Bug#793845: nsf: FTBFS on Linux due to test suite errors
[Dropping pkg-tcltk, which bounces without -devel.] Stefan Sobernig stefan.sober...@wu.ac.at writes: Thx for this help! Did you test on the very same buildd/porterbox which reports the issue: hasse.debian.org; or any other armhf/mipsel box? I don't have access to the autobuilders, on which logins are restricted to their actual maintainers. Rather, my test builds were on harris.debian.org (armhf) and eder.debian.org (mipsel). -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Thx for this help! Did you test on the very same buildd/porterbox which reports the issue: hasse.debian.org; or any other armhf/mipsel box? I don't have access to the autobuilders, on which logins are restricted to their actual maintainers. Rather, my test builds were on harris.debian.org (armhf) and eder.debian.org (mipsel). On debian-admin I was told that abel.debian.org has the same hardware as hasse.debian.org, and is a porterbox. It will take some time till I can sign up and get sponsoring for a guest account there, any chance that you can give it a try on abel.debian.org and report back (maybe including a core dump)? Thank you for considering it. Stefan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Stefan Sobernig stefan.sober...@wu.ac.at writes: I would still appreciate it if you could test on abel. I did just now, and had a successful build there too. I wonder what went wrong originally. Anyway, thanks again! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Thx for this help! Did you test on the very same buildd/porterbox which reports the issue: hasse.debian.org; or any other armhf/mipsel box? I don't have access to the autobuilders, on which logins are restricted to their actual maintainers. Rather, my test builds were on harris.debian.org (armhf) and eder.debian.org (mipsel). On debian-admin I was told that abel.debian.org has the same hardware as hasse.debian.org, and is a porterbox. It will take some time till I can sign up and get sponsoring for a guest account there, any chance that you can give it a try on abel.debian.org and report back (maybe including a core dump)? Thank you for considering it. 2.0.0-2 builds successfully for armhf/mipsel on boxes other than hasse?! https://buildd.debian.org/status/package.php?p=nsfsuite=unstable And there is no change to 2.0.0-2 other than disabling the HTTP tests. I would still appreciate it if you could test on abel. Thx. //stefan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Hi Aaron! How can I best simulate the environment at the buildd box? Is there a chance to obtain a core dump from the buildd box, alternatively? I'm not aware of a way to get core dumps off of autobuilders, though perhaps if you asked their maintainers nicely, they might be able to help you out. I tried building the package on a physical porterbox of each architecture, but wasn't able to reproduce the error on either, so I'm not sure why the automatic builds failed. Thx for this help! Did you test on the very same buildd/porterbox which reports the issue: hasse.debian.org; or any other armhf/mipsel box? Stefan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Builds of nsf on Linux architectures failed with test suite errors, of two types. (There was also a kFreeBSD failure, which I'll report separately.) On armhf and mipsel, the linearization test segfaults when done: We will certainly look at these a.s.a.p. I tried to reproduce the issue, but sadly I failed in doing so. In my (emulated) armhf environment, building the package succeeds. Details on my build environment (QEMU-powered Debian sid ARMHF guest on Ubuntu host): # cat /etc/debian_version stretch/sid # uname -a Linux debian-armhf 3.2.0-4-vexpress #1 SMP Debian 3.2.51-1 armv7l GNU/Linux # apt-show-versions build-essential debhelper tcllib gcc tcl tcl-dev tcl8.6 build-essential:armhf/sid 11.7 uptodate debhelper:all/sid 9.20150628 uptodate gcc:armhf/sid 4:4.9.2-4 uptodate tcl:armhf/sid 8.6.0+8 uptodate tcl-dev:armhf/sid 8.6.0+8 uptodate tcl8.6:armhf/sid 8.6.4+dfsg-2 uptodate tcllib:all/sid 1.17-dfsg-1 uptodate What I did: apt-get source nsf apt-get build-dep nsf cd nsf-2.0.0 dpkg-buildpackage -us -uc gives me -rw-r--r-- 1 root root 303662 Jul 30 11:49 ../nsf_2.0.0-1_armhf.deb -rw-r--r-- 1 root root 25394 Jul 30 11:49 ../nsf-dev_2.0.0-1_armhf.deb -rw-r--r-- 1 root root 20870 Jul 30 11:49 ../nsf-shells_2.0.0-1_all.deb How can I best simulate the environment at the buildd box? Is there a chance to obtain a core dump from the buildd box, alternatively? Stefan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Stefan Sobernig stefan.sober...@wu.ac.at writes: How can I best simulate the environment at the buildd box? Is there a chance to obtain a core dump from the buildd box, alternatively? I'm not aware of a way to get core dumps off of autobuilders, though perhaps if you asked their maintainers nicely, they might be able to help you out. I tried building the package on a physical porterbox of each architecture, but wasn't able to reproduce the error on either, so I'm not sure why the automatic builds failed. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Hi Aaron! Thx for reporting, I had already noticed the failing builds. Builds of nsf on Linux architectures failed with test suite errors, of two types. (There was also a kFreeBSD failure, which I'll report separately.) On armhf and mipsel, the linearization test segfaults when done: We will certainly look at these a.s.a.p. On the remaining Linux architectures, the http test times out, perhaps due to idiosyncracies of buildd network configuration: !!!::req1 (2): Connection refused by host 'localhost' port '8086' error flushing sock949f8a8: socket is not connected!!! xocomm/test.001: incorrect result for 'Trying to load image logo-100.jpg ... ', expected: '1', got 0 Mmmh. I tested in my build environments in various configurations (amd64, i386). the connection tests suceed. How can I learn more about these buildd limitations you hinted at? Are certain ports locked down etc.? If I cannot ship around those limitations, what is the authoritative approach then: disable these network I/O tests? Cheers, Stefan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
On the remaining Linux architectures, the http test times out, perhaps due to idiosyncracies of buildd network configuration: !!!::req1 (2): Connection refused by host 'localhost' port '8086' error flushing sock949f8a8: socket is not connected!!! xocomm/test.001: incorrect result for 'Trying to load image logo-100.jpg ... ', expected: '1', got 0 Mmmh. I tested in my build environments in various configurations (amd64, i386). the connection tests suceed. How can I learn more about these buildd limitations you hinted at? Are certain ports locked down etc.? If I cannot ship around those limitations, what is the authoritative approach then: disable these network I/O tests? https://wiki.debian.org/buildd says one should not rely on any network (interface) being available, so (conditionally) disabling those tests is the way to go? Stefan -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Stefan Sobernig stefan.sober...@wu.ac.at writes: says one should not rely on any network (interface) being available, so (conditionally) disabling those tests is the way to go? Yes, I reckon so. Thanks for looking into it! Incidentally, I should clarify that I don't manage any autobuilders myself, just keep an eye out for new binary packages that appear on amd64 but not i386 or vice versa. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Source: nsf Version: 2.0.0-1 Severity: serious Justification: fails to build from source Builds of nsf on Linux architectures failed with test suite errors, of two types. (There was also a kFreeBSD failure, which I'll report separately.) On armhf and mipsel, the linearization test segfaults when done: linearization/boat.076: boat nx local-order violated ::wheelboat ::wheelboat ok make[1]: *** [test-core] Segmentation fault Makefile:533: recipe for target 'test-core' failed make[1]: Leaving directory '/«PKGBUILDDIR»' dh_auto_test: make -j1 test returned exit code 2 make: *** [build-arch] Error 2 debian/rules:11: recipe for target 'build-arch' failed On the remaining Linux architectures, the http test times out, perhaps due to idiosyncracies of buildd network configuration: TCL_LIBRARY=`echo /usr/include/tcl8.6/tcl-private/library` LD_LIBRARY_PATH=.:/usr/lib: PATH=.:/usr/lib:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games TCLLIBPATH=. . /usr/bin/tclsh8.6 ./library/xotcl/tests/xocomm.test -libdir `echo ./unix` -testlog ./__test.log /usr/bin/tclsh8.6 ./library/xotcl/tests/../apps/comm/webserver.xotcl -root ./library/xotcl/tests/../doc -pkgdir ./library/xotcl/tests/../library Warning: Non-positional parameter -verbose was passed more than once (::xocomm/test.001 method configure xocomm/test.001 -count 1 -verbose 0 -msg {Trying to load image logo-100.jpg ... } -verbose 1 -pre {puts starting...} -setResult {expr {[r0::sink set contentLength] == 1706}} -cmd {SimpleRequest r0 -url http://localhost:8086/logo-100.jpg}) pre: puts starting... run: SimpleRequest r0 -url http://localhost:8086/logo-100.jpg starting... !!!::req1 (2): Connection refused by host 'localhost' port '8086' error flushing sock949f8a8: socket is not connected!!! xocomm/test.001:incorrect result for 'Trying to load image logo-100.jpg ... ', expected: '1', got 0 in test file ./library/xotcl/tests/xocomm.test *** bgerror Test was exited with code -1 Test was exited with code -1 while executing $example run (procedure run line 4) invoked from within nx::test run (after script)* make: *** wait: No child processes. Stop. make: *** Waiting for unfinished jobs make: *** wait: No child processes. Stop. E: Caught signal ‘Terminated’: terminating immediately make[1]: *** [test-http] Terminated Makefile:589: recipe for target 'test-http' failed Build killed with signal TERM after 150 minutes of inactivity Could you please take a look? Thanks! -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#793845: nsf: FTBFS on Linux due to test suite errors
Aaron M. Ucko u...@debian.org writes: (There was also a kFreeBSD failure, which I'll report separately.) Never mind; a closer look at the build log reveals that configuration on kfreebsd-amd64 failed for the simple reason that the autobuilder simply didn't have enough disk space, which indicates nothing one way or the other about nsf's portability beyond Linux. -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org