[Idle-dev] Re: merge DS_RPC_BRANCH back to main?

[Stephen M. Gava]

> It would be good to get more eyes on it.

Continuing all work on main now would accomplish this.

> I was thinking that since it's been a year since there was a release of the
> Visual Python version of Idlefork, that they would appreciate another release
> of the DS_RPC_BRANCH which contains all the work you have done to date and the
> merges of Python Idle bugfixes.  

Hmm, I think there is a fundanmental misunderstanding here. There is no
vpython version of idlefork, and there hasn't been since the 0.8.1
release. I'm not talking semantics here, I mean that any pretense of a
semi-stable relationship to that earlier codebase was dropped long ago.
That was the last idlefork with no drastic changes from previous
still-relatively-close-to-vpython versions. Since then there have been
_major_ changes to many areas of idlefork that so far are untested
except for use by developers and a small number of cvs checkout users.
The entire configuration backend has been rewritten and replaced as the
underpinning for the entire new configuration frontend, for instance. I
expect there are minor (but possibly major, hopefully not) breakages all
over the place as a result of all this new code.

> So the cleanup work you are doing is quite
> appropriate for a "final" release.  

My intention is that I am tidying up (and also still developing) for a
(hopefully reasonably near future) series of alpha/testing releases,
which will include the new rpc stuff and the new config engine.

> The new merge from Python Idle you
> mentioned (PEP263??), would that be useful to the VP branch?

All these merges need to be in main so everyone can work there and test
by default. I don't really think that with a developer base of our size
(really only two who are very active) we can indefinitely maintain two
development streams.

> Once we merge to MAIN, I suppose there would only be critical bug fixes on the
> DS_RPC_BRANCH.  Then we need to convince Scherer/Sherwood et. al. (via config
> options?) that the MAIN line of development is a good way to continue for their
> purposes.

The vpython folk are in the same boat as everyone else, when there is a
new idlefork release that is at least beta quality they will be able to
consider whether to adopt it or not. The intention is clearly that
idlefork will eventually become python/idle and then python/idle will be
suitable for all idle users including vpython. The idea is to end the
forking of production idle, and the need for any forking, not to keep it
going. 

We agreed on the DS_RPC_BRANCH as a place for non new-rpc development to
move to while (if?) things were drastically broken by moving in the new
rpc stuff. My impression is that it's moved in now and into ongoing
development (correct me if I'm wrong though). I don't want to keep our
tiny development team (of mostly two) spread indefinitely over two
branches. We need to both be focusing on one version now (or asap at
least) so it can all come together for some test releases. 

Stephen.
-- 
Stephen M. Gava <elguavas@python.net> http://python.net/crew/elguavas
IDLEfork http://idlefork.sourceforge.net "just like IDLE, only crunchy"