Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Help install gsll/fsbv on OSX

10 views
Skip to first unread message

Leo

unread,
Jan 16, 2010, 11:01:29 PM1/16/10
to
Hello,

I am in the process of installing gsll and stuck with fsbv.

Following the readme.html in fsbv source dir, I run 'pkg-config
--cflags-only-I libffi' and it returns

-I/usr/local/opensource/libffi-3.0.9/include

The full installation looks like this

/usr/local/opensource/libffi-3.0.9/
|-include/
| +-ffi/
|-lib/
| +-pkgconfig/
+-share/
|-info/
+-man/
+-man3/

then I modify libffi-unix.lisp to have:

@@ -11,7 +11,7 @@
;; When installed through Mac Ports, FSBV include files
;; will be found in /opt/local/include.
#+darwin
-(cc-flags "-I/opt/local/include/")
+(cc-flags "-I/usr/local/opensource/libffi-3.0.9/include/")

#+darwin
(include "ffi/ffi.h")


Finally I try loading fsbv from slime with CCL and get the following
error: http://openpaste.org/en/18921/

Any idea how to resolve that? Thanks.

Leo

D Herring

unread,
Jan 17, 2010, 12:25:10 AM1/17/10
to
Leo wrote:
> I am in the process of installing gsll and stuck with fsbv.
>
> Following the readme.html in fsbv source dir, I run 'pkg-config
> --cflags-only-I libffi' and it returns
>
> -I/usr/local/opensource/libffi-3.0.9/include
...

> then I modify libffi-unix.lisp to have:
>
> @@ -11,7 +11,7 @@
> ;; When installed through Mac Ports, FSBV include files
> ;; will be found in /opt/local/include.
> #+darwin
> -(cc-flags "-I/opt/local/include/")
> +(cc-flags "-I/usr/local/opensource/libffi-3.0.9/include/")
>
> #+darwin
> (include "ffi/ffi.h")
>
>
> Finally I try loading fsbv from slime with CCL and get the following
> error: http://openpaste.org/en/18921/

The important part is
# Unknown Grovel syntax: CFFI-GROVEL::CC-FLAGS
# [Condition of type SIMPLE-ERROR]

[BTW see http://paste.lisp.org/]


> Any idea how to resolve that? Thanks.

Are you using the latest CFFI? It looks like cc-flags was introduced
around 2009-08-23 (after the latest tarball release).

If you don't want to use darcs to get it from the official site, you
can grab a snapshot from my git mirror.
http://git.libcl.com/upstream/?p=cffi.git;a=summary
http://git.libcl.com/upstream/?p=cffi.git;a=snapshot;h=2fe41d4c409e7a697c4ca15f4fd60487a6cc97e1;sf=tgz

- Daniel

Leo

unread,
Jan 17, 2010, 6:57:18 AM1/17/10
to
D Herring <dher...@at.tentpost.dot.com> writes:

> Are you using the latest CFFI? It looks like cc-flags was introduced
> around 2009-08-23 (after the latest tarball release).
>
> If you don't want to use darcs to get it from the official site, you
> can grab a snapshot from my git mirror.
> http://git.libcl.com/upstream/?p=cffi.git;a=summary
> http://git.libcl.com/upstream/?p=cffi.git;a=snapshot;h=2fe41d4c409e7a697c4ca15f4fd60487a6cc97e1;sf=tgz

Thanks. This was the problem. I used the latest release but actually a
recent snapshot is required. GSLL now loads and runs.

I am doing (lisp-unit:run-tests) but it produce an error logged here:

http://paste.lisp.org/display/93531

so the unit-testing didn't finish. Any idea what's wrong here?

Thanks.

Leo

D Herring

unread,
Jan 17, 2010, 9:59:16 AM1/17/10
to
Leo wrote:

> I am doing (lisp-unit:run-tests) but it produce an error logged here:
>
> http://paste.lisp.org/display/93531
>
> so the unit-testing didn't finish. Any idea what's wrong here?

That one's obscure to me. It looks like there's a structure (named
COMBINATION ?) with a c-pointer member, and the member was never
initialized. You'll have to look at the code or ask on the GSLL
lists. They may want you to reproduce the backtrace and select
"--more--" at the end; the good stuff may have been cut off at frame 20.

If you're lucky, the test might have printed an error/warning message
before this point.

Later,
Daniel

Leo

unread,
Jan 17, 2010, 7:17:33 PM1/17/10
to
D Herring <dher...@at.tentpost.dot.com> writes:

Thanks. Coincidentally the developer of gsll noted the same problem
this morning. It was fixed in git just now. The summary of the unit
testing is:

TOTAL: 1455 assertions passed, 46 failed, 0 execution errors.

which seems too high.

Leo

0 new messages