On Friday 13 December 2002 9:29 am, Lars Gullik Bjønnes wrote: > Angus Leeming <[EMAIL PROTECTED]> writes: > | On Thursday 12 December 2002 9:50 pm, Lars Gullik Bjønnes wrote: > | > Angus Leeming <[EMAIL PROTECTED]> writes: > | > | Incidentally, I have patches for many of the xforms bugs that > | > | currently affect LyX. I really think that rolling out an xforms tree > | > | is a good idea. Once we've proved they work, it'll be easy to get > | > | them rolled into the official xforms source. > | > | > | > | Lars, could you set one up? > | > > | > Sure I can do that, but I do not want this is our lyx repository, so I > | > will instead setup a separate repository. > | > > | > Link to sources? > | > | Sounds fine to me. Sources at: > | ftp://ncmir.ucsd.edu/pub/xforms/OpenSource/xforms-1.0-release.tgz > > But I also share some of Jean-Marcs worries. > (I do not want this to turn into a fork of XForms, it just isn't the > nice thing to do...)
[ Angus goes away and reads JMarc's posting... ] On Friday 13 December 2002 9:05 am, Jean-Marc Lasgouttes wrote: > I think that instead of maintaining a tree, it would be better to > maintain a set of patch, so that we can feed them easier to spl. > > We could of course distribute xforms-1.0lyx[123...] on our web site. > If we have a cvs tree, we'll maybe be tempted to fix too much in it :) Ok, I may as well answer both of you at the same time, so I'll send this back to the list. You saw my posting to the xforms list in which I asked how the future would pan out. I think that it can be summarised as "interested but no time for active development". That means also that anyone interested in fixing the bugs in the xforms source should make absolutely sure that those fixes don't actually break things. So, as such an interested "fixer" I have to ensure that my patches work. I'm not sufficiently confident in my coding abilities to say that I get everything right first time, but I do think that I'm not hopeless. That being said, more than one tester is always a good thing. The LyX project has an active interest in fixing bus in xforms. The LyX project is active and the xforms one is not. We should be able to produce squeaky clean patches to xforms bugs that can then get rolled back into the official xforms source without any problems. I too do not want this proposed tree to turn into a fork. I envisage it as a "sort of" branch of the official xforms cvs tree with working patches merged back into the official source. Note that no such official xforms cvs tree exists and who knows when it will. Steve has said he'll set one up, but his time is limited. I don't envisage the two sources diverging much. I really don't think that there's that much to patch. By and large, I'm proposing bug fixes and code clean-ups rather than major structural changes. Logistically, a tree seems easier to maintain than a set of patches that must be applied in a certain order. Given the above, are there any real reasons against doing this? (It is Friday the 13th though. Maybe I have a hidden, devel-ish agenda. Sorry, bad pun ;-) Angus
