Jon Martin Solaas wrote:
Created SKER4766 in scarab:
http://sesat.no/scarab/issues/curmodule/10001/tqk/0/template/ViewIssue.vm/id/SKER4766/posinlist/2
This fix worked beautifully yesterday, and several hours today, but ...
Turns out that the Fast synonym syntax has changed, due to updates to
synonym lists, today. So the suggested query that appeared as "oslo"
yesterday now appears as "1###oslo" today. Which is a little closer to
the format AbstractSimpleFastSearchCommand.collectRelevantQueries would
accept initially (e.g. contains some #-chars ...), but the value is
interpreted as "1" and not "oslo".
If this method is not used by any code except catalogue-command
geography-expansion, maybe we could write a temporary version to just
handle the current format that we know of, and update it later when we
have a better overview? For instance collectRelevantQueries() looks for
'@'-characters to perform string splits during parsing, but it is
somewhat unclear what the @ is for and if/when/why it is present in our
synonym lists. Or does anyone have the full picture and time to fix it
once-and-for-all, or at least can provide me with enough info so that I
will know how to handle "1###oslo" without breaking anything?
If no, I may end up writing a monkey-version that looks for presence of
"!###" and absence of "@" and "#!#", and kick off some mumbojumbo code
to get hold of just "oslo" ... but it's not pretty ...
--
martin s.
begin:vcard
fn:Jon Martin Solaas
n:Solaas;Jon Martin
email;internet:[EMAIL PROTECTED]
tel;work:+47 92468511
tel;home:+47 92468511
tel;cell:+47 92468511
version:2.1
end:vcard
_______________________________________________
Kernel-development mailing list
[email protected]
http://sesat.no/mailman/listinfo/kernel-development
