[PEAK] Trellis future (forking)

[Sergey Schetinin]

Thanks for the responses everyone. I wish there were more Trellis
users out there though :) Is it possible they are not subscribed to
this list? Does it make sense to set up a new, better visible one on
google groups?

Also thanks for the pointer to wxpita, I like looking into various GUI
frameworks internals.

AFAICT the interest is mostly in application of Trellis to GUI,
networking and ORM as well as some more docs and examples. I know I
spent absolutely *huge* amount of time trying to understand Trellis in
its entirety and I'm pretty sure it can be made easier to get into
both as a user and as a developer of the core / new cell types etc.
Also Phillip kind of dives right into magic with his TempConverter
example while a more gradual introduction would certainly have saved
myself from a lot of issues and wasted effort. I think I can do that
for others.

While I have absolutely no plans to work on ORM integration or
anything of that kind I do have a lot of code to make GUI and
networking easier. I will not open-source all of it for now, but will
certainly move some of the building blocks to Trellis-fork. For
example there will be a much easier way to integrate with wx event
handling but not an entire GUI library. I'll also probably move my
select()-based eventloop to the fork and would love to see people
implement poll, kqueue and libevent based ones.

As of current status, if anyone is interested -- sure, go ahead and
look into the fork differences, but after I complete a few more
planned changes I'll write a detailed explanations of what's different
and how to adapt your code to it, something which should make the
changes in the source clearer. After that I'll dive back into my
products for some time (which will probably mean yet more changes to
trellis-fork) and return to writing docs and examples when I have a
more free schedule.


Also, a question for Phillip (if he has read this far :), may I change
the license to MIT? Google Code Hosting seems to require projects to
use one of predefined licenses and PSF and ZPL are not on the list.



On 2009-06-17, Grant Baillie <grant at osafoundation.org> wrote:
>
>  On 12 Jun, 2009, at 04:22, Sergey Schetinin wrote:
>
>
> > Hi,
> >
> > I was thinking about it for months, but I think it's time to fork
> > Trellis -- Phillip is busy with other things and I don't want to
> > bother him with endless requests. Trellis the way it is now is a great
> > start but it could use a lot more work -- I'm willing to put that work
> > in and I think Andrew Svetlov would participate as well.
> >
> > I'm about to launch a number of projects into production and I want to
> > make sure all of the dependencies are well maintained -- for Trellis
> > that seems to mean it needs forking. I also think Trellis needs more
> > that just bugfixing, so I'm not sure if there's way not to fork it.
> > Well, maybe except for taking over it completely, but it's unlikely
> > anyone would want that and I still want to see where Phillip want to
> > take Trellis himself -- clearly I can't pretend I'm as good as he is.
> >
> > To avoid conflicts I'd change the import names to
> >
> > from trellis import *
> > from trellis.activity import *
> >
> > etc.
> >
> > Some things I think I'll do right away after forking:
> > * apply the bugfixes I've sent before
> > * drop py2.3 support
> > * add more comments to source code to make it easier to understand
> > for people just learning the system (and when coming back to it after
> > a while), I'd also rewrite some parts for clarity but without
> > compromising performance.
> > * add some more practical examples
> > * reorganize source code:
> >  * comment out some classes not reported to be used in actual
> > projects: testing is good but not enough.
> >  * add utility modules, for example wx utilities (WXEventLoop would
> > move there as well)
> >  * reorganize test suite
> >
> > There's plenty more, but that's enough for a start I think.
> >
> > So my questions are:
> > * any reasons I shouldn't do this?
> > * is "trellis" as package name ok? (it would be trellis-fork on pypi
> > and similar)
> > * can this mailing list be used for discussion of this fork?
> >
> >
> > I also wonder how many users Trellis has. If you use it, please say a
> > few words about your project at least if it's in a hobby / academia /
> > business software / other class. Thanks.
> >
>
>  Hi, Sergey
>
>  I just caught up on this thread today: trellis-fork looks pretty awesome.
> I'll have a look at the changes you posted earlier today, and send some
> feedback.
>
>  I have a couple of projects that use trellis (and that I'll look into
> porting over to trellis-fork). One is the original client for trellis:
> <http://chandler2.osafoundation.org>, that was funded as
> far as the "clunky prototype phase" :). I'm pursuing that, and another
> desktop GUI project of my own, as mostly hobby projects at this point.
>
>  --Grant
>
>
>  _______________________________________________
>  PEAK mailing list
>  PEAK at eby-sarna.com
>  http://www.eby-sarna.com/mailman/listinfo/peak
>


-- 
Best Regards,
Sergey Schetinin

http://s3bk.com/ -- S3 Backup
http://word-to-html.com/ -- Word to HTML Converter