[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: ‘iconv’ fails to build on NetBSD 5.1
From: |
Ludovic Courtès |
Subject: |
Re: ‘iconv’ fails to build on NetBSD 5.1 |
Date: |
Thu, 07 Jul 2011 15:47:44 +0200 |
User-agent: |
Gnus/5.110017 (No Gnus v0.17) Emacs/24.0.50 (gnu/linux) |
Hi Bruno,
Bruno Haible <address@hidden> skribis:
>> The ‘iconv’ modules fails to build on NetBSD 5.1 (see gcc70 of the GCC
>> Compile Farm):
>
> Can you show the gnulib-tool invocations that you used when builting
> the tarball of your package? In particular, which of the modules
> 'iconv', 'iconv-h', 'iconv_open' are used, and which --avoid options
> are used?
>From ‘gnulib-cache.m4’:
# gnulib-tool --import --dir=. --lib=libgnu --source-base=lib --m4-base=m4
--doc-base=doc --tests-base=tests --aux-dir=build-aux --lgpl=3
--no-conditional-dependencies --libtool --macro-prefix=gl --no-vc-files accept
alignof alloca-opt announce-gen autobuild bind byteswap canonicalize-lgpl ceil
close connect duplocale environ extensions flock floor fpieee frexp full-read
full-write func gendocs getaddrinfo getpeername getsockname getsockopt
git-version-gen gitlog-to-changelog gnu-web-doc-update gnupload havelib
iconv_open-utf inet_ntop inet_pton isinf isnan ldexp lib-symbol-versions
lib-symbol-visibility libunistring listen locale log1p maintainer-makefile
malloc-gnu malloca nproc open pipe2 putenv recv recvfrom rename send sendto
setsockopt shutdown socket stat-time stdlib strftime striconveh string sys_stat
trunc verify vsnprintf warnings wchar
So it’s only ‘iconv_open-utf’, but it should drag the other iconv*
modules AIUI.
> This shouldn't be either. Can you 'grep serial aclocal.m4' ?
# iconv.m4 serial 11 (gettext-0.18.1)
This is older than the one in the Gnulib repo I used.
(Time passes...)
The bug originates in Nixpkgs/NixOS, which has a wrapper around
‘aclocal’ to make it honor $ACLOCAL_PATH (this is necessary with Nix
since each package is installed in its own path.) The problem is that
the wrapper would *prepend* the -I flags corresponding to $ACLOCAL_PATH
instead of appending them, hence the bug.
Sorry for the noise!
Thanks,
Ludo’.