Jonas Smedegaard wrote:
> On 06-05-2005 14:04, Jacob Sparre Andersen wrote:
> > Jonas Smedegaard wrote:

> >>The newest releases of the Debian package use dsdo as
> >>source. Is it wrong of me to assume that dsdo should be
> >>able to reproduce the content of myspell?
> > 
> > Yes.  The up-stream package "dsdo" does not generate
> > proper versions of the various application-specific
> > packages (at the moment at least).
> 
> I'd like to patch it instead of maintaining several
> sources, if possible.
> 
> Do you know if there is things more complicated than the
> change of filenames of the *.dat files that I discovered
> now?

Yes.

<http://da.speling.org/filer/building-scripts+data.20050423.zip>
contains the complete building scripts and data for
generating the various versions of the packages.  Feel free
to use them - even if just to get an idea about the
complications involved in generating the different packages.

> I will test before uploading. Anything clever I could use
> for testing apart from looking up "cykelstativernes"?

This script should give you a random collection of composite
words:

   for first in $(lynx -dump -source 
"http://da.speling.org/random_words/?count=10"; | egrep '^[[:lower:]]+$'); do
      for second in $(lynx -dump -source 
"http://da.speling.org/random_words/?count=11"; | egrep '^[[:lower:]]+$'); do
         echo ${first}${second}
      done
   done | sort -u

Most of them will probably me nonsense words, but most of
them will technically be correct Danish words.  If your
spell-checker rejects any of those nonsense words which
technically are correct, then we may have a problem.

And if you don't get any technically correct words the first
time around, feel free to run the script again.  Or use the
words as inspiration for creating technically correct
nonsense compositions.

Unfortunately, it is likely that it also will accept some
incorrect combinations.  This is due to the design of the
handling of composite words in most spell-checkers.

Jacob

PS: Some technically correct nonsense (I hope) compositions:
     + klisterpotteankerspil
     + klientrolleklamphugger
     + straffesagsdeodorant
-- 
Ada - the programming language still ahead of its time.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to