Jacob Nordfalk <[email protected]>
writes:

> 2015-05-11 14:41 GMT+02:00 Kevin Brubeck Unhammer
> <[email protected]>:

[...]

>     * If I introduce an error on e.g. line 70190 (or something way
>     down in
>     the file), it jumps to the wrong line number on Validate. You can
>     try
>     this in the bidix on nno-nob; just scroll to the bottom and change
>     e.g. "rpar" to "bug" and it'll jump to line 65535. Oh, that number
>     looks familiar :)
>     
>
> This is a bug in xmlling which apertium-validate-dictionary uses.
> xmlling it gives the wrong line number. If I introduce an error in
> line 70190, and rund
>
> $ apertium-validate-dictionary apertium-nno-nob.nno-nob.dix
>
> I will get:
> apertium-nno-nob.nno-nob.dix:65535: Schemas validity error : Element
> 's': No match found for key-sequence ['xxxxrpar'] of keyref
> 'symbolNameRef'.
> apertium-nno-nob.nno-nob.dix fails to validate
>
> See also
> http://stackoverflow.com/questions/2410806/xml-schema-validation-for-large-schema-files
>  
>
> I guess we'll just have to wait for the fix for xmllint.....

9 year old bug report with lots of arguing and an unapplied patch.
Looks promising :(

>     * If I click Edit Source for an .hfst file it tries to edit the
>     binary …
>     
>
>     * apertium-$lang.$lang.lexc – the lexicon and pardefs
>     * apertium-$lang.$lang.twol – morphophonological rules (like "turn
>     -y
>     into -ie before -s")
>     I'm guessing most of the time, people will be editing the lexc. I
>     don't
>     know if there's an obvious way to assign >1 source per binary; it
>     might
>     not be worth it if the lexc is edited most of the time (and so far
>     automorf.hfst's are the only binaries with >1 source). Maybe some
>     actual
>     HFST users should chime in first at least :-)
>
> Ive added support for opening the .lexc file in SVN.
> I havent tested it, as I dont have the tools for HFST installed.
> Could you provide me with some example pairs and the file names
> generated (ls -l in the pair).

https://svn.code.sf.net/p/apertium/svn/trunk/apertium-kaz-tat

- kaz-tat.autogen.hfst – created from
  languages/apertium-tat/*{lexc,twol}

- kaz-tat.autogen.bin – also created from
  languages/apertium-tat/*{lexc,twol}, but this time compiled with
  lt-comp

- kaz-tat.automorf.bin – created from
  languages/apertium-kaz/*{lexc,twol}, compiled with lt-comp

So … "automorf.bin" means it has to be read with lt-proc instead of
hfst-proc, but doesn't say where it came from originally. On the other
hand, if apertium-kaz/apertium-kaz.kaz.dix doesn't exist, but .lexc
does, then I suppose that should be good enough for auto-detection.


-Kevin


Attachment: signature.asc
Description: PGP signature

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Apertium-stuff mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/apertium-stuff

Reply via email to