Well... I haven't actually tried to _use_ FFI, I was just getting everything compiled. :) I'm pretty sure that the darcs patch I posted contains the right fix for a Makefile problem I didn't explicitly mention, namely, giving it the right path to find some of the files. YHC-hat was also having some issues, so I just commented it out since I'm not worried about it at the moment. (Though, I don't know what YHC-Hat has to do with FFI makefiles.) Best, Leon On 6/2/06, Thomas Shackell wrote:

Hi Leon,

The bug is likely in the libffi which I lifted from redhat.org, you'd hope it might be correct but I guess you can't have everything ;-)

I'll have a go at fixing it :-)

Thanks

Tom

Leon Smith wrote:

...

Hello all. I'm the student working on the typechecker for Google's SoC.

I ran into a few build problems with the repository. I've attached a patch for them, though two problems are not covered in this patch. These problems both concern

yhc-devel/src/runtime/libtool

which is a generated script. First, something tries to run it directly, without running sh on it. Second, the configuration script does not set the variable $max_cmd_len on my platform (Darwin). Both these cause the build process to fail.

I worked around these issues by setting the executable bit and editing the generated script myself to set the variable. However the true fixes are beyond my expertise.

Congrats on self hosting. I did download and try out YHC sometime in January. It compiled without problems, but I didn't get to use it. I don't remember what issues I was having with YHC and the program I was writing at the time.

Best Wishes, Leon

------------------------------------------------------------------------

_______________________________________________ Yhc mailing list Yhc@haskell.org http://www.haskell.org//mailman/listinfo/yhc

_______________________________________________ Yhc mailing list Yhc@haskell.org http://www.haskell.org//mailman/listinfo/yhc