Re: [libreoffice-l10n] How does one find specific strings in the po files?

On Sat, Dec 17, 2011 at 12:55 PM, Sérgio Marques <smarquespt@gmail.com> wrote:
> One thing that would also be useful is the ability of ignore accelerators.
>  If one wants to find Open and the word is O~pen, the second won´t appear
> on the search.
>
> This is definitely annoying and contra producing. This feature is a must
> have for search function. What do You guys think about it?
>
>
> > I think this is already filed as
> > http://bugs.locamotion.org/show_bug.cgi?id=327 ?
>
> Yes, I´ve already saw this entry but I think that it´s the opposite. The
> title says " Ensure that searching in Pootle *takes accelerators into
> account" *(my bold)

Note that this won't be 100% accurate, as LibreOffice assigns
mnemonics automatically when they are not set - and there is no way
for pootle to know what mnemonic LO did choose...
(i.e when searching for ~Open, only return results with ~Open (& no
accelerator at all because of the above stated reason), but no results
with O~pen or Op~en or Ope~n.

However, (and that is stated in the bugreport already): when just
searching for "Open"
it should return all variants: no accelerator, ~Open, O~pen, Op~en and Ope~n.

Whether this is possible depends on the indexer that is being used (or
when no indexer is used) - pootle can use lucene and xapian as
indexer. Our pootle is using xapian.

So the requirements are:
* There must be a tokenizer that preserves the mnemonics and doesn't
flatten them out (i.e. make Open out of O~pen - O~pen must be stored
in the index)
* The search must be able to return all matches when no mnemonic is
given. This probably is harder, and probably the only way to do is is
to search for all possible variants. And this implies that a new
option "ignore mnemonics" is needed.

This sounds ugly, thus a custom tokenizer that will flattens out the ~
and additionally stores a "MNEMONIC_O~pen" or a similar construct
sound cleaner.
So searching without mnemonic (the default) will just search the
flattened out string (i.e. ~ removed), and when user activates "[x]
search mnemonic" it will only find the MNEMONIC_O~pen one (depending
on whether ~ is a special character in xapian or not, imagine another
character at the corresponding position instead) - it will not return
the instances where LO did define the mnemonic.
This might as well confuse people, as they see the mnemonic in the UI,
and thus expect search for mnemonic to turn it up, but IMHO this is a
nicer solution.

It is very very hard to please everyone though....

ciao
Christian

-- 
Unsubscribe instructions: E-mail to l10n+help@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/l10n/
All messages sent to this list will be publicly archived and cannot be deleted