Lars Aronsson napisał(a):
Is there any way to extract all Swedish text used in the Swedish
version of OpenOffice.org, including menues, error messages and
help texts? Since the current spell checking dictionary is very
primitive, I want to improve it. And one step would be to make
sure that all the Swedish help texts are recognized by the
spell checker.
It's trivially simple. I use gawk script on sdf files. The procedure is
simple enough:
1. Unpack the l10n tarball to a directory.
2. Run #ls -R | gawk -f create.awk > extract.sh
I extract filenames from file1 and build a bash script using a gawk
oneliner "create.awk":
/^\..*\:/{gsub(":","",$0); prevline=$1}
/localize\.sdf/{print "gawk -f print_pl.awk " prevline"/"$0}
The script print_pl.awk can be adapted for any language:
BEGIN {FS="\t"}
{if ($10=="pl") print $11}
Change "pl" to a language you want to check, and use
3. # extract.sh > complete_translation.txt
You can import complete_translation.txt. You'd get a lot of false alarms
on StarBasic but we can build a user dictionary with StarBasic keywords
(that's trivially easy) for spell-checking the whole help. Is there any
other language we should cover?
Is such a test part of the normal procedure?
It seems it could make sense for every language.
See my proposal:
http://wiki.services.openoffice.org/wiki/Automating_Translation_QA
Best,
Marcin
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
