[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: CVS automake testsuite failures under Tru64 unix
From: |
Alexandre Duret-Lutz |
Subject: |
Re: CVS automake testsuite failures under Tru64 unix |
Date: |
20 May 2002 22:59:02 +0200 |
User-agent: |
Gnus/5.0808 (Gnus v5.8.8) Emacs/20.7 |
>>> "adl" == Alexandre Duret-Lutz <address@hidden> writes:
>>> "Akim" == Akim Demaille <address@hidden> writes:
adl> [...]
>>> > Nicolas> AC_INIT(sinclude, 0.0)
adl> [...]
Akim> I have not read the whole thread, but in the case of a non literal
Akim> (i.e., not an Sh-free expression), it shoudn't matter. In the case of
Akim> an M4 literal, be sure to *double* quote, as you don't want sinclude
Akim> to be evaluated.
Akim> AC_INIT([[sinclude]], ...
Akim> AC_INIT([[m4_define]], ...
adl> Thanks. I'm checking this in.
[...]
adl> -AC_INIT($me, 1.0)
adl> +AC_INIT([[$me]], [1.0])
Holy crap! I commited this before waiting for the Automake
test suite to complete (it was so obvious it would work...).
Now we have a handful of failures due to
AC_INIT([[foobar]], [1.0])
being transformed into
PACKAGE_NAME='foobar'
PACKAGE_TARNAME='-foobar-'
PACKAGE_VERSION='1.0'
PACKAGE_STRING='foobar 1.0'
PACKAGE_BUGREPORT=''
hence the tarball is named `-foobar--1.0.tar.gz' instead
of `foobar-1.0.tar.gz'.
Apart from renaming the problematic test case, the only other
idea I have is to uses AC_INIT's fourth parameter:
AC_INIT([[$me]], [1.0], address@hidden, [[$me]])
this seems to works. However we would have to raise the
Autoconf version requirement.
Another idea?
--
Alexandre Duret-Lutz