[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: GNU AutoGen Version 5.11.6/AutoOpts Version 34.0
From: |
Ludovic Courtès |
Subject: |
Re: GNU AutoGen Version 5.11.6/AutoOpts Version 34.0 |
Date: |
Tue, 22 Feb 2011 21:42:44 +0100 |
User-agent: |
Gnus/5.110011 (No Gnus v0.11) Emacs/23.2 (gnu/linux) |
Hi Bruce,
Bruce Korb <address@hidden> writes:
> On 02/22/11 05:27, Ludovic Courtès wrote:
>> It seems there’s still an issue:
>>
>> --8<---------------cut here---------------start------------->8---
>> gcc -std=gnu99 -DHAVE_CONFIG_H -I. -I.. -I.. -I../autoopts -pthread
>> -I/nix/store/x8n3bs4jyy1im0jw1xikkqhfijgm36hz-libunistring-0.9.3/include
>> -I/nix/store/7z0h50prh7iyqwylr9sikgkhn4kqsh2q-guile-2.0.0/include/guile/2.0
>> -I/nix/store/j78gdcwhr5q38c2f5ipcr2h4xfnjr0gj-boehm-gc-7.2pre20110122/include
>> -g -O2 -c -o autogen-ag.o `test -f 'ag.c' || echo './'`ag.c
>> In file included from
>> /nix/store/7z0h50prh7iyqwylr9sikgkhn4kqsh2q-guile-2.0.0/include/guile/2.0/libguile.h:47:0,
>> from autogen.h:35,
>> from ag.c:5:
>> /nix/store/7z0h50prh7iyqwylr9sikgkhn4kqsh2q-guile-2.0.0/include/guile/2.0/libguile/eval.h:82:0:
>> warning: "scm_primitive_eval_x" redefined
>> ../config.h:530:0: note: this is the location of the previous definition
>
> This gets inserted into config.h, presumably, when trying to use
> the function fails in a config script macro:
>
>> AC_DEFUN([AG_LINK_EVAL_STRING],[
>> AC_MSG_CHECKING([whether scm_primitive_eval_x links])
>> AC_CACHE_VAL([ag_cv_link_eval_string],[
>> ag_save_CPPFLAGS="${CPPFLAGS}"
>> CPPFLAGS="${ag_cv_with_libguile_cflags} ${CPPFLAGS}"
>> ag_save_LIBS="${LIBS}"
>> LIBS="${ag_cv_with_libguile_libs} ${LIBS}"
>> AC_TRY_LINK(address@hidden:@include <guile/gh.h>
>> @%:@include <libguile.h>],
>> [SCM res = scm_primitive_eval_x( SCM_UNDEFINED );],
>> [ag_cv_link_eval_string=yes],[ag_cv_link_eval_string=no]
>> ) # end of TRY_LINK
>> CPPFLAGS="${ag_save_CPPFLAGS}"
>> LIBS="${ag_save_LIBS}"
>> ]) # end of AC_CACHE_VAL for ag_cv_link_eval_string
>> AC_MSG_RESULT([${ag_cv_link_eval_string}])
>> if test "X${ag_cv_link_eval_string}" = Xno
>> then
>> AC_DEFINE([scm_primitive_eval_x], [scm_eval_x],
>> [Define this if no scm_primitive_eval_x])
>> AC_DEFINE([scm_t_port], [scm_port], [Define this if no scm_t_port])
>> fi
>>
>> ]) # end of AC_DEFUN of AG_LINK_EVAL_STRING
>
> At this point, I *THINK* I can presume Guile 1.4 to be dead and gone.
> Thus I *THINK* I can presume that a working libguile has scm_primitive_eval_x.
> Consequently, I can simply remove this test and fix this part of the problem.
> Confirmation would be appreciated. Still, this macro ought to have worked
> correctly. Oh, wait, maybe the "#include <guile/gh.h>" messed things up?
> Is that true?
Yes, GH was deprecated in 1.8 (and 1.6, IIRC), and it’s no longer in 2.0.
>> In file included from ag.c:17:0:
>> expState.c: In function 'find_entry_value':
>> expState.c:155:18: error: 'scm_listofnull' undeclared (first use in this
>> function)
>
> So, what happened to scm_listofnull? Is this another interface change?
Oops, yes. It apparently vanished unintentionally as a side-effect of
e7efe8e793fa51ea898aea4477939c598b3e9fac in Dec. 2009, which is why NEWS
doesn’t mention it.
All I can recommend is another AC_LINK_IFELSE, or better yet, using your
own variable instead:
scm_listofnull = scm_list_1 (SCM_EOL);
Hope this helps,
Ludo’.