/testScript

1 0

cvs commit: hugs98/src static.c
by Sigbjorn Finne 15 Jul '02

15 Jul '02

sof 2002/07/15 20:44:54 PDT Modified files: src static.c Log: resolveImportList(): when computing the 'subentities' of a type constructor or a class, gracefully handle dcons/members that might not yet have been added to the current module's name table. Revision Changes Path 1.77 +25 -20 hugs98/src/static.c

1 0

cvs commit: hugs98/src machdep.c hugs98/src/unix aclocal.m4 configure
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 14:18:39 PDT Modified files: src machdep.c src/unix aclocal.m4 configure Log: Alistair Cloete's fixes for Darwin Revision Changes Path 1.57 +3 -3 hugs98/src/machdep.c 1.9 +4 -3 hugs98/src/unix/aclocal.m4 1.27 +32 -27 hugs98/src/unix/configure

1 0

cvs commit: hugs98/src storage.c
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 10:46:34 PDT Modified files: src storage.c Log: Avoid double-freeing of dlls Revision Changes Path 1.47 +5 -2 hugs98/src/storage.c

1 0

cvs commit: hugs98/src/unix configure configure.in
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 05:54:47 PDT Modified files: src/unix configure configure.in Log: Fixed byacc again and added dynamic linking flag -r to the list of those worth trying Revision Changes Path 1.26 +154 -1 hugs98/src/unix/configure 1.29 +2 -2 hugs98/src/unix/configure.in

1 0

cvs commit: hugs98/src/unix aclocal.m4 configure configure.in
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 04:08:32 PDT Modified files: src/unix aclocal.m4 configure configure.in Log: Tried to make dynamically loaded file a bit trickier (i.e., make it contain some relocations) to test the linker a bit harder. Fixed the byacc test to work on Solaris. Revision Changes Path 1.8 +5 -1 hugs98/src/unix/aclocal.m4 1.25 +32 -16 hugs98/src/unix/configure 1.28 +1 -1 hugs98/src/unix/configure.in

1 0

cvs commit: hugs98/src builtin.c
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 04:05:26 PDT Modified files: src builtin.c Log: Make ffi error message match RC4 terminology Revision Changes Path 1.27 +3 -3 hugs98/src/builtin.c

1 0

cvs commit: hugs98/tests/ffi FileIO_aux.c
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 03:50:22 PDT Modified files: tests/ffi FileIO_aux.c Log: braino Revision Changes Path 1.2 +1 -1 hugs98/tests/ffi/FileIO_aux.c

1 0

cvs commit: hugs98/tests testScript hugs98/tests/ffi FileIO_aux.c FileIO.hs
by Alastair Reid 09 Jul '02

09 Jul '02

reid 2002/07/09 03:46:14 PDT Modified files: tests testScript tests/ffi FileIO.hs Added files: tests/ffi FileIO_aux.c Log: Apply Sven's patch to testScript so it works on Solaris. Replace totally bogus hardwired constants with O_WRONLY | O_CREAT | O_TRUNC I'd foolishly assumed that constants that old would be completely portable. Revision Changes Path 1.9 +3 -3 hugs98/tests/testScript 1.3 +3 -3 hugs98/tests/ffi/FileIO.hs

1 0

cvs commit: hugs98/tests testScript
by Alastair Reid 06 Jul '02

06 Jul '02

reid 2002/07/06 04:03:04 PDT Modified files: tests testScript Log: testScript no longer overrides Hugs' default path since testScript now assumes you did a make install Revision Changes Path 1.8 +1 -1 hugs98/tests/testScript

1 0

cvs commit: hugs98/src ffi.c storage.c storage.h hugs98/tests testScript
by Alastair Reid 06 Jul '02

06 Jul '02

reid 2002/07/06 03:52:01 PDT Modified files: src ffi.c storage.c storage.h tests testScript Log: Minor tweaks to make it compile and pass ffi tests again Revision Changes Path 1.17 +3 -3 hugs98/src/ffi.c 1.46 +5 -3 hugs98/src/storage.c 1.40 +2 -3 hugs98/src/storage.h 1.7 +5 -5 hugs98/tests/testScript

1 0

cvs commit: hugs98/src server.c
by Sigbjorn Finne 05 Jul '02

05 Jul '02

sof 2002/07/05 05:55:49 PDT Modified files: src server.c Log: shutdownHugsServer: as part of shutdown sequence, call everybody(EXIT) Revision Changes Path 1.17 +3 -2 hugs98/src/server.c

1 0

cvs commit: hugs98/src server.h
by Sigbjorn Finne 05 Jul '02

05 Jul '02

sof 2002/07/05 05:54:25 PDT Modified files: src server.h Log: added typedef for f-ptr to shutdownHugsServer() Revision Changes Path 1.7 +4 -2 hugs98/src/server.h

1 0

cvs commit: hugs98/src connect.h machdep.c storage.c storage.h
by Sigbjorn Finne 05 Jul '02

05 Jul '02

sof 2002/07/04 22:21:27 PDT Modified files: src connect.h machdep.c storage.c storage.h Log: Have Scripts track their associated primInfoDef (if any). In turn, the primInfoDef for a module now tracks its corresponding DLL handle. This is all done so that upon unloading a script, we can properly let go of the underlying DLL (and primInfoDef.) Revision Changes Path 1.40 +3 -2 hugs98/src/connect.h 1.56 +112 -31 hugs98/src/machdep.c 1.45 +99 -23 hugs98/src/storage.c 1.39 +6 -2 hugs98/src/storage.h

2 1

cvs commit: hugs98/src/msc config.h
by Sigbjorn Finne 05 Jul '02

05 Jul '02

sof 2002/07/04 22:23:01 PDT Modified files: src/msc config.h Log: allegedly made up-to-date wrt unix/config.h.in Revision Changes Path 1.10 +45 -0 hugs98/src/msc/config.h

1 0

cvs commit: hugs98/src Makefile.in config.h.in ffi.c hugs98/src/unix linking_notes acconfig.h aclocal.m4 configure configure.in hugs98/tests testScript
by Alastair Reid 30 Jun '02

30 Jun '02

reid 2002/06/30 17:44:49 PDT Modified files: src Makefile.in config.h.in ffi.c src/unix acconfig.h aclocal.m4 configure configure.in tests testScript Added files: src/unix linking_notes Log: Autoconf changes to support new ffi. These have only been tested on Linux so far - help testing on other platforms would be appreciated. To test, search hugs98/src/unix/configure.in for HUGS_TRY_DYNLINK and add lines to test for the flags your platform needs. Changes: Added some GHC autoconfery for figuring out alignments, sizes, etc. (The results make it into header files but are currently ignored by all ffi code.) Added HUGS_TRY_DYNLINK to aclocal.m4 This tries to build a loadable library, tries loading it, and figures out whether to use a leading underscore. The loading test is a simplified version of the code in machdep.c. Typical uses: HUGS_TRY_DYNLINK("/LD") HUGS_TRY_DYNLINK("-shared") HUGS_TRY_DYNLINK("-shared -nostdlib") HUGS_TRY_DYNLINK("-bundle") HUGS_TRY_DYNLINK("-bundle -lc") Killed the old test for leading underscores - it is subsumed by the new autoconf code (which performs a more direct test). Merged the different pieces of compile and link code in ffi.c into a single piece of code (by moving all configuration into autoconf). Revision Changes Path 1.14 +0 -2 hugs98/src/Makefile.in 1.13 +214 -1 hugs98/src/config.h.in 1.16 +19 -23 hugs98/src/ffi.c 1.6 +182 -0 hugs98/src/unix/acconfig.h 1.7 +157 -59 hugs98/src/unix/aclocal.m4 1.24 +1040 -532 hugs98/src/unix/configure 1.27 +49 -30 hugs98/src/unix/configure.in 1.6 +5 -5 hugs98/tests/testScript

1 0

cvs commit: hugs98/src ffi.c machdep.c
by nordland＠glass.cse.ogi.edu 24 Jun '02

24 Jun '02

nordland 2002/06/24 10:00:57 PDT Modified files: src ffi.c machdep.c Log: Small MacOS X tweak. Revision Changes Path 1.15 +6 -2 hugs98/src/ffi.c 1.55 +3 -3 hugs98/src/machdep.c

1 0

Recent ffi commits
by Alastair Reid 23 Jun '02

23 Jun '02

You'll be glad to know that I'm going to let the repository recover for a couple of days while I hang out with my family. This also means I probably won't be able to fix anything I've broken for a couple of days so let's hope I didn't break anything too badly. Top items on my agenda when I get back to it are: - Making sure that the wrapper and dynamic import forms work The code is there - just need to check it works. - Test, test, test Sigbjorn promised to whack on it with HDirect generated code. Sounds like a great stress test. I should try building HGL using --target ffi. - Add something to the test suite. Has to be written so that the tests pass if ffi is disabled or not implemented. - Further contemplation of how to invoke the ffi generation and how/whether that should invoke the C compiler The current story is that hugs +G will generate a fresh .c file and link it to get a .so file for every .hs file it loads which contains ffi code. This will not look so clever when people install their code in read-only directories. - Autoconfiscation Mostly porting/stealing code from the GHC autoconf stuff. Code most needing fixed are HsFFI.h and the function compileAndLink in ffi.c - More libraries (CError, CTypes, ...) Would be easier if we had the hierarchial modules up and running since the GHC versions are close to what we want. Hmmm, didn't Sigbjorn commit a script to suck the tree over? - Finish the implementation of foreign export. The difficult thing about foreign export is that you can export a function at a more specific type than its Haskell type. For example: foreign export ccall "plus_Int" (+) :: Int -> Int -> Int The way to do this is to treat it as though the user had written: plus_Int :: Int -> Int -> Int plus_Int = (+) foreign export ccall plus_Int :: Int -> Int -> Int My problem is that I am completely baffled by the typechecker - I can't figure out how to do this. (Help with typeForeignExport in type.c very welcome.) (I'd feel bad about the length of this list of undone things if I didn't feel that the ffi is much closer to being usable than it was before I started.) I'm not sure how fast I'll get through this list though - I'm supposed to be choosing and booking a wedding venue and I did hardly any real paying work this week. So if anyone feels like claiming some of these tasks as their own, I'd be one happy Haggis. -- Alastair ps There's some longer term goals like: - Stripping the Hugs-specific code out of GreenCard - Removing any primitives that can be implemented by the ffi - Cleaning up internal naming schemes in Hugs. Hugs still talks about MallocPtrs - terminology that died at least 5 years ago. Addr is gone too. - differentiate between primitives (could be associated with any module) and foreign functions (associated with a very particular module). Take the opportunity to cleanup registerPrims which is pretty gross. But these can wait until the new ffi has proved itself.

1 1

cvs commit: hugs98 Credits hugs98/docs ffi-notes.txt server.html server.tex hugs98/src machdep.c hugs98/src/unix README hugs98/tests Readme
by Alastair Reid 22 Jun '02

22 Jun '02

reid 2002/06/22 12:47:38 PDT Modified files: . Credits docs ffi-notes.txt server.html server.tex src machdep.c src/unix README tests Readme Log: Redirected all references to my Yale and Utah email addresses and URLs to point to alastair(a)reid-consulting-uk.ltd.uk http://www.reid-consulting-uk.ltd.uk/alastair/ Revision Changes Path 1.5 +2 -1 hugs98/Credits 1.6 +44 -47 hugs98/docs/ffi-notes.txt 1.2 +3 -3 hugs98/docs/server.html 1.4 +5 -7 hugs98/docs/server.tex 1.54 +3 -3 hugs98/src/machdep.c 1.3 +1 -1 hugs98/src/unix/README 1.8 +3 -1 hugs98/tests/Readme

1 0

cvs commit: hugs98/src static.c storage.c
by Alastair Reid 22 Jun '02

22 Jun '02

reid 2002/06/22 10:01:35 PDT Modified files: src static.c storage.c Log: Fixed rather foolish bug in subText: subText builds a Text from a sub-string of a String. It did this by inserting a '\0' in the string, calling findText and restoring the string. This failed rather miserably if the String was the result of textToStr. Revision Changes Path 1.76 +4 -4 hugs98/src/static.c 1.44 +22 -8 hugs98/src/storage.c

1 0

cvs commit: hugs98/src MkInstal.in
by Alastair Reid 21 Jun '02

21 Jun '02

reid 2002/06/21 16:49:53 PDT Modified files: src MkInstal.in Log: Tweaked install process so that make install doesn't try to run the ffi (so I hopefully don't scrog those on unsupported platforms - which is most platforms). Use 'make install_ffi' to do a normal install and then run 'hugs +G' on any foreign import code in the installed copy. Revision Changes Path 1.9 +6 -6 hugs98/src/MkInstal.in

2 1

cvs commit: hugs98/tests testScript
by Alastair Reid 21 Jun '02

21 Jun '02

reid 2002/06/21 16:40:16 PDT Modified files: tests testScript Log: Added -W flag to test so that MarshalUtils.with doesn't get stomped on by the old-style implicit variable syntax. Can we make -W the default? Revision Changes Path 1.5 +2 -2 hugs98/tests/testScript

1 0

cvs commit: hugs98/tests/ffi types_aux.h
by Alastair Reid 21 Jun '02

21 Jun '02

reid 2002/06/21 16:35:49 PDT Added files: tests/ffi types_aux.h Log: forgeto

1 0

cvs commit: hugs98/lib/exts CString.hs Storable_aux.h .cvsignore CTypes.hs MarshalAlloc.hs MarshalUtils.hs Storable.hs Storable_aux.c hugs98/src HsFFI.h MkInstal.in connect.h ffi.c hugs.c machdep.c static.c hugs98/tests Readme config.in testScript ...
by Alastair Reid 21 Jun '02

21 Jun '02

reid 2002/06/21 16:22:01 PDT Modified files: lib/exts .cvsignore CTypes.hs MarshalAlloc.hs MarshalUtils.hs Storable.hs Storable_aux.c src HsFFI.h MkInstal.in connect.h ffi.c hugs.c machdep.c static.c tests Readme config.in testScript tests/ffi FileIO.hs FileIO.output Forms.hs Sin.hs Types.hs forms_aux.h types_aux.c Added files: lib/exts CString.hs Storable_aux.h Log: Big changes: 1) Hugs +G only generates code for files specified on command line. 2) Hugs now #includes the header file if specified. 3) Hugs will report use of lib specs as an error. (You can turn this off if you #define SILENTLY_IGNORE_FFI_LIB_SPECS) 4) Hugs now ignores lib specs 5) The +L flag can be used to specify things you want added to the compilation command line. This would typically be the things you'd have specified in your lib specs but it might also be compiler flags, preprocessor flags, etc. 6) Hugs uses a single global namespace for any C code called by ffi code. 7) make install applies hugs +G to files in lib/exts which contain ffi code (The list of files is maintained manually both to keep the makefile portable and because some libraries need +L arguments.) build_libexts :: hugs$(EXE) install_include ./hugs +q -P../lib:../lib/exts Storable +G +L"Storable_aux.c" < /dev/null ./hugs +q -P../lib:../lib/exts MarshalAlloc +G < /dev/null ./hugs +q -P../lib:../lib/exts MarshalUtils +G < /dev/null The need to use </dev/null suggests that +G should imply quitting before reading commands and maybe +q too. 8) You now have to make install before running the testsuite Also: Added CString.hs Made CString.hs marginally more like what it should be. Revision Changes Path 1.2 +3 -0 hugs98/lib/exts/.cvsignore 1.2 +8 -1 hugs98/lib/exts/CTypes.hs 1.2 +3 -3 hugs98/lib/exts/MarshalAlloc.hs 1.2 +2 -2 hugs98/lib/exts/MarshalUtils.hs 1.5 +51 -51 hugs98/lib/exts/Storable.hs 1.3 +18 -18 hugs98/lib/exts/Storable_aux.c 1.5 +18 -19 hugs98/src/HsFFI.h 1.8 +8 -3 hugs98/src/MkInstal.in 1.39 +3 -2 hugs98/src/connect.h 1.14 +41 -36 hugs98/src/ffi.c 1.85 +20 -7 hugs98/src/hugs.c 1.53 +5 -5 hugs98/src/machdep.c 1.75 +18 -16 hugs98/src/static.c 1.7 +2 -2 hugs98/tests/Readme 1.2 +5 -0 hugs98/tests/config.in 1.4 +4 -8 hugs98/tests/testScript 1.2 +17 -15 hugs98/tests/ffi/FileIO.hs 1.2 +9 -11 hugs98/tests/ffi/FileIO.output 1.2 +9 -17 hugs98/tests/ffi/Forms.hs 1.2 +2 -2 hugs98/tests/ffi/Sin.hs 1.2 +16 -16 hugs98/tests/ffi/Types.hs 1.2 +2 -1 hugs98/tests/ffi/forms_aux.h 1.2 +19 -3 hugs98/tests/ffi/types_aux.c

1 0

cvs commit: hugs98/src static.c
by Sigbjorn Finne 21 Jun '02

21 Jun '02

sof 2002/06/21 13:53:33 PDT Modified files: src static.c Log: checkExport: for Ty(..), add the in-scope constructors onto the export list also Revision Changes Path 1.74 +6 -2 hugs98/src/static.c

1 0