[OPEN-ILS-DEV] z39.50 again

[Dan Scott]

On Tue, 2010-09-07 at 16:55 -0300, Grant Johnson wrote:
> Thanks Dan,
> 
> Same versions as indicated.
> 
>  yaz-config --version
> 3.0.47
> 
> perl -MNet::Z3950::Simple2ZOOM -e 'print
> > $Net::Z3950::Simple2ZOOM::VERSION . "\n"'
> 1.04
> 
>  
> I added  inputcharset="utf-8"  like so... to no avail... should this be
> marc-8?

Nope, everything coming out of Evergreen is UTF-8.

If I've counted correctly, the record in question is
http://islandpines.roblib.upei.ca/opac/en-CA/skin/roblib/xml/rdetail.xml?r=580632&ol=4&t=medicine&tp=keyword&l=4&d=2&o=60&hc=4019&rt=keyword

To look at that a little closer:
http://islandpines.roblib.upei.ca/opac/extras/supercat/format/marcxml/record/580632

And even closer, let's look at the leader:

<leader>01607cam  2200373 a 4500</leader>

Is LDR/09 an 'a', indicating UTF8 encoding? No sir! Does that confuse
current versions of MARC::Record / MARC::File::XML no end, in that it
sees that the record claims to be MARC-8 encoded (due to the space
instead of the 'a' in LDR/09) and dutifully tries to convert the record
from MARC8 to UTF8 (even though it's already in UTF8)?

Down in the 505 $a, we have some \x2019 Unicode characters.

Could that result in immense pain? Indeed! You can see where we ran into
some of that pain in the past in the diffs at
http://svn.open-ils.org/trac/ILS/changeset?new=16250%40branches%
2Frel_1_6_0%2FOpen-ILS%2Fsrc%2Fperlmods%2FOpenILS%2FWWW%
2FSuperCat.pm&old=13582%40branches%2Frel_1_6_0%2FOpen-ILS%2Fsrc%
2Fperlmods%2FOpenILS%2FWWW%2FSuperCat.pm - specifically in this instance
forcing LDR/09 to 'a' before passing it through the potentially
character-corrupting as_xml_record().

Make sure you have an up-to-date copy of this
(http://svn.open-ils.org/trac/ILS/browser/branches/rel_1_6_0/Open-ILS/src/perlmods/OpenILS/WWW/SuperCat.pm) in your /openils/lib/perl5/OpenILS/WWW/SuperCat.pm directory, restart apache, and give it a go again. I've got my fingers crossed that you'll have a better result.