Paul Vriens wrote:
Paul Vriens wrote:
Alexandre Julliard wrote:
Paul Vriens paul.vriens.wine@gmail.com writes:
Something like the attached splits the loop. We now need to pass that array (string) to ver.pl.
Or do you have some other idea's how this could/should be solved.
I don't think we want to add even more information to pass between scripts, there is already too much of it IMO. We should probably try to merge scripts instead.
Also I'm not sure about skipping of the kernel32 locale files; it seems to me we could just as well use that as the reference list, and display every known language and sublanguage. At least it would make the index page much more impressive ;-)
I actually was thinking the same about both items.
The downside of using the locale stuff is of course that our overall translations percentage goes down the drain ;) The upside is that we only need to do one --verify-translation run on kernel32 to get our list of possible translations.
I will have a look at merging checkmakefile.pl with ver.pl (already did that while testing).
I played around a bit and some of the results are shown on the attached screenshots.
Is this what you mean? The index page will (in the end) not show any neutral languages anymore.
We probably need some nicer sorting, but that looks trivial. Another idea is to have English (United States) as some kind of header instead of adding it to the table? All other translations are after all compared to English (US).
Please don't look at the columns/numbers, the graphs are correct.