Re: Compile issues for 1.8.0 on Solaris 10

[Neil Jerram]

Charles Gagnon <address@hidden> writes:

> Those are compile issues but not really deve issues. Maybe just a
> few things to fix in the configure.in scripts.
>
> I got guile-1.8.0 to compile on my Solaris 10 box with gcc 3.4.3 but
> I had to fix a few things first:
>
> 1. I had to replace the "-pthread" option in all the Makefile with
>    "-pthreads". I think this is just a Linux vs Solaris thing, one
>    gcc uses pthread and the other pthreads. -pthread reports invalid
>    option.

Thanks.  This seems to be an issue in our ACX_PTHREAD definition (in
acinclude.m4), which has been fixed in the latest version on
http://autoconf-archive.cryp.to, so I'll update our acinclude.m4 to
use the latest version.

> 2. per BIll Schottstaet (reported for 1.7.91), I changed filesys.c
>    around line 860 cause I was getting an error in function
>    scm_readdir:
>
>    I changed filesys.c line 860 (or thereabouts) from:
>
>     union {
>       struct dirent ent;
>       char pad1 [sizeof(struct dirent) + NAME_MAX];
>       char pad2 [offsetof (struct dirent, d_name) + NAME_MAX + 1];
>     } u;
>     SCM_SYSCALL (readdir_r ((DIR *) SCM_CELL_WORD_1 (port), &u.ent, &rdent));
>
>     to:
>
>     struct dirent ent;
>     SCM_SYSCALL (readdir_r ((DIR *) SCM_CELL_WORD_1 (port), &ent));

This code has since changed in CVS (both HEAD and the 1.8.x branch),
so I think that means this problem has been fixed and will be OK in
the next 1.8.x release.

> 3.  I also removed -Werror from libguile/Makefile cause I was getting:
>
>     In file included from eval.c:5969:
>     eval.c: In function `deval':
>     eval.c:3248: warning: implicit declaration of function `alloca'
>     make[2]: *** [libguile_la-eval.lo] Error 1
>     make[2]: Leaving directory `/home/bil/test/guile-1.7.91/libguile'
>     make[1]: *** [all-recursive] Error 1

Where is alloca() declared in Solaris 10?  Is it in <malloc.h>?  If it
is, we already have the fix for this in place, so it will be in the
next release.

> Thanks.

Many thanks for your report.

     Neil