Thanks, Daniel.
In Catalan there is a very special case because LS is not a correct word
but can be a token as part of a contraction (dels = de + els, als = a +
els...). "LS" gets accepted from suggestions like this, after tokenizing:
<message>Cal escriure: <suggestion>dels</suggestion>.</message>
(Suggestions outside messages seem to be ignored).
One or two grammar rules can be added to solve the issue. Nothing more is
needed.
I don't find specially usefull the method getAllIgnoreWords(). There are
some risks. I think correct words should be in the dictionaries. What
problem are we solving with this? Moreover, I don't understand why it's
triggered only when some rule is enabled or disabled.
In the SuggestionExtractor, this line:
private static final Pattern BACK_REFERENCE_PATTERN =
Pattern.compile("\\\\" + "\\d+");
should be:
private static final Pattern BACK_REFERENCE_PATTERN = Pattern.compile(".*"
+ "\\\\" + "\\d+" + ".*");
to make it more robust. I have found some segments of words converted
unexpectedly in accepted words.
Regards,
Jaume Ortolà
2014-11-29 11:05 GMT+01:00 Daniel Naber <daniel.na...@languagetool.org>:
> On 2014-11-28 23:46, Jaume Ortolà i Font wrote:
>
> > I have found a strange bug. Take the non-existent word "ls" (LS). This
> > happens in Catalan:
>
> This is what happens: unlike the command line LT, the web version
> disables a rule by default (WHITESPACE_RULE). This triggers
> JLanguageTool.getAllIgnoreWords() which loads ignore words from the
> rules suggestions. Rule ANTIPODES has this suggestion:
>
> <suggestion><match no="1"/>ls antípodes</suggestion>
>
> Thus "ls" gets accepted. The idea once was to not have errors on words
> that have been suggested by a different rule. Ideas for fixes welcome...
>
> Regards
> Daniel
>
>
>
> ------------------------------------------------------------------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
>
> http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
> _______________________________________________
> Languagetool-devel mailing list
> Languagetool-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/languagetool-devel
>
------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
_______________________________________________
Languagetool-devel mailing list
Languagetool-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/languagetool-devel
