Bug#575002: marked as done (std::bad_alloc thrown in regression tests [armel powerpc s390])
Your message dated Mon, 12 Apr 2010 19:42:35 -0400 with message-id and subject line Version: 2.2-13 has caused the Debian Bug report #575002, regarding std::bad_alloc thrown in regression tests [armel powerpc s390] to be marked as done. This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact ow...@bugs.debian.org immediately.) -- 575002: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=575002 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems --- Begin Message --- Source: cvc3 Version: 2.2-11 Severity: serious Justification: FTBFS Hi, your package FTBFS in various ways, on various archs: armel: | Running API test, regress level = 0 | | test(): { | Lambda application: (LAMBDA (x: INT): x)(1) | Simplified: 1 | | } | test1(): { | Query: TRUE | Valid | | Query: FALSE | Invalid | | Query: (p OR NOT p) | Valid | | Query: ((x = y) => (f(x) = f(y))) | make[2]: *** [regress] Error 1 | make[1]: *** [regress0] Error 2 kfreebsd-amd64: | Java API test | * | /usr/bin/make -C java test 2>&1 \ | | tee -a regressions.log; [ ${PIPESTATUS[0]} -eq 0 ] | make[3]: Entering directory `/build/buildd-cvc3_2.2-11-kfreebsd-amd64-5_aMGU/cvc3-2.2/java' | /usr/lib/jvm/default-java/bin/java -d64 -Djava.library.path=../java/lib/x86_64-kfreebsd-gnu -ea -jar ../java/lib/cvc3test.jar | make[3]: *** [test] Bus error (core dumped) kfreebsd-i386: | Query: ((x = y) => (f(x) = f(y))) | terminate called after throwing an instance of 'std::bad_alloc' | what(): std::bad_alloc | make[2]: *** [regress] Error 1 s390: | Running API test, regress level = 0 | | test(): { | Lambda application: (LAMBDA (x: INT): x)(1) | Simplified: 1 | | } | test1(): { | Query: TRUE | Valid | | Query: FALSE | Invalid | | Query: (p OR NOT p) | Valid | | Query: ((x = y) => (f(x) = f(y))) | make[2]: *** [regress] Error 1 | make[1]: *** [regress0] Error 2 | make: *** [debian/stamp-makefile-check] Error 2 Full build logs: https://buildd.debian.org/status/package.php?p=cvc3 Mraw, KiBi. --- End Message --- --- Begin Message --- Version: 2.2-13 --- End Message ---
Bug#575002:
This appears to be caused by the following assignment in src/sat/minisat_solver.cpp, line 1582: d_assigns[var] = toInt(lbool(p.sign())); d_assigns has type vector. On the affected architectures, char is unsigned. When p.sign() is -1, the result is that d_assigns[var] is assigned 255, an unsigned char, which causes the following assertion to fail, because the value of l_False is -1, a signed int (getValue(c[z]) just coerces the d_assigns value for c[z].var() to int): DebugAssert(getValue(c[z]) == l_False, "MiniSat::Solver:::propagate: Unit Propagation"); The offending assignment exists in the upstream MiniSat code, but it's not clear whether it leads to a bug in the absence of the changes made for CVC3 (the assertion is one such change). A simple fix is to change the type of d_assigns to vector. A new version of the package with this fix is pending. -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#575002: cvc3: FTBFS on various archs/in various ways
Hi again, Nice bug descriptions. :) Christopher L Conway wrote: > What should be the disposition of this bug? I guess it should be merged with one of the two offshoots. Happy debugging, Jonathan -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#575002: cvc3: FTBFS on various archs/in various ways
I've requested removal of the failing binaries here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=576333 The regression failures appear to fall into two classes. I've cloned a bug for the std::bad_alloc error here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=576334 The JVM crash on kfreebsd-amd6 is here: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=576335 I will take a look at armel using qemu. What should be the disposition of this bug? Thanks, Chris -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#575002: cvc3: FTBFS on various archs/in various ways
Hi Chris, Christopher L Conway wrote: > the problem architectures are not directly supported by > upstream and I don't have access to testing machines for these > architectures, I'm not sure what the correct resolution is. Should I > remove these architectures in debian/control? I’d suggest something like this: 1. First, request the removal of the broken packages currently in testing. You can do this with a bug against the release.debian.org package [1]. Once this finishes, the bug is no longer release critical (though it is still important). 2. Clone this bug into separate bugs tracking information about each collection of regressions that you think will have the same fix. 3. Add information about how to track them down (a good command to run, a patch to try, whatever) to the bug log. 4. Help anyone who is interested in working on the bugs. If you have time, you can investigate further without access to a porterbox: * armel can be emulated with qemu [2]. * kfreebsd-amd64 and kfreebsd-i386 can be run natively on common hardware. Or you can run them in a virtual machine (e.g., KVM), still fairly quickly. * s390 can be emulated with hercules [3], though rather slowly. On a slow machine, this can try one’s patience, though. No need to get carried away. DDs have access to some machines [4] for porting. This is a good reason to become a DD. :) If you have any questions, please don’t hesitate to ask. HTH, Jonathan [1] http://wiki.debian.org/ftpmaster_Removals [2] http://www.aurel32.net/info/debian_arm_qemu.php [3] http://www.josefsipek.net/docs/s390-linux/hercules-s390.html [4] http://db.debian.org/machines.cgi -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#575002: cvc3: FTBFS on various archs/in various ways
Kibi, Thanks for reporting this. I've already file a bug upstream for these errors. As the problem architectures are not directly supported by upstream and I don't have access to testing machines for these architectures, I'm not sure what the correct resolution is. Should I remove these architectures in debian/control? Note also that bug also likely affects package version 2.2-7, which didn't run the regressions in debian/rules. Is there a mechanism to revoke the affected binary packages from testing? -Chris On Mon, Mar 22, 2010 at 3:26 PM, Cyril Brulebois wrote: > Source: cvc3 > Version: 2.2-11 > Severity: serious > Justification: FTBFS > > Hi, > > your package FTBFS in various ways, on various archs: > > armel: > | Running API test, regress level = 0 > | > | test(): { > | Lambda application: (LAMBDA (x: INT): x)(1) > | Simplified: 1 > | > | } > | test1(): { > | Query: TRUE > | Valid > | > | Query: FALSE > | Invalid > | > | Query: (p OR NOT p) > | Valid > | > | Query: ((x = y) => (f(x) = f(y))) > | make[2]: *** [regress] Error 1 > | make[1]: *** [regress0] Error 2 > > kfreebsd-amd64: > | Java API test > | * > | /usr/bin/make -C java test 2>&1 \ > | | tee -a regressions.log; [ ${PIPESTATUS[0]} -eq 0 ] > | make[3]: Entering directory > `/build/buildd-cvc3_2.2-11-kfreebsd-amd64-5_aMGU/cvc3-2.2/java' > | /usr/lib/jvm/default-java/bin/java -d64 > -Djava.library.path=../java/lib/x86_64-kfreebsd-gnu -ea -jar > ../java/lib/cvc3test.jar > | make[3]: *** [test] Bus error (core dumped) > > kfreebsd-i386: > | Query: ((x = y) => (f(x) = f(y))) > | terminate called after throwing an instance of 'std::bad_alloc' > | what(): std::bad_alloc > | make[2]: *** [regress] Error 1 > > s390: > | Running API test, regress level = 0 > | > | test(): { > | Lambda application: (LAMBDA (x: INT): x)(1) > | Simplified: 1 > | > | } > | test1(): { > | Query: TRUE > | Valid > | > | Query: FALSE > | Invalid > | > | Query: (p OR NOT p) > | Valid > | > | Query: ((x = y) => (f(x) = f(y))) > | make[2]: *** [regress] Error 1 > | make[1]: *** [regress0] Error 2 > | make: *** [debian/stamp-makefile-check] Error 2 > > Full build logs: > https://buildd.debian.org/status/package.php?p=cvc3 > > Mraw, > KiBi. > > > -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#575002: cvc3: FTBFS on various archs/in various ways
Source: cvc3 Version: 2.2-11 Severity: serious Justification: FTBFS Hi, your package FTBFS in various ways, on various archs: armel: | Running API test, regress level = 0 | | test(): { | Lambda application: (LAMBDA (x: INT): x)(1) | Simplified: 1 | | } | test1(): { | Query: TRUE | Valid | | Query: FALSE | Invalid | | Query: (p OR NOT p) | Valid | | Query: ((x = y) => (f(x) = f(y))) | make[2]: *** [regress] Error 1 | make[1]: *** [regress0] Error 2 kfreebsd-amd64: | Java API test | * | /usr/bin/make -C java test 2>&1 \ | | tee -a regressions.log; [ ${PIPESTATUS[0]} -eq 0 ] | make[3]: Entering directory `/build/buildd-cvc3_2.2-11-kfreebsd-amd64-5_aMGU/cvc3-2.2/java' | /usr/lib/jvm/default-java/bin/java -d64 -Djava.library.path=../java/lib/x86_64-kfreebsd-gnu -ea -jar ../java/lib/cvc3test.jar | make[3]: *** [test] Bus error (core dumped) kfreebsd-i386: | Query: ((x = y) => (f(x) = f(y))) | terminate called after throwing an instance of 'std::bad_alloc' | what(): std::bad_alloc | make[2]: *** [regress] Error 1 s390: | Running API test, regress level = 0 | | test(): { | Lambda application: (LAMBDA (x: INT): x)(1) | Simplified: 1 | | } | test1(): { | Query: TRUE | Valid | | Query: FALSE | Invalid | | Query: (p OR NOT p) | Valid | | Query: ((x = y) => (f(x) = f(y))) | make[2]: *** [regress] Error 1 | make[1]: *** [regress0] Error 2 | make: *** [debian/stamp-makefile-check] Error 2 Full build logs: https://buildd.debian.org/status/package.php?p=cvc3 Mraw, KiBi. -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org