On 07/06/2012 10:26 AM, Stefano Lattarini wrote:
> Hi Diego.
>
> On 07/06/2012 12:10 AM, Diego Elio Pettenò wrote:
>> Il 05/07/2012 11:26, Stefano Lattarini ha scritto:
>>> How so? Removal of $(mkdir_p) is only planned for Automake 1.13, that is
>>> still unreleased.
>>
>> Ehm Stefano, that's definitely not the case, I've been hitting that
>> issue in Gentoo in many packages.
>>
>> https://bugs.gentoo.org/show_bug.cgi?id=automake-1.12
>>
>> mkdir_p is gone in automake-1.12.1 for what we're concerned — I also
>> listed it in the 1.12 instead of "planned for 1.13" in my own guide
>> http://goo.gl/bda6l as I thought it was intentional.
>>
> Ouch, I can reproduce this :-/ The issue is that the call to
> AM_PROG_MKDIR_P in AM_INIT_AUTOMAKE has been substituted by a
> call to AC_PROG_MKDIR_P, which doesn't define $(mkdir_p). This
> was an unintended regression.
>
> I will soon push the attached patches to take care of the problem.
> Note that they only re-introduce the $(mkdir_p) variable (because
> is very easy to do so), but not the @mkdir_p@ substitution. Anyone
> knows if this can create problems in practice?
>
Grrr, the second patch doesn't work on systems (like NetBSD 5.1) where
"mkdir -p" is not deemed good enough, and "install-sh -d" is used
instead:
./configure
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... ./install-sh -c -d
checking for gawk... no
checking for mawk... no
checking for nawk... no
checking for awk... awk
checking whether make sets $(MAKE)... yes
checking that generated files are newer than configure... done
configure: creating ./config.status
config.status: creating Makefile
+ make test
./install-sh -c -d . dir1/dir2
(mkdir x && cd x && ./install-sh -c -d .. y/z ../foo)
sh: ./install-sh: not found
*** Error code 127
Stop.
make: stopped in /home/slattarini/tmp/automake-1.12.1a/t/mkdir_p.dir
+ exit_status=1
Anyway, I've verified that this is *not* a regression (at least w.r.t
Automake 1.11.2), so I'll simply adjust the tests accordingly (using
$(mkdir_p) in a $(SUBDIR) Makefile, instead of fooling around with
chdir in make recipes).
I'll post the adjusted patch soon.
Regards,
Stefano
