[Bioperl-l] Bioperl 1.5.2 RC5

Chris Fields cjfields at uiuc.edu
Mon Nov 27 11:41:53 EST 2006

On Nov 27, 2006, at 10:18 AM, Sendu Bala wrote:

> Chris Fields wrote:
>> On Nov 26, 2006, at 3:00 PM, Conrad Halling wrote:
> [snip]
>>> t/SearchIO...................ok
>>> 1/1340
>>> -------------------- WARNING ---------------------
>>> MSG: XML::SAX::Expat not currently supported; must have local   
>>> copies of
>>> NCBI DTD docs!
>>> ---------------------------------------------------
>>> -------------------- WARNING ---------------------
>>> MSG: error in parsing a report:
>>> 404 File
>>> `/Users/challing/.cpan/build/bioperl-1.5.2_005-RC/  
>>> NCBI_BlastOutput.dtd'
>>> does not exist
>>> file:///Users/challing/.cpan/build/bioperl-1.5.2_005-RC/  
>>> NCBI_BlastOutput.dtd
>>> Handler couldn't resolve external entity at line 2, column 82,  
>>> byte  104
>>> error in processing external entity reference at line 2, column  
>>> 82,  byte
>>> 104 at /test/perl/lib/site_perl/5.8.8/darwin-2level/XML/ 
>>> Parser.pm  line 187
> [snip]
>> There is an eval{} block set up to catch that error but apparently  
>> it  doesn't work.  I'll have a look at it.
> The test script has nothing to catch in the eval because  
> blastxml.pm catches the error and emits a warning. You'll either  
> have to make blastxml.pm throw (instead of returning undef + warn),  
> or figure something else out.
> It's a behaviour change, but would anyone care if it threw thanks  
> to this problem?

Ah, that's my fault.  I had changed that to a throw but didn't commit  
it.  In a nutshell, we can't/won't support XML::SAX::Expat for a  
number of reasons.  One is the problem with the BLAST DTDs (no  
explicit web address, so XML validation fails).  There may be a way  
to turn validation off.  However, the other, more serious issue is  
that it is no longer maintained (it hasn't been updated in three  
years).  There were plans a year ago for XML::SAX::ExpatXS to take  
over the XML::SAX::Expat namespace but no one has bothered to make  
the changeover, apparently.

I have committed the change from warn to throw and made the error  
message more explicit.  You can look it over and merge to 1.5.2 at  
your discretion.


Christopher Fields
Postdoctoral Researcher
Lab of Dr. Robert Switzer
Dept of Biochemistry
University of Illinois Urbana-Champaign

More information about the Bioperl-l mailing list