IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

purified seg-fault data

Dave Matthews matthews at GREENGENES.CIT.CORNELL.EDU
Sat Mar 9 19:08:03 EST 1996

Hi Richard, Jean,

I'm having trouble with segmentation faults while loading ace4_1 from
rawdata/.ace files.  

This has been going on for a few weeks now, ever since I set up a copy of
the database on a Solaris machine that I reinitialize and reload every
night via tace.  Since it always fails on only one of my 115 rawdata files
I thought it might be some error in the file, so I've been trying to isolate
the problem.  Now that I've purified it I think my file is innocent, but
maybe some of the observations I've made would be helpful to you.

The problem file ("p.ace") is about 90 kbytes.  The crash always occurs at
the same point in p.ace, unless I've made changes in other files that are
loaded prior to it.  Here is the record that crashes it under one set of

Locus : "alpha-Amy-D2"
Chromosome 7D
Chromosome_arm 7DL

I.e. if I delete everything in the file after this point (about 60% of it),
it still crashes.  If I remove this whole record there is no crash, and in
addition the entire p.ace can then be loaded in the next tace session with
no crash.  Removing just one or both of the data values,

Chromosome 7D
Chromosome_arm 7DL

gives an intermediate result:  there is no immediate crash but loading the
whole p.ace in the next session crashes.

The Locus object "alpha-Amy-D2" does not already exist in the database at
the time this is loaded.

Both tags, Chromosome and Chromosome_arm are ?Text.  I notice that if I
query "find text 7d" I get an error ">?Text should be followed by at least
3 characters".  Could this be part of the problem?  In the completely loaded
database there are 233 "7D"s and 88 "7DL"s.

No error is reported by tace in log.wrm.  tace's exit $status is always

The problem seems to be the same on two different Solaris machines.  I just
tested under SunOS and got a seg-fault in a different file instead. 

A weirdness:  After I've loaded all 115 files (with p.ace duly crashing),
I can then load it with no problem.  But if I try to load it immediately 
after it has crashed it only crashes again.

... Hmm, maybe my file wasn't so innocent after all.  I just changed those
two tags from ?Text to Text and reinitialized.  No crash!! 

So should I change all my ?Texts to Texts?

- Dave

More information about the Acedb mailing list

Send comments to us at biosci-help [At] net.bio.net