GNU bug report logs - #11438
make installcheck

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 11438 in the body.
You can then email your comments to 11438 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Jeff Janes <jeff.janes <at> gmail.com>
To: bug-coreutils <at> gnu.org
Subject: make installcheck
Date: Tue, 8 May 2012 12:23:26 -0700

Section 5 of the "INSTALL" file for coreutils says that "make
installcheck" will repeat the self-checks, but using the binaries in
their final installed locations.

I cannot figure out what "make installcheck" is doing, but surely it
is not doing that.  It doesn't seem to even visit the final installed
location of the binaries at all.

I tried this against versions 8.16 (tarball), 8.14 (tarball), and the
git HEAD, on both openSUSE 12.1 and on Cygwin.

I do all the work as non-root, and for ./configure, I always specified
a --prefix that specifies a not-currently-existent directory name
within my home directory (and no other options)

Thanks,

Jeff

Message #8 received at 11438 <at> debbugs.gnu.org (full text, mbox):

From: Jim Meyering <jim <at> meyering.net>
To: Jeff Janes <jeff.janes <at> gmail.com>
Cc: 11438 <at> debbugs.gnu.org
Subject: Re: bug#11438: make installcheck
Date: Tue, 08 May 2012 21:46:03 +0200

Jeff Janes wrote:
> Section 5 of the "INSTALL" file for coreutils says that "make
> installcheck" will repeat the self-checks, but using the binaries in
> their final installed locations.
>
> I cannot figure out what "make installcheck" is doing, but surely it
> is not doing that.  It doesn't seem to even visit the final installed
> location of the binaries at all.
>
> I tried this against versions 8.16 (tarball), 8.14 (tarball), and the
> git HEAD, on both openSUSE 12.1 and on Cygwin.
>
> I do all the work as non-root, and for ./configure, I always specified
> a --prefix that specifies a not-currently-existent directory name
> within my home directory (and no other options)

Thanks for the report.

However, "make installcheck" is a no-op for most packages.  Besides,
it would not be useful to coreutils, unless you use a non-default
configuration that would install su, which is normally set-UID (su
is not normally installed).  In other words, "make check" is just fine,
since none of these programs act differently depending on whether they
are installed or in the build directory.

Message #11 received at 11438 <at> debbugs.gnu.org (full text, mbox):

From: Jeff Janes <jeff.janes <at> gmail.com>
To: Jim Meyering <jim <at> meyering.net>
Cc: 11438 <at> debbugs.gnu.org
Subject: Re: bug#11438: make installcheck
Date: Tue, 8 May 2012 14:59:17 -0700

On Tue, May 8, 2012 at 12:46 PM, Jim Meyering <jim <at> meyering.net> wrote:
> Jeff Janes wrote:
>> Section 5 of the "INSTALL" file for coreutils says that "make
>> installcheck" will repeat the self-checks, but using the binaries in
>> their final installed locations.
>>
>> I cannot figure out what "make installcheck" is doing, but surely it
>> is not doing that.  It doesn't seem to even visit the final installed
>> location of the binaries at all.
>>
>> I tried this against versions 8.16 (tarball), 8.14 (tarball), and the
>> git HEAD, on both openSUSE 12.1 and on Cygwin.
>>
>> I do all the work as non-root, and for ./configure, I always specified
>> a --prefix that specifies a not-currently-existent directory name
>> within my home directory (and no other options)
>
> Thanks for the report.
>
> However, "make installcheck" is a no-op for most packages.  Besides,
> it would not be useful to coreutils, unless you use a non-default
> configuration that would install su, which is normally set-UID (su
> is not normally installed).  In other words, "make check" is just fine,
> since none of these programs act differently depending on whether they
> are installed or in the build directory.

OK, thanks for the explanation.  I thought that if a package did not
implement an option it would give me a "No rule to make target"
message, rather than being a no-op.  But I guess that that would
complicate "make distcheck".

Cheers,

Jeff

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.