[armel] gpgme1.0 1.2.0-1.2ubuntu1 fails to build in maverick & natty

Bug #607804 reported by Ulrich Weigand
10
This bug affects 1 person
Affects Status Importance Assigned to Milestone
gpgme1.0 (Ubuntu)
Fix Released
High
Jani Monoses
Natty
Fix Released
High
Jani Monoses
pth (Ubuntu)
Fix Released
High
Unassigned
Natty
Fix Released
High
Unassigned

Bug Description

checking for GNU Pth... *** longjmp causes uninitialized stack frame ***: ./conftest terminated
*FAILED*
 | Found GNU Pth 2.0.7 under /usr, but
 | was unable to perform a sanity execution check. This usually
 | means that the GNU Pth shared library libpth.so is present
 | but $LD_LIBRARY_PATH is incomplete to execute a Pth test.
 | In this case either disable this test via --without-pth-test,
 | or extend $LD_LIBRARY_PATH, or build GNU Pth as a static
 | library only via its --disable-shared Autoconf option.
 | We used the following build environment:
 | CC="gcc"
 | CFLAGS="-Wall -g -O2 -I/usr/include"
 | LDFLAGS="-Wl,-Bsymbolic-functions -L/usr/lib"
 | LIBS=" -lpth -ldl -lnsl"
 | See config.log for possibly more details.
make: *** [configure-stamp] Error 1
dpkg-buildpackage: error: debian/rules build gave error exit status 2

Tags: armel
Revision history for this message
Scott Kitterman (kitterman) wrote :

It's now FTBFS in the archive, so suggestions appreciated on how to fix this.

Changed in gpgme1.0 (Ubuntu):
importance: Undecided → High
milestone: none → natty-alpha-2
status: New → Confirmed
Revision history for this message
Scott Kitterman (kitterman) wrote :

I tried a no-change rebuild of pth on armel and now the failure is:

checking for GNU Pth... *** longjmp causes uninitialized stack frame ***: ./conftest terminated
*FAILED*
 | Found GNU Pth 2.0.7 under /usr, but
 | was unable to perform a sanity execution check. This usually
 | means that the GNU Pth shared library libpth.so is present
 | but $LD_LIBRARY_PATH is incomplete to execute a Pth test.
 | In this case either disable this test via --without-pth-test,
 | or extend $LD_LIBRARY_PATH, or build GNU Pth as a static
 | library only via its --disable-shared Autoconf option.
 | We used the following build environment:
 | CC="gcc"
 | CFLAGS="-Wall -g -O2 -I/usr/include"
 | LDFLAGS="-Wl,-Bsymbolic-functions -lpth -L/usr/lib"
 | LIBS=" -lpth -ldl -lnsl"
 | See config.log for possibly more details.
make: *** [configure-stamp] Error 1

re-assigning to pth.

affects: gpgme1.0 (Ubuntu Natty) → pth (Ubuntu Natty)
Revision history for this message
Scott Kitterman (kitterman) wrote :

config.log from the second failed build of gpgme1.0 (after the pth rebuild).

Revision history for this message
Matthias Klose (doko) wrote :

the reassignment is at least incomplete. m4/pth.m4 is found in gpgme1.0

summary: - [armel] gpgme1.0 1.2.0-1.2ubuntu1 fails to build in maverick
+ [armel] gpgme1.0 1.2.0-1.2ubuntu1 fails to build in maverick & natty
Changed in gpgme1.0 (Ubuntu Natty):
importance: Undecided → High
milestone: none → natty-alpha-2
status: New → Confirmed
Changed in pth (Ubuntu Natty):
status: Confirmed → New
Revision history for this message
Scott Kitterman (kitterman) wrote :

As an experiment, I updated the gpgme1.0 m4/pth.m4 with the one from libpth-dev and it still fails.

Revision history for this message
Scott Kitterman (kitterman) wrote :

<doko> ScottK: we know that we don't want pth, so maybe just disable this autoconf check? I didn't analyze the build failure
<ScottK> doko: If you think that's best, I can do that.
<ScottK> Gotta go. Airplane is about to land.
<doko> it's not a fix, but a workaround which should work for all archs. subscribe Janimo

Jani Monoses (jani)
Changed in gpgme1.0 (Ubuntu Natty):
assignee: nobody → Jani Monoses (jani)
Revision history for this message
Jani Monoses (jani) wrote :

@doko: we don't want pth ? I see libgpgme provides both pth and pthread based shared libraries.

/usr/lib/libgpgme-pth.so.11

Or should the pth variant be dropped on armel only?

pth has issues on ARM, possibly due to the fact that user contexts are not available on that architecture, and causes the vwdial related crashes and FTBFS But this longjump bug may be unrelated to that.
 https://bugs.launchpad.net/ubuntu/+source/glibc/+bug/383975

Revision history for this message
Jani Monoses (jani) wrote :

builds with the latest pth

Changed in gpgme1.0 (Ubuntu Natty):
status: Confirmed → Fix Released
Changed in pth (Ubuntu Natty):
status: New → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.