bzr-explorer-dev team mailing list archive
-
bzr-explorer-dev team
-
Mailing list archive
-
Message #00029
Re: [qbzr] explorer 1.0 planning & quncommit
Ian Clatworthy пишет:
Hi all,
We aren't far off bzr 2.0 now so I'm going to think seriously about what
timeframe makes the most sense for calling something Bazaar Explorer
1.0. There are still some important things I want to land scope-wise,
e.g. an import wizard/dialog [1], but my gut feel is that qbzr may end
up being the bottleneck. Do others feel the same?
For me bottleneck is RL.
Don't get me wrong: qbzr is progressing in leaps and bounds. On the
positive side, I hope explorer usage has helped increase the testing of
what if offers. On the negative side though, explorer also tends to
highlight what qbzr is missing still. Javier has plugged quite a few of
those gaps in recent weeks (e.g. qsend) but some of his work still needs
review IIUIC. If anyone has time to help with that this week, please do!
I'm planning to send review on Javier's qsend dialog. At first look it
require many tweaks re code style.
So after Javier's stuff is reviewed, tweaked and landed, what's still
missing in qbzr that explorer needs? I believe the list is something
like this:
* quncommit
It sounds like very easy dialog. But it's better to see sketch or draft
of intended UI.
* qshelve
* qunshelve
You talking about builtin shelve? Do you aware that it still does not
work on Windows because of OS locks? So I'm not positive about putting
efforts here.
* qplugins
?
* qcustom (or whatever name we talked about - I forget right now)
qrun maybe?
Of those, I suspect the top one is "mandatory", the next 3 could easily
come later, and the last one falls into both camps. :-)
qrun sounds like simple and useful universal commands launcher, so I'd
rather put it into "mandatory" section. And based on shelve brokenness
on OS of my choice, I think shelve/unshelve will going to the end of the
line.
If others agree, is there *any* chance someone could volunteer to take a
crack at quncommit real soon now? Even a *really* simple version would
be better than nothing I suspect.
As I said before it's very simple to make simple quncommit, but let's
try to get some sketch first (mock in Qt Designer is enough to start).
Ian C.
[1] 10 days ago, I did have plans for coding that on the weekend just
gone but life ended up being more complex than I hoped.
I hope you're well now.
References