MozillaTranslator v5.23
Ricardo Palomares Martínez
After about a month using it myself and adding/solving nitpicks, I've
finally managed to pack and upload a new version of MozillaTranslator,
v5.23:
* MozillaTranslator no longer complains on empty region JAR files
when writing XPIs for products without region JAR files (bug
445558).
* Added preference to set fonts for Edit Phrase dialog and table
views, as some L10n teams were having problems with DialogInput and
Dialog fonts (bug 445557).
* MozillaTranslator now can check for different ending characters in
original vs. translated strings (bug 439352).
** WARNING **: this MAY raise a number of errors in your
translation. Fixing them may require filing bugs to request
approval, since L10n for Firefox 3 is frozen.
* MozillaTranslator properly handles now newline escape sequences
(bug 439349). Properties file escape sequences in the form \r\n or
\n were wrongly removed and parsed as real newlines, which was
wrong, despite it works and looks nicer in the Edit Phrase dialog.
This has been solved.
** WARNING **: this WILL raise a number of modified strings in the
properties files. You will need to fix the corresponding
translations, which may require filing bugs to request approval,
since L10n for Firefox 3 is frozen.
* MozillaTranslator now parses INI files as Properties files (bug
420450).
** WARNING **: this means that, if you had translated the file as a
"MT_UnknownFormatFile" node, you will lose the translation. To
solve it, make sure you have exported the translated file before
updating MT, and import the translated file after updating MT to
v5.23.
* Images no longer appear as modified on every Update Product run
unless they are really modified (bug 415975)
* MozillaTranslator now properly manages weird Firefox/Thunderbird
ab-CD.JAR structures (bug 383914)
You can get the ZIP containing the new JAR and updated userguide at:
http://sourceforge.net/project/showfiles.php?group_id=18842
Ricardo
Axel Hecht
> Hi,
>
> After about a month using it myself and adding/solving nitpicks, I've
> finally managed to pack and upload a new version of MozillaTranslator,
> v5.23:
Hi Ricardo,
thanks for the updated version, nice to see those fixes get out to
localizers.
And also for the note on getting the resulting patches into bugs and
into my approval queue ;-)
Axel
flod (Francesco Lodolo)
I've just installed MozillaTranslator 5.23 (thanks for your great work!)
and I have a question about the QA query "Check Variables".
Since I don't know if a similar change could be handy for other
localizers, I thought it was better to ask here before creating a new
useless bug.
Could it be possible to exclude from this check html entities like &?
Francesco
Ricardo Palomares Martínez
You can do that in File -> Setup -> General -> QA related preferences
-> DTD entities ignored in Check Variables.
You can provide a list of entities to ignore for original strings and
another list for translated strings. This way you can exclude "&"
for original strings, since "&" in en-US won't usually translate to
"&" in other language (that's certainly the case for es-ES and I guess
also for it and fr, for example). In a similar way, if you want to use
entities for some specific characters to your language (like "¿" and
"¡" in Spanish), you can exclude them from translated strings checks.
Hope this helps and you don't need to file a new bug. ;-)
Ricardo
flod (Francesco Lodolo)
> You can do that in File -> Setup -> General -> QA related preferences
> -> DTD entities ignored in Check Variables.
>
> You can provide a list of entities to ignore for original strings and
> another list for translated strings. This way you can exclude "&"
> for original strings, since "&" in en-US won't usually translate to
> "&" in other language (that's certainly the case for es-ES and I guess
> also for it and fr, for example). In a similar way, if you want to use
> entities for some specific characters to your language (like "¿" and
> "¡" in Spanish), you can exclude them from translated strings checks.
>
> Hope this helps and you don't need to file a new bug. ;-)
Francesco
Michael Wolf
thank you for your great work.
But, there is still a bug with QA Queries --> Redundant Strings. In the
java/MS-DOS window following error message appears:
> Exception in thread "Thread-3" java.util.MissingResourceException: Can't find re
> source for bundle java.util.PropertyResourceBundle, key button.copy_to_all.label
>
> at java.util.ResourceBundle.getObject(Unknown Source)
> at java.util.ResourceBundle.getString(Unknown Source)
> at org.mozillatranslator.kernel.L10nManager.translate(L10nManager.java:6
> 0)
> at org.mozillatranslator.kernel.Kernel.translate(Kernel.java:139)
> at org.mozillatranslator.action.CopyToAllInRedundantView.<init>(CopyToAl
> lInRedundantView.java:42)
> at org.mozillatranslator.gui.RedundantViewToolbar.getToolbar(RedundantVi
> ewToolbar.java:49)
> at org.mozillatranslator.gui.ComplexTableWindow.<init>(ComplexTableWindo
> w.java:100)
> at org.mozillatranslator.runner.ViewRedundantRunner.run(ViewRedundantRun
> ner.java:103)
Seems, that the ressource key "button.copy_to_all.label" isn't found
which is defined in file
mozillatranslator.action.CopyToAllInRedundantView.java in line 42.
Regards
Michael
Ricardo Palomares Martínez
> Hi Ricardo,
>
> thank you for your great work.
>
> But, there is still a bug with QA Queries --> Redundant Strings. In the
> java/MS-DOS window following error message appears:
> Seems, that the ressource key "button.copy_to_all.label" isn't found
> which is defined in file
> mozillatranslator.action.CopyToAllInRedundantView.java in line 42.
Thanks for the pointer. I don't usually run Redundant Strings query,
so I hadn't noticed. Funny thing is, I don't really understand which
is CopyToAllInRedundantView.java used for (Redundant Strings query is
actually implemented in
org.mozillatranslator.runner.ViewRedundantRunner.java). It is called
to implement a button ("Redundant Strings"), but I'm not sure the
associated action is useful at all.
Even funnier and more interesting, while debugging those classes I've
realized that traversing the whole tree is waaaaay quicker than I ever
thought it would be (at least in Linux with JDK 1.6.0_7). Implementing
some kind of auto-complete suggestions or similar results could be easy.
So, unless you or someone else desperately needs accessing Redundant
Strings query, I could spend some more time digging into that code.
Ricardo
Michael Wolf
> Thanks for the pointer. I don't usually run Redundant Strings query,
> so I hadn't noticed. Funny thing is, I don't really understand which
> is CopyToAllInRedundantView.java used for (Redundant Strings query is
> actually implemented in
> org.mozillatranslator.runner.ViewRedundantRunner.java). It is called
> to implement a button ("Redundant Strings"), but I'm not sure the
> associated action is useful at all.
Seems that the toolbar button "Copy to all" is defined in
CopyToAllInRedundantView.java line 42. I compared this file with that of
MT 513 where the feature "Redundant Strings" was still working. For any
reason the line
super(Kernel.translate("copy_to_all"), null);
was changed to
super(Kernel.translate("button.copy_to_all.label"), null);
But, in the file resource/userinterface.properties the string
copy_to_all=Copy to All
remained unchanged and button.copy_to_all.label wasn't defined.
Therefore MT doesn't find the resource. The new line
super(Kernel.translate("button.copy_to_all.label"), null);
should be changed back to the old one
super(Kernel.translate("copy_to_all"), null);
Then the button "Copy to all" will appear in the MT toolbar. Its task is
to copy a certain string value for multiple strings to the other
occurrences of a string so the translations will be uniform.
Example:
In Upper Sorbian you can translate Spanish by adjective (španiski) abo
by substantive (španišćina) - "español" resp. "el español". To make the
translation uniform you can choose one of them to copy it to all other
occurrences of the string "Spanish".
Michael
Ricardo Palomares Martínez
> Seems that the toolbar button "Copy to all" is defined in
> CopyToAllInRedundantView.java line 42. I compared this file with that of
> MT 513 where the feature "Redundant Strings" was still working. For any
> reason the line
>
> super(Kernel.translate("copy_to_all"), null);
>
> was changed to
>
> super(Kernel.translate("button.copy_to_all.label"), null);
Yes, I tried to set a naming standard for resources, but I soon
decided it was too much work for no real gain.
> Then the button "Copy to all" will appear in the MT toolbar. Its task is
> to copy a certain string value for multiple strings to the other
> occurrences of a string so the translations will be uniform.
I see now. I had chosen another string to name the button, which
completely misleaded me. Thanks again for the pointer.
Ricardo
Robert Kaiser
> Strings query, I could spend some more time digging into that code.
I really like that query from time to time, but don't desperately need
it right now. Better have a right fix in a few weeks than a somewhat-fix
right now ;-)
Robert Kaiser
Ricardo Palomares Martinez
Actually, I have fixed it already (although there is no release yet
with it). I changed the way it looked it up for duplicated original
strings and I'm pretty sure you, with your overpopulated glossary,
will appreciate the performance gain. :-)
Ricardo
--
If it's true that we are here to help others,
then what exactly are the OTHERS here for?
Robert Kaiser
> (bug 439349). Properties file escape sequences in the form \r\n or
> \n were wrongly removed and parsed as real newlines, which was
> wrong, despite it works and looks nicer in the Edit Phrase dialog.
> This has been solved.
There is an additional problem - if you have \r (carriage return)
characters in your localized string, MT will not show them in any way on
the screen. Still, you can delete them if you know where to find them. :)
Robert Kaiser
Ricardo Palomares Martínez
Honestly, I detected those while running cvs diff before committing
(at least in Unix, they show up as weird characters), and solved it on
case-by-case basis.
If there are too many affected people having problems to solve it that
way, I could look for a programmatic solution, but I'd rather work in
new features than coding something that will be almost one-use. What
do MT users think?
Ricardo