What's missing for 4.0

Andreas Pakulat apaku at gmx.de
Wed Sep 16 20:06:26 UTC 2009


On 16.09.09 16:08:21, David Nolden wrote:
> Once again, I think it would be good to re-evaluate what features we need to 
> implement before we can release KDevelop 4.0.

I've tried to tag things in bugzilla with Target 1.0.0 and 4.0.0
respectively, would be good to go over that list and check what might be
missing or what might be possible to drop.

I haven't been able to go through the wishes again (as I did for bugs
recently) though.
 
> So, the stuff that I think we need:
> - User friendly session management ("Close Session" instead of "Close All 
> Projects", etc.)
> - A project-model selection mechanism (the dolphin stuff) that is merged with 
> the build-set, has a default-selection based on the currently open file, and 
> is used used as default for running/debugging
> 
> And of course massive bug fixing (including the debugger).
> 
> What's your list?

I don't really know. I guess I'm using kdevelop too little. 
One thing we definetly need - IMHO - is better integration between
project tree and vcs support (deleting a folder should delete it from
the VCS, adding a file to vcs should add it to the project). Another
vcs-related thing is being able to checkout a project from a repository,
possibly with a repo-browser-toolview.

Another item is the grepview, it works, but it doesn't play well
together with projects. I don't have a clear plan on that though.

Providing EditorContext for the editor context menu so vcs actions are
in it (Or does that work already?) and more importantly provide a
FileContext for the filemanager so we get a fully-loaded context menu
there.

Another thing we have to decide is what to do about mercurial and git
support. I know mercurial got some love a few months ago, but I have no
idea what its state really is. I do know that git's state is pretty bad
and I also know that if I'd work on it we'd need another 3-5 months to
finish up as I'd definetly want to rewrite the code in parts.

Then there's documentview, coverage, the whole veritas/xtest stuff,
valgrind and cvs. I never tried cvs and I think its a plugin that should
rather be in extragear - who uses cvs these days anyway ;) valgrind
plain doesn't work (AFAIK) and only the memcheck is properly supported.
veritas/xtest is unmaintained since almost a year, its hard to use and
IMHO clear candidate for removal. Same with coverage I think. Leaves
documentview, it works and has only few (if any) bugs AFAIK, but should
move to the platform IMHO.

Thats about it. I think spicing up grepview would take 1 full weekend,
the context-menu-stuff only a day at most. The vcs stuff is a bit more,
probably doable in 2-3 weekends (features, not fixing up bugs found
along the way ;)

I'd also like to point out that I think we'll need about 3 months of
feature freeze for bugfixing. There's already quite a pile of bugreports
and not all of them are so easy to fix. So I don't think a release this
year is very probable, rather january next year. This means we could aim
at releasing with KDE 4.4 (and potentially even requiring it if there
are important features/bugfixes to use).

Andreas

-- 
You will overcome the attacks of jealous associates.




More information about the KDevelop-devel mailing list