openerp-doc team mailing list archive

[Olivier Dony <odo@xxxxxxxxxxx>]

Le 21/02/11 20:06, Don Kirkby a écrit :
> What's up, docs?

;-)


> I'm working on merging the 5.0 branch into 6.0 for the first time, and
> it's a huge mess!

I expected as much, which is why I hadn't found time yet to do it. 
Instead I had done a lame cherry-pick to port Thibaut's contribution as 
discussed in bug 722612 - and apparently my push failed, so it's 
actually fine because you did the proper thing :-)

Thanks a lot for the merge, outstanding job there!


> Can anybody tell me what the heck happened to the
> source/book/2/2_5_CRM images folders on the 5.0 branch? It looks like
> a bunch of stuff happened and then got rolled back. I propose to just
> revert the images folders to let the merge succeed, and then someone
> else can cherry pick the changes if there actually are any that apply
> to version 6.0.

I think what happened is a merge error at revision 437, where the image 
directory was removed by mistake. The next commits (until 441) were 
apparently attempts to restore the original images, with various success.
The thing is that new people started working on our documentation at 
that time, and they have a writer/editor/translator background rather 
than technical. That kind of background is excellent for 
writing/improving documentation, but unfortunately our current system is 
more suited for people with a technical background, as you risk getting 
quickly exposed to VCS/Bazaar intricacies.

Reverting that directory and ignoring it during the merge sounds like 
the most sensible solution to me.


> To avoid this kind of mess when we start working on a separate
> documentation branch for version 6.1, please announce your plans ahead
> of time, and keep new work on the public trunk branch. Then we can
> make a bunch of small merges instead of doing a big one like this.

Definitely agree there. Our new documentation contributors are learning 
bzr gradually, and we will also make sure to assist them in the process 
at all steps, to avoid this kind of mess again.

We also do not despair of finding a tool to make the documentation 
accessible to non-technical profiles, having all the dirty work done for 
them in a user-friendly manner by some integrated tool, including 
suggesting translations, versioning etc.
Unfortunately Rosetta is not up to the job at the moment, and using 
plain PO files for storing the translations sounds quite limited too, 
unless a proper tool can do the reconstruction of translated sections 
transparently.