Re: [wgs-assembler-users] Assertion failed in MANode2Array, August 5th SVN

[Walenz, B. <bw...@jc...>]

Hi, Ole-

Quite amusingly, I hit the same assert on a bacterial test assembly this afternoon.  No rhyme or reason here; we’ve run a fish or two, and I know of a few plants that have run too.

Looks like I slacked off and didn’t implement an array resize.  I’ll claim that I was thinking of replacing this array with an STL vector, but likely I was just lazy and then forgot to do it.

It’s fixed in r4380.

b




On 8/6/13 4:32 AM, "Ole Kristian Tørresen" <o.k...@ib...> wrote:

Hi,
I'm getting an assertion fail while running recent code from the SVN. I saw this with both July 31th and August 5th snapshots:
Working on unitig 258682 (0 unitigs and 2840 fragments)
utgcns: MultiAlignUnitig.C:99: int MANode2Array(MANode*, char***, int32***, int32): Assertion `0' failed.

Failed with 'Aborted'

Backtrace (mangled):

/projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns(_Z17AS_UTL_catchCrashiP7siginfoPv+0x2a)[0x40dcfa]
/lib64/libpthread.so.0[0x3fefc0f500]
/lib64/libc.so.6(gsignal+0x35)[0x3fef0328a5]
/lib64/libc.so.6(abort+0x175)[0x3fef034085]
/lib64/libc.so.6[0x3fef02ba1e]
/lib64/libc.so.6(__assert_perror_fail+0x0)[0x3fef02bae0]
/projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns(_ZN15unitigConsensus17generateConsensusEv+0x4fc)[0x42ddbc]
/projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns(_Z16MultiAlignUnitigP11MultiAlignTP7gkStoreP11CNS_OptionsPi+0xa89)[0x42fa19]
/projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns(main+0xa11)[0x409e51]
/lib64/libc.so.6(__libc_start_main+0xfd)[0x3fef01ecdd]
/projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns[0x409089]

Backtrace (demangled):

[0] /projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns::AS_UTL_catchCrash(int, siginfo*, void*) + 0x2a  [0x40dcfa]
[1] /lib64/libpthread.so.0() [0x3fefc0f500]
[2] /lib64/libc.so.6::(null) + 0x35  [0x3fef0328a5]
[3] /lib64/libc.so.6::(null) + 0x175  [0x3fef034085]
[4] /lib64/libc.so.6() [0x3fef02ba1e]
[5] /lib64/libc.so.6::(null) + 0  [0x3fef02bae0]
[6] /projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns::unitigConsensus::generateConsensus() + 0x4fc  [0x42ddbc]
[7] /projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns::MultiAlignUnitig(MultiAlignT*, gkStore*, CNS_Options*, int*) + 0xa89  [0x42fa19]
[8] /projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns::(null) + 0xa11  [0x409e51]
[9] /lib64/libc.so.6::(null) + 0xfd  [0x3fef01ecdd]
[10] /projects/cees/bin/celera/wgs-August5_2013_pacbio/Linux-amd64/bin/utgcns() [0x409089]

GDB:


If I understand this code correctly (http://sourceforge.net/p/wgs-assembler/svn/HEAD/tree/trunk/src/AS_CNS/MultiAlignUnitig.C#l98), if ir + 1 == depth you'll always hit the assertion.

I guess I can run this with older code, from before June 17th...

I've attached the layout of the unitig, if that's helpful. The assembly is a combination of Illumina, 454 and error corrected PacBio reads.

Thank you.

Ole