bug#19222: Some test failures with LibTool-2.4.4 on non GNU/Linux systems

[Gary V. Vaughan]

Hi Assaf,

Thanks for the reports and results.

> On Nov 30, 2014, at 1:56 AM, Assaf Gordon <address@hidden> 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)