axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] Re: [Gcl-devel] Possible GCL 2.6.7 for axiom


From: Camm Maguire
Subject: [Axiom-developer] Re: [Gcl-devel] Possible GCL 2.6.7 for axiom
Date: 28 Apr 2005 19:09:35 -0400
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

Greetings!

"Mike Thomas" <address@hidden> writes:

> Hi Camm.
> 
> | > | 1) reenable run-process
> 
> I checked into HEAD yesterday a fix which works with the test snippet you
> provided in a separate email to Tim.  I note that the interface (on Unix
> also I believe) is not robust in the sense that it doesn't check the type or
> number of arguments to the function, but otherwise it looks fine.
> 

Thanks!  Had forgotten there was so much difference with mingw here,
but at least this appears to imply that there is some windows version
of fork() :-).  At some point, perhaps we can boil the ifdefs down to
a minimal set of primitives.

> 
> | >   5) Something has recently broken the path handling on Windows
> | so the CVS
> | > HEAD GCL build lisp compilation fails.
> | >
> |
> | OK, more detail here would be helpful, but this is not relevant to
> | 2.6.x for axiom, no?
> 
> No, but for the record it's caused by the following change:
> 
> http://savannah.gnu.org/cgi-bin/viewcvs/gcl/gcl/pcl/defsys.lisp.diff?r1=1.4&;
> r2=1.5

OK, this was along the path of supporting unique init function names
for our profiling users.  When gprof is enabled, the patch (yet
committed in entirety) will place the full filename in the C init
function name, so that ld will not complain of duplication.  This
obviously poses a challenge for the likes of unixport/sys_foo.c in
attempting to find the init function name, so I was trying to
centralize all GCL supplied system files with a pathname-directory of
nil passed to compile file.  This boils down to the current directory
on Linux, but I guess not on Windows.  Can you see why not?  Is there
anyway to make this behavior identical?

> 
> the purpose of which escapes me.  The problem is that on Windows the name of
> the file being passed to COMPILE-FILE in "pcl/defsys.lisp" is incomplete
> when that new version of the file is used.
> 
> Reversing that change then leads to a further problem which I have yet to
> investigate.

Thanks as always for your tremendous contributions!

Take care,

> 
> Cheers
> 
> Mike Thomas.
> 
> 
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah




reply via email to

[Prev in Thread] Current Thread [Next in Thread]