On 21 February 2013 18:27, Andrea Pescetti <pesce...@apache.org> wrote:
> On 20/02/2013 Jürgen Schmidt wrote: > >> 04/08 - 05/15 translation -> I guessed ~5 weeks for updating the new >> strings. This can change when we know exactly how many new strings we >> will have. But translation for new languages can already work on the >> existing po files for 3.4.1. >> > > Looks good, but if the new infrastructure allows it by that moment we > should aim at either having a usable Pootle instance (I mean, one where new > volunteers can work) or all the PO files ready, so that we can point > volunteers to PO files with no need to prepare them explicitly. > Status on l10tools is that I am testing generation of po files at the moment. It is a very timeconsuming work, because I need to be 100% sure that genLang extract at least the same information as the current tools. genLang extracts a bit more information in some cases e.g.: - there a a number of messages marked [en-US] that are currently not extracted - some keyparts are not extracted correctly, typically missing root part of key - Itemlist, is only extracted with itemnumber, but often the items have names For these cases I modify the sdf file (manually correcting the error if you will), but that has to be done for all languages, so that I can later match keys. Once this part is finished, genLang can generate po files directly. genLang can already read po files, but need to be able to merge (extract from source, merge with existing po files) and write updated po files for all languages. The part is only rudimentary programmed and finally genLang will write new source files, containing the selected languages, this part is more or less programmed, but totally untested. Once the community has accepted genLang, we (I cannot do that alone) need to integrate it into the build process. so as you can see, I have a full schedule ahead of me, but I am still confident that I will finish in time for 4.0. Please remember that as soon as the extract works, we have po files directly available in svn...and if everything fails, we can use the current method to convert them back to source files. have a nice day. Jan I. Ps. the l10n commits at the moment is NOT the full picture, but parts I need tested on different platforms, once l10n is ready for others I will announce it on the list. I've see some l10n-related commits (good!), but I don't know where we are > precisely. > > > Important is from my pov that we start talking about the improvements >> that we make in public. >> > > I agree. A couple of screenshot and a brief description are often enough, > and I would use the posts to recruit some new development volunteers too. > > Regards, > Andrea. >
