> 2) I'd like to make sure that one can build cleanly from the soon to
> be release 2.5.4 release. Could you please checkout version 2.5.4
> from cvs, and post the minimal steps required to get your build?
> (cvs -d:pserver: address@hidden :/cvsroot/gcl co -r
> Version_2_5_4 gcl)
Version 2.5.4 compiles fine. I've tried using the environment variables
C_INCLUDE_PATH and LIBRARY_PATH the way I would for CPPFLAGS or LDFLAGS
for instance, but without success. How do you use them ? When I know
how to use them, I shall post the minimal steps required to get my
build.
> 5) Can you please check that you can build an ansi image too
> (--enable-ansi), cd into ansi-tests, execute
> ../unixport/saved_ansi_gcl, (load "gclload.lsp") and report your
> results?
Ok, here are the results for this. Note that I had to preallocate
memory to pass the tests. (I've merely followed the instructions in
the readme file to enlarge the initial memory allocation.)
By the way, I've spotted a possible typo in sfasli.c (there's not
a single HAVE_ALLOC except one in sfasli.c, but there are plenty
of HAVE_ALLOCA elsewhere) :
--- sfasli.c.old Mon Aug 18 22:37:05 2003
+++ sfasli.c Mon Aug 18 22:37:07 2003
@@ -45,7 +45,7 @@
FEerror("Cannot add self symbols\n",0);
if ((u=bfd_get_symtab_upper_bound(bself))<0)
FEerror("Cannot get self's symtab upper bound",0);
-#ifdef HAVE_ALLOC
+#ifdef HAVE_ALLOCA
q=(asymbol **)alloca(u);
#else
q=(asymbol **)malloc(u);