[OPEN-ILS-DEV] direct_ingest.pl, biblio_fingerprint.js and Unicode chars

[Dan Scott]

2009/11/30 Dan Wells <dbw2 at calvin.edu>:
> Hi Dan,
>
> I am still listening, and would be quite happy to see this resolved somehow.  I also want to note that the marc2bre.pl changed back to using composed characters here:
>
> http://svn.open-ils.org/trac/ILS/changeset/12985/trunk/Open-ILS/src/extras/import/marc2bre.pl
>

Heh, right, that was the first thing I pointed out back in my original
novel-length reply :)

> When doing my lastest loads, I have some records which fail using composed characters while others fail using decomposed.  Other than noting these failed records, I haven't been able to do any troubleshooting to determine a precise cause.

It would be great to have one of the records that fail for you in
decomposed mode to add to our test datasets, so we can avoid
flip-flopping the decomposed/composed mode in the future and nail the
real cause of the problem.

I suppose the next logical step would be to rewrite the
OpenILS::Application::Ingest fingerprint methods to avoid the
ScriptRunner server-side JavaScript biblio_fingerprint.js fun and see
if that resolves the problem.