[Expat-bugs] [ expat-Bugs-3175583 ] segfault with large CALS table

[SourceForge.net]

Bugs item #3175583, was opened at 2011-02-08 10:46
Message generated for change (Comment added) made by jcs32
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110127&aid=3175583&group_id=10127

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: jcs32 (jcs32)
Assigned to: Nobody/Anonymous (nobody)
Summary: segfault with large CALS table

Initial Comment:
I get a segfault without additional errors when trying to parse a large CALS table from a US patent document using XML::Twig which uses XML::Parser and expat underneath. I assume this may be the right buglist as a segfault should indicate the problem is (at least in effect) not on the perl side. libexpat is in version 1.5.2 (ubuntu 10.04). Attached is the offending table. There may be a problem in the xml itself, but usually that throws a meaningful parser error.

----------------------------------------------------------------------

>Comment By: jcs32 (jcs32)
Date: 2011-04-20 14:40

Message:
For the records, this was a problem of the the shell stack size and
occurred if the xml doc contained too many elements of one type (or
possibly of one type within the same parent tag). So ulimit -s solved this
issue.

----------------------------------------------------------------------

Comment By: jcs32 (jcs32)
Date: 2011-02-08 16:39

Message:
I just tested the entire file (containing the demo table) with xmlwf (a
compiled Expat application testing for well formedness), which seems to
work fine. This suggest that this is no Expat bug after all, although it
could be somewhere in the perl bindings. My apologies.
btw, could be that my library version is 2.0.1, that is what the package
manager claims. The link in /lib is libexpat.so.1.5.2 (has been updated),
however. Using XML::Twig I have had a second segfault, this time with more
than ~25000 <p> tags in one parent node (a similar condition as for the
table rows in the sample file). I cannot test more, had to build a
workaround. Sorry again for any inconvenience if this should be is a perl
only problem.

----------------------------------------------------------------------

Comment By: Fred L. Drake, Jr. (fdrake)
Date: 2011-02-08 15:01

Message:
That's a very old version of Expat; a number of bugs have been fixed since
then.

Please try to reproduce with a current version of libexpat.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110127&aid=3175583&group_id=10127