Sanity check messages are wrongly exported with every translation

This is probably due to inheritance of osv objects... we'll see if a worthy fix can be found.

On Friday 29 October 2010, you wrote:
> This is probably due to inheritance of osv objects... we'll see if a
> worthy fix can be found.

I believe it is hard to fix.^H^H^H

On a second thought, a patch like this would have proven me wrong on my first
sentence..

Other meaningless strings occurring in every translation :
"The Object name must start with x_ and not contain any special character !"
"Error ! You can not create recursive Menu"
"Invalid model name in the action definition."
It would ease our proofreading duty if those were removed too.
This of course is not an emergency and you may defer it until after v6.0.
Lionel

On Wednesday 17 November 2010, you wrote:
> Other meaningless strings occurring in every translation :
> "The Object name must start with x_ and not contain any special character
> !" "Error ! You can not create recursive Menu"
> "Invalid model name in the action definition."
> It would ease our proofreading duty if those were removed too.
> This of course is not an emergency and you may defer it until after v6.0.
> Lionel
>

Got them, all.

Many spurious terms including ORM constraints messages have been removed from POT files, and the code has been fixed to avoid exporting them (revision 3042 and some additional ones coming soon as well).
Expect all these terms to disappear from the Rosetta web interface soon, if not already gone.

The useless strings are still there in Launchpad Translations (at least for French). Can you do something about it ?
Lionel.

Hello Lionel,

I have double-checked again, and these spurious terms are gone from all the *.pot files now (only one remained in l10n_ch).
I also notice that Launchpad is removing them bit by bit in the PO files, during the daily commits.
These disabled terms are put in comments at the end of the PO files.

As an example, the one in base_calendar/i18n/fr.po was removed by Launchpad in revision 3938 (2010-11-19), and the term had been removed from the POT file in revision 3937 the day before (2010-11-18).
Other files have not been updated yet, for example warning/i18n/fr.po, even though the corresponding POT file was fixed at the same time in revision 3937.

So I am not sure what exactly is going on, perhaps it depends partly on the processing delay for updated POT files, or it waits until another actual translation change is done?

In any case, we should take this question to the Launchpad Rosetta project, and close the current bug, which is indeed fixed: the spurious terms are not exported anymore. I propose to deal with this in question 137354.

missing link to question #137354: https://answers.launchpad.net/rosetta/+question/137354