Some cleaning :
- removing dirty hack getMARCurl in SearchMarc.pm & replacing it by some Perl in Search.pm : if repeated, urls are stored in biblioitems.url, separated by a |. Now, the urls are splitted in an array. WARNING : the template has been modified & NPL MUST modify their templates as well, should be really trivial (open http://bureau.paulpoulain.com:9010/cgi-bin/koha/opac-detail.pl?bib=1 to see the result)
- using getMARCsubjcts and getMARCnotes in css templates as well. WARNING (specially for NPL) : the MARC notes & subjects were outside from BIBLIO_RESULTS array. As notes & subjects were in this <!-- TMPL_LOOP --> i had to add them IN the loop. I think it's better, so I commented (line 73 and 74 of opac-detail) 2 lines that are useless for me but useful for you. My opinion is that it would be better to modify your template to put notes & subjects in the <TMPL_LOOP> (should be trivial) than uncommenting those lines (= create duplicate code)
However, my real opinion is that we have no good reason to use <!-- TMPL_LOOP name="BIBLIO_RESULTS" --> as we always have only 1 biblio here. But that's something we should fix in HEAD, not in 2.2