+1 for your comments, Caio!
Rafaella
Caio Tiago Oliveira wrote:
On 7/26/07, Thorsten Ziehm <[EMAIL PROTECTED]> wrote:
> Translation is here a different part. Checking the
>> translation in new features, does not need test cases for
functionality.
>> It is only needed to know, where the changes were done in UI.
> Where we get this information compressed for testing?
The Office is too complex. I cannot tell you, where changes were made.
This is a big problem, but I do not know how this can be solved.
I might to be oversimplifying this, but...
there are a lot of new strings or dialogs.
The same way as the l10n tests are made for the functionality we have
today, there is the need to add *l10n* test sometimes.
One example for 2.3 is the database report. Would make sense to add a
TCM l10n test for this for the OOo Base tests.
Should this test be done before each release? No.
But would be nice to do these tests before the first release after its
integration.
Why? A consistent sdf file does not mean that the translation is right.
Even if the l10n responsible run pofilter on the files, may occur errors.
If not, he could even submit strings partially translated, for
instance (gsicheck would catch only when missing tags or so, it can't
find too short strings).
So... would be nice to someone look at these dialogs and say they are
correctly translated before integrating.
l10n tests on TCM would help to find these cases.
In general, it's nice to review the new strings before releasing.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
