[Bioperl-l] perl 5.10 released

Chris Fields cjfields at uiuc.edu
Wed Dec 19 17:08:31 UTC 2007


On Dec 19, 2007, at 10:44 AM, Dave Messina wrote:

> Hi everyone,
>
>
> Perl 5.10 builds fine and passes all tests on my PB G4 running OS X  
> 10.5.1. Piece o' cake.
>
> Here are results of testing BioPerl on this virgin install:
>
> I downloaded the latest CVS tarball. I did 'perl Build.PL', which  
> used CPAN to install a bunch of dependencies. I then did 'Build  
> test'. For the most part everything was fine.
>
> - Bio::Biblio::IO::medlinexml throws an exception because  
> XML::Parser isn't installed.

XML::Parser used to be shipped with a number of perl distros even  
though it isn't core.  We should add a require to these.

> - RNA_SearchIO fails a few tests.

These are very likely from recent commits I made re:GenericHSP and use  
of bits(), raw_score(), etc. (the fails look like missing/switched  
vals with these method tests).  I'll fix these post-svn migration, but  
I don't think these are related to 5.10.

> - Bio::Ontology::SimpleGOEngine::GraphAdaptor throws an exception  
> because Graph::Directed isn't installed.

Odd, that should be caught out before tests are run.  Needs to be  
fixed, but one would think this would fail as well under 5.8.

> - Spidey fails one test.

Passes for me.  Is it dependency-related?

> And of course without the optional dependencies installed, many  
> tests were skipped.
>
> I'll now go back and install the optional dependencies and do the  
> network tests, but it looks like for the most part we play nice with  
> the new Perl.
>
> Dave

Not sure, but it seems a bit faster.  Maybe it's just me but it would  
be nice to see some benchmarks comparing perl 5.8 vs 5.10.  I agree,  
it was a very fast and easy install.

I'll start a page on the wiki for test fails using perl 5.10.  I'm  
seeing a few fails;  I'm getting the following with everything  
installed (including DBD::mysql, DBI, etc) using perl 5.10, Mac OS X  
10.5.1 (note Test::Harness now gives TODO's, so some of these are  
actually passing).  Note the entrezgene.t and DB.t fails; I looked  
into these and I think they are related to the odd 'pseudohashes are  
deprecated' warnings we were getting in perl 5.8 tests, so there may  
be something legitimately buggy.

Test Summary Report
-------------------
t/Annotation.t                (Wstat: 0 Tests: 112 Failed: 0)
   TODO passed:   96
t/BioGraphics.t               (Wstat: 256 Tests: 35 Failed: 1)
   Failed test number(s):  4
   Non-zero exit status: 1
t/DB.t                        (Wstat: 65280 Tests: 106 Failed: 0)
   Non-zero exit status: 255
   Parse errors: Bad plan.  You planned 116 tests but ran 106.
t/DBCUTG.t                    (Wstat: 1024 Tests: 33 Failed: 4)
   Failed test number(s):  29-31, 33
   Non-zero exit status: 4
t/RNA_SearchIO.t              (Wstat: 2048 Tests: 496 Failed: 8)
   Failed test number(s):  291, 338, 372-374, 395, 455, 486
   Non-zero exit status: 8
t/entrezgene.t                (Wstat: 65280 Tests: 648 Failed: 0)
   Non-zero exit status: 255
   Parse errors: Bad plan.  You planned 1422 tests but ran 648.
Files=255, Tests=15066, 435 wallclock secs ( 3.15 usr  1.72 sys +  
124.87 cusr 13.29 csys = 143.03 CPU)
Result: FAIL
Failed 5/255 test programs. 13/15066 subtests failed.


chris



More information about the Bioperl-l mailing list