On August 14, 2004 04:23 pm, you wrote: > > There are several outstanding issues with the livedocs code. As I said > > in my post, last night, HEAD seems broken. Also, Nuno has a number of > > patches available at http://livedocs.aborla.net/ > > It would be nice if Ilia (and/or Wez) could review these patches, and > apply / comment on them.
I reviewed the patches and applied 7 of them, thanks guys. The Cygwin patch was not applied because I have not way to test it and the last person I spoke to who build the manual on win32 via Cygwin had no issues. What in HEAD is broken? > > My point? I'd like to determine the proper procedure for patching > > (getting patches approved for) livedocs, or find an alternative > > development method. Follow, Nuno example :-). He is doing great work, which for the most part gets into CVS once Wez or I have the time to review it. > > I also understand that the original authors of livedocs don't want their > > code messed with. Isn't this the beauty of CVS, though Why commit if it'll be reverted. > > I suspect this idea won't be well-received, but if we're not > > willing/able to keep livedocs HEAD up to date, could we not branch the > > module for testing purposes? Karma could be re-granted on a per-case > > basis, with the condition that non-core developers do not commit to > > HEAD, but to the dev branch. My understanding of the "staleness" of > > livedocs is that Ilia and Wez are busy on PHP 5. I understand this. > > People like Nuno, though, have demonstrated a knowledge of livedocs > > code, and should be able to contribute, IMHO. I'd also like to see > > livedocs tagged regularly so we can easily "rollback" to a previous > > version (last night, while getting livedocs running, and failing > > miserably, I was tempted to start pulling random dates out for cvs up -D > > ). There are some critical bugs in the code in particular dealing with pregeneration logic that fails to create some pages. While it's mostly self contained inside pregenerate.php there are some spillbacks to other files. Until that is resolved I'd like to keep once branch only, once that is fixed we can tag a "stable" release as that's the only major issue I am aware of at this point. Ilia
