Bugs‎ > ‎

Nasty, easy-to-fix and easily-forgotten details

In the old days, when most people had no way of entering accented characters into search forms, we at the ARTFL Project developed a method that we called post-fix. Basically, a user could (and still can in fact) enter "c,", "e/" or "a^" and those strings would be converted on the way in, to "ç", "é" and "à" respectively. This causes a problem when one does a search for say "Balzac, Honoré de" in the author bibliographic search field. This search becomes "Balzaç Honoré de". The fix is relatively simple. The postfix2UTF8 subroutine in philosubs.pl would need the following change:

$l =~ s/c\,/\xc3\xa7/g;

should be changed to

$l =~ s/c\,([a-zA-Z\177-\377])/\xc3\xa7$1/g;

This simple fix has managed to slip under the radar (and out of the CVS) in our previous distributions and although we hope to get it in there next time around, if it eludes us yet again, feel free to make the fix yourself!

Comments