GNU bug report logs -
#19222
Some test failures with LibTool-2.4.4 on non GNU/Linux systems
Previous Next
To reply to this bug, email your comments to 19222 AT debbugs.gnu.org.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-libtool <at> gnu.org:
bug#19222; Package
libtool.
(Sun, 30 Nov 2014 01:57:02 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
Assaf Gordon <assafgordon <at> gmail.com>:
New bug report received and forwarded. Copy sent to
bug-libtool <at> gnu.org.
(Sun, 30 Nov 2014 01:57:03 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
Hello,
Trying the recently released libtool-2.4.4 on non gnu/linux systems give some failures.
=== FreeBSD 10.0 ====
95: versioning FAILED (versioning.at:153)
169: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:48)
==== NetBSD 6.1.4 =====
55: debug tracing FAILED (help.at:159)
65: Link order test FAILED (link-order.at:89)
66: Link order of deplibs FAILED (link-order2.at:138)
70: static linking flags for programs FAILED (static.at:196)
86: deplib in subdir FAILED (deplib-in-subdir.at:126)
95: versioning FAILED (versioning.at:179)
97: DESTDIR with in-package deplibs FAILED (destdir.at:101)
143: C++ exception handling FAILED (exceptions.at:392)
164: deplibs without file command FAILED (deplibs-mingw.at:64)
169: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:48)
==== OpenBSD 5.6 ====
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 26 27 28 29 30 31 33 34 35 36
38 39 40 41 42 48 49 50 51 52 98 99 100 101 102 117 118 119 120 121
122 123 124 125 126 127 128 129 133 134 135 144 145 146 148 149 150
151 161
==== Minix R3.3.0 ===
55 58 65 66 70 78 86 95 97 106 113 143 147 156 157 164 168
====
The logs are large (including the 'testsuite.dir'), so attached are links:
http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-freebsd10-tests.tar.gz
http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-MinixR330-tests.tar.gz
http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-netbsd614.tar.gz
http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-openbsd56.tar.gz
No test failures on GNU-Hurd 0.5 (i386/Debian) and Mac OS X 10.9.5.
Also,
On OpenBSD 5.6, for some reason, after running "./configure", "make" tried to re-build some files using autoconf/automake - which required automake/autoconf despite this being a release tarball. Not sure if this is normal or not.
Log:
http://files.housegordon.org/libtool-2.4.4/openbsd56-configure-log.txt
Regards,
- Assaf
Information forwarded
to
bug-libtool <at> gnu.org:
bug#19222; Package
libtool.
(Wed, 03 Dec 2014 22:06:02 GMT)
Full text and
rfc822 format available.
Message #8 received at 19222 <at> debbugs.gnu.org (full text, mbox):
Hi Assaf,
Thanks for the reports and results.
> On Nov 30, 2014, at 1:56 AM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
>
> Hello,
>
> Trying the recently released libtool-2.4.4 on non gnu/linux systems give some failures.
>
> === FreeBSD 10.0 ====
> 95: versioning FAILED (versioning.at:153)
> 169: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:48)
> ==== NetBSD 6.1.4 =====
> 55: debug tracing FAILED (help.at:159)
> 65: Link order test FAILED (link-order.at:89)
> 66: Link order of deplibs FAILED (link-order2.at:138)
> 70: static linking flags for programs FAILED (static.at:196)
> 86: deplib in subdir FAILED (deplib-in-subdir.at:126)
> 95: versioning FAILED (versioning.at:179)
> 97: DESTDIR with in-package deplibs FAILED (destdir.at:101)
> 143: C++ exception handling FAILED (exceptions.at:392)
> 164: deplibs without file command FAILED (deplibs-mingw.at:64)
> 169: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:48)
> ==== OpenBSD 5.6 ====
> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 26 27 28 29 30 31 33 34 35 36
> 38 39 40 41 42 48 49 50 51 52 98 99 100 101 102 117 118 119 120 121
> 122 123 124 125 126 127 128 129 133 134 135 144 145 146 148 149 150
> 151 161
> ==== Minix R3.3.0 ===
> 55 58 65 66 70 78 86 95 97 106 113 143 147 156 157 164 168
> ====
>
> The logs are large (including the 'testsuite.dir'), so attached are links:
> http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-freebsd10-tests.tar.gz
> http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-MinixR330-tests.tar.gz
> http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-netbsd614.tar.gz
> http://files.housegordon.org/libtool-2.4.4/libtool-2.4.4-openbsd56.tar.gz
I can see immediately that most of the failures on your OpenBSD system are caused
by not having a GNU M4 installed on the path for libtoolize. I should have been a
lot more vocal about that change rather than hiding it in the 2.4.3 release notes,
as it has tripped a lot of people up. The latest master revision now checks for a
suitable M4 at configure time, and complains if there is nothing suitable on your
command search PATH, which I think will help.
May I ask if you would kindly rerun then test suite on that machine and reply with a
link to the updated logs, after having installed GNU M4 first of course :)
I'll look into the logs of the other failing machines in the days ahead...
> No test failures on GNU-Hurd 0.5 (i386/Debian) and Mac OS X 10.9.5.
Great! Thanks :)
> Also,
> On OpenBSD 5.6, for some reason, after running "./configure", "make" tried to re-build some files using autoconf/automake - which required automake/autoconf despite this being a release tarball. Not sure if this is normal or not.
> Log:
> http://files.housegordon.org/libtool-2.4.4/openbsd56-configure-log.txt
No that's not normal, and is usually caused either by clock-skew on a network filesystem,
low clock resolution in the build filesystem (so that tar unpacks files with the same
timestamp when in fact those files were a fraction of a second apart), or even a buggy
tar program mangling the timestamps during unpack.
Cheers,
--
Gary V. Vaughan (gary AT gnu DOT org)
Information forwarded
to
bug-libtool <at> gnu.org:
bug#19222; Package
libtool.
(Thu, 04 Dec 2014 14:01:01 GMT)
Full text and
rfc822 format available.
Message #11 received at 19222 <at> debbugs.gnu.org (full text, mbox):
Hi Assaf,
> On Nov 30, 2014, at 1:56 AM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
>
> Trying the recently released libtool-2.4.4 on non gnu/linux systems give some failures.
>
> === FreeBSD 10.0 ====
> 95: versioning FAILED (versioning.at:153)
> 169: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:48)
I think current master fixes these. Please let me know if not, as I plan to put out
another release fairly quickly for this and a couple of other issues with 2.4.4.
Cheers,
--
Gary V. Vaughan (gary AT gnu DOT org)
Information forwarded
to
bug-libtool <at> gnu.org:
bug#19222; Package
libtool.
(Thu, 04 Dec 2014 21:25:02 GMT)
Full text and
rfc822 format available.
Message #14 received at 19222 <at> debbugs.gnu.org (full text, mbox):
On 12/04/2014 09:00 AM, Gary V. Vaughan wrote:
>> On Nov 30, 2014, at 1:56 AM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
>>
>> Trying the recently released libtool-2.4.4 on non gnu/linux systems give some failures.
>>
>> === FreeBSD 10.0 ====
>> 95: versioning FAILED (versioning.at:153)
>> 169: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:48)
>
> I think current master fixes these. Please let me know if not, as I plan to put out
> another release fairly quickly for this and a couple of other issues with 2.4.4.
>
With the latest git revision ( 2.4.4.8-9f52 ), no test failures on FreeBSD 10 and FreeBSD 9.3,
when building from tarball (that is: 'git clone && bootstrap && make dist' on a GNU/Linux box, then 'configure && make && make check' on the FreeBSD VM.
- Assaf
Information forwarded
to
bug-libtool <at> gnu.org:
bug#19222; Package
libtool.
(Thu, 04 Dec 2014 21:34:02 GMT)
Full text and
rfc822 format available.
Message #17 received at 19222 <at> debbugs.gnu.org (full text, mbox):
On 12/03/2014 05:05 PM, Gary V. Vaughan wrote:
>> On Nov 30, 2014, at 1:56 AM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
<...>
> I can see immediately that most of the failures on your OpenBSD system are caused
> by not having a GNU M4 installed on the path for libtoolize. I should have been a
> lot more vocal about that change rather than hiding it in the 2.4.3 release notes,
> as it has tripped a lot of people up. The latest master revision now checks for a
> suitable M4 at configure time, and complains if there is nothing suitable on your
> command search PATH, which I think will help.
>
> May I ask if you would kindly rerun then test suite on that machine and reply with a
> link to the updated logs, after having installed GNU M4 first of course :)
On OpenBSD 5.6, with GNU M4 installed as "/usr/local/bin/gm4",
-1-
Building the latest git ( 2.4.4.8-9f52 ) from a tarball (that is, "git clone && bootstrap && configure && make dist" on a GNU/Linux box, then "configure && make && make check" on the OpenBSD VM), two tests fail:
===
33: demo.at:548 hardcoding library path
libtoolize automake autoconf
101: old-ltdl-iface.at:35 Makefile.inc
libtoolize automake autoconf
===
log here:
http://files.housegordon.org/libtool/openbsd56-libtool-2.4.4.8-9f52.tar.gz
-2-
When bootstrapping on the OpenBSD VM (i.e. "git clone && bootstrap etc") there are more errors, indicating perhaps some bad code generation?
===
$ make check
...
32: force non-PIC objects /tmp/libtool/tests/testsuite: /tmp/libtool/tests/testsuite.dir/at-groups/32/test-source[70]: syntax error: `}' unexpected
testsuite: WARNING: unable to parse test group: 32
testsuite: WARNING: A failure happened in a test group before any test could be
testsuite: WARNING: run. This means that test suite is improperly designed. Please
testsuite: WARNING: report this failure to <bug-libtool <at> gnu.org>.
...
====
Logs are here:
http://files.housegordon.org/libtool/openbsd56-libtool-git.make-check.log
http://files.housegordon.org/libtool/openbsd56-libtool-git-9f52.tar.gz
- Assaf
Information forwarded
to
bug-libtool <at> gnu.org:
bug#19222; Package
libtool.
(Fri, 05 Dec 2014 11:36:02 GMT)
Full text and
rfc822 format available.
Message #20 received at 19222 <at> debbugs.gnu.org (full text, mbox):
Hi Assaf,
Thanks for the feedback.
On Dec 4, 2014, at 9:33 PM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
> On 12/03/2014 05:05 PM, Gary V. Vaughan wrote:
>>> On Nov 30, 2014, at 1:56 AM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
>
> <...>
>
>> I can see immediately that most of the failures on your OpenBSD system are caused
>> by not having a GNU M4 installed on the path for libtoolize. I should have been a
>> lot more vocal about that change rather than hiding it in the 2.4.3 release notes,
>> as it has tripped a lot of people up. The latest master revision now checks for a
>> suitable M4 at configure time, and complains if there is nothing suitable on your
>> command search PATH, which I think will help.
>>
>> May I ask if you would kindly rerun then test suite on that machine and reply with a
>> link to the updated logs, after having installed GNU M4 first of course :)
>
> On OpenBSD 5.6, with GNU M4 installed as "/usr/local/bin/gm4",
>
> -1-
> Building the latest git ( 2.4.4.8-9f52 ) from a tarball (that is, "git clone && bootstrap && configure && make dist" on a GNU/Linux box, then "configure && make && make check" on the OpenBSD VM), two tests fail:
> ===
> 33: demo.at:548 hardcoding library path
> libtoolize automake autoconf
> 101: old-ltdl-iface.at:35 Makefile.inc
> libtoolize automake autoconf
> ===
>
> log here:
> http://files.housegordon.org/libtool/openbsd56-libtool-2.4.4.8-9f52.tar.gz
According to the comment at the top of old-ltdl-iface.at, I'm supposed to have
removed that test already, so you can safely ignore that failure :->
It's not at all obvious from the logs why the test cases are not working here :(
But...
> -2-
> When bootstrapping on the OpenBSD VM (i.e. "git clone && bootstrap etc") there are more errors, indicating perhaps some bad code generation?
>
> ===
> $ make check
> ...
> 32: force non-PIC objects /tmp/libtool/tests/testsuite: /tmp/libtool/tests/testsuite.dir/at-groups/32/test-source[70]: syntax error: `}' unexpected
> testsuite: WARNING: unable to parse test group: 32
> testsuite: WARNING: A failure happened in a test group before any test could be
> testsuite: WARNING: run. This means that test suite is improperly designed. Please
> testsuite: WARNING: report this failure to <bug-libtool <at> gnu.org>.
> ...
> ====
>
> Logs are here:
> http://files.housegordon.org/libtool/openbsd56-libtool-git.make-check.log
> http://files.housegordon.org/libtool/openbsd56-libtool-git-9f52.tar.gz
This looks like the bootstrap process and/or Autotest is making assumptions about
the shell utilities that are not true with the OpenBSD implementations.
You might be able to figure out where that is by diffing the generated testsuite
files between a Linux bootstrap and an OpenBSD bootstrapped version.
Cheers,
--
Gary V. Vaughan (gary AT gnu DOT org)
This bug report was last modified 10 years and 193 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.