SuperCat SRU double-encodes UTF8 characters, does not set character encoding

[Dan Scott]

* Evergreen master
* Ubuntu 12.04

Tested with a 2.7 production system, the HTTP header was setting the charset to ISO-8859-1 (per Apache defaults I believe) because the SuperCat SRU methods were not themselves setting an explicit header. For an example, try:

curl -I 'https://laurentian.concat.ca/opac/extras/sru/LUSYS/holdings?version=1.1&operation=searchRetrieve&query=societes&maximumRecords=11'

(replacing hostname / library shortname / search query as necessary).

Also, perhaps due to changes in Encode or MARC::XML behaviour on Ubuntu 12.04, it seems that the encode_utf8() call for $marc->as_xml_record() is no longer necessary; in fact, it corrupts any non-ASCII characters. Which isn't good.

[Dan Scott (denials)]

See http://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/dbs/lp1431541_supercat_sru_encoding for a very simple fix, tested and in production on an Ubuntu 12.04 system that has a whole ton of non-ASCII characters.

[Mike Rylander (mrylander)]

Dan, that looks entirely sane. Seems this is relevant for backport to 2.6+ ... no?

[Dan Scott (denials)]

Mike, yes, I believe this should be backported to 2.6. I'll set the milestones accordingly!

[Ben Shum (bshum)]

Fix pushed to master and backported to rel_2_8 and rel_2_7.