August 2005 - C2hs - Haskell.org

Re: c2hs feature request: {# alignof #} hook
by Manuel M T Chakravarty 21 Aug '05

21 Aug '05

Duncan Coutts: > Writing Storable instances would be more convenient if c2hs supported an > {# alignof #} hook, just like it currently supports the {# sizeof #} > hook. > > I think the syntax would be exactly the same, ie it'd just take a C type > name. Good idea. Manuel

1 0

Gtk+cairo working :-)
by Paolo Martini 18 Aug '05

18 Aug '05

-----BEGIN PGP SIGNED MESSAGE----- Hash: RIPEMD160 Hello, I'm proud to announce that we got Gtk integration initially working :-) I don't know if you follow the site/blog[1] but there we go: http://haskell.org/gtk2hs/archives/2005/08/17/code/#comment-35 Thanks much people, you guys rock. Cheers, Paolo. P.S. the cairobindings darcs repo is now ``outdated'' by the gtk2hs one I'm setting up as soon as an issue with Gtk2Hs's version of c2hs will be resolved. I'll blog-post about that. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDBICPotUQQOWB3WQRA8SlAJ0d4mecBKH5JJH2ch6IT1/4fExK0ACeK4E+ WW6WDY1kk3qeqcT0fgsPh38= =4oRp -----END PGP SIGNATURE-----

1 1

Re: [C2hs] Re: [Gtk2hs-devel] hsffig & c2hs
by Manuel M T Chakravarty 15 Aug '05

15 Aug '05

Duncan Coutts: > On Sun, 2005-08-07 at 22:06 +1000, Manuel M T Chakravarty wrote: > > > * There will be a command line option that permits to override that > > choice and to explicitly specify the platform for which bindings are > > generated. C2hs will have a database of platforms it knows about and > > all we need to do is to ensure that Visual-C gets an entry in that > > database. > > That sounds like a sensible approach. The latest release already has that database in "C2HSConfig.hs.in". > > All I need is a for somebody to compile and run the attached C functions > > in Visual C and let me know the results. > > I'll try and give that a go some time. (And then check to see if it > passes the Gtk+ ABI init check) Great - thanks! Manuel

1 0

Questions
by Paolo Martini 15 Aug '05

15 Aug '05

Hello, First of all, thanks much for the good tool you're providing :-) I have some questions. Why doesn't {#enum ..#} hooks define the default marshallers to be cFromEnum/cToEnum? Furthermore, is it possible to define the marshallers functions for a give type? (IMO it would make the .chs files more readable) If not, would it be possible to extend the syntax, maybe adding "... inmarsh outmarsh#}" at the end of the hooks? For example I'm using "id" as the marshaller for {#pointer ..#} hooks, (in {#fun name {id `pointer'} -> `pointer' id#}) is it correct, or am I missing something? I can't find an example of a C struct type which isn't defined as a pointer. How should one go to bind structures like this? typedef struct { unsigned long index; double x; double y; } cairo_glyph_t; And then, how to pass them to functions getting a *cairo_glyph_t? Moreover, what if they are getting an array of them? (e.g. void cairo_show_glyphs (cairo_t *cr, cairo_glyph_t *glyphs, int num_glyphs); Thank you, Paolo. --

2 1

c2hs-config
by Paolo Martini 14 Aug '05

14 Aug '05

There seem to be no c2hs-config in the source package, is the documentation[1] outdated, and it isn't needed anymore for some reason, or its missing have some other meaning? Also, I can't find how to register the c2hs package into GHC, any clue? Thank you, Paolo

2 1

RE: [C2hs] ANN: C->Haskell version 0.14.3 "Travelling Lightly"
by Simon Marlow 14 Aug '05

14 Aug '05

On 14 August 2005 14:53, Paolo Martini wrote: > The problem is that c2hs.cabal is missing the field: "Build-Depends: > base,haskell98" (right after "Category:", preserving the newline > before "Executables:") > > I don't exacly know if it is a workaround or a solution to a bug, but > it does make c2hs compile -- hooray :-) That's the right solution. Cheers, Simon

2 1

RE: [C2hs] ANN: C->Haskell version 0.14.3 "Travelling Lightly"
by Simon Marlow 14 Aug '05

14 Aug '05

On 12 August 2005 16:06, Manuel M T Chakravarty wrote: > Gour: >> Manuel M T Chakravarty (chak(a)cse.unsw.edu.au) wrote: >> >>> CForeign is a module from the FFI standard. What version of GHC >>> are you using? >> >> The Glorious Glasgow Haskell Compilation System, version >> 6.4.1.20050801 > > This must be a 6.4.1 issue. In 6.4, CForeign is visible, but in > 6.4.1. it doesn't seem to be visible anymore. Hmm, shoudln't > patchlevel releases not change the interface? CForeign is still in 6.4.1. What problem are you having? Cheers, Simon

2 1

Beijing Chinese Translation
by Beijing Chinese Translation 12 Aug '05

12 Aug '05

1 0

ANN: C->Haskell version 0.14.3 "Travelling Lightly"
by Manuel M T Chakravarty 12 Aug '05

12 Aug '05

I like to announce version 0.14.3 of C->Haskell, which brings the following advances over 0.14.1: * gcc's asm construct is supported, which is apparently important for some libraries on Mac OS X (thanks to Duncan Coutts for the patch); * C->Haskell supports cross compilation (see details below); [NB: C->Haskell now generates completely portable Haskell code (needs only H98 + FFI) and can target C compilers on platforms that cannot build C->Haskell itself; eg, PalmOS. In particular, C->Haskell generated bindings are suitable for nhc98 and Hugs, too.] * hierarchical modules syntax can be used in import hooks (again thanks to Duncan for the patch); and * two new built-in translation functions "upcaseFirstLetter" and "downcaseFirstLetter" are supported to generated suitable Haskell names from C names (eg, for variants of enums). Source code is available as a tar ball from http://www.cse.unsw.edu.au/~chak/haskell/c2hs/ or via Darcs (binary packages may appear later). Happy Hacking, Manuel -=- Cross compilation -=- C->Haskell needs to take system-specific parameters of the C compiler used to compile the libraries for which an interface is generated into account (most notably when bitfields are used in this library). By default C->Haskell uses the conventions of the C compiler that was used to build C->Haskell itself. From this version, C->Haskell accepts the --platform=PLATFORM flag that enables the explicit specification of the target platform. This works only for platforms that are in C->Haskell's platform specification database. At the moment, that's only i686-pc-linux-gnu and m68k-palmos. However, it is very easy to determine the platform specification for any platform on which you can build C->Haskell: Just invoke it as `c2hs --version' and it will tell you. If you send me the output of `c2hs --version', I'll happily add it to the database. (If you'd like to target a platform on which you cannot build C->Haskell, such as an embedded system, let me know.)

2 6

ANN: C->Haskell 0.14.2
by Manuel M T Chakravarty 07 Aug '05

07 Aug '05

I just put version 0.14.2 of C->Haskell at http://www.cse.unsw.edu.au/~chak/haskell/c2hs/ It includes Duncan Coutts' asm and hierarchical module patch and also fixes the way the new shell wrappers pass command line arguments. Manuel

1 0