Weekly status roundup:

*What Got Done?*
Sadly, this list is not nearly as long as we'd like.

   - Basic proficiency with git has been obtained. It's going to take a
   while before I'm fully comfortable with it.
   - The bitbucket repositories have been fully moved to sourceforge.net,
   and deleted from bitbucket.org.
   - The website migration is underway, though I have some style questions
   for everybody. See blow.
   - The SF.net Tracker has not been configured. I'm still waiting for the
   Migration API to be publicly documented. Mark: Do we have any status on
   this?
   - The SF.net add-ons have been activated, so we can look at the process
   for migrating Trac for TG1 at least.

*Where are we now?*
This is a mixed bag. Since we have the repositories on sf.net, we have made
progress. The web area is a bit of a mess, though, and I'll explain that
below in its own section. It also appears that we have a significant issue
occurring in SQLAlchemy 0.7beta and TG2.1. I have a proposed solution that
I'll document below.

*What is on the slate for next week?*
The primary task on the plate for next week is getting the website migrated
over. Hopefully, by the end of the week, we'll also see documentation on the
SF.net Tracker Migration API, so we can begin planning this move.

*Web Style Question:*
So, I've downloaded the current website's source files, and have discovered
some things that surprised me.  For one thing, the pypi for the various
versions of TG2 is actually 650M. The docs are their own (fairly large)
fileset. What I'd like to do is split out these components into their own
projects, and make them easily sync'ed to SF.net's web space. End result is
that we'll see a URL layout lke this:

* http://pylonsproject.sourceforge.net/
** tg2
*** downloads
**** 2.0
**** 2.1
*** docs
**** 2.0
**** 2.1

This will result in the simplest set of repositories, I think: one for
downloads, one for docs, one for the main webpages. I'm also going to switch
the downloads around so that we're using symlinks, instead of full files.
I'm pretty sure we can get the total size down to a mere 40-50 megs, total.

*TG2.1 / SQLAlchemy Bug*
So, it seems that, with 2.1 and SQLAlchemy 0.7 beta, we have a bug. For now,
the short term fix is to make sure to get users to install 0.6.6. The only
question is *how* to make this happen.

I've seen suggestions that we go with a modification that will make SA 0.6.6
into a required version. I'm not fond of this idea, since it leaves open
other possible areas of issue (for instance, if some other library gets an
incompatible upgrade).

My initial response was to tell people to use our private PyPI. The
immediate response is "They don't do it now, and we tell them to do so, why
will they start listening?" So, I have a different proposed solution: We
change our setup.py to include our private PyPI in the list of places to
look. As much as possible, we even force it to be the *first* place to look.

Now, people can make end runs around this issue, and it won't fix existing
applications, but if we put this into our quickstart, it will take care of
the issues pretty nicely. As an added bonus, it also prevents future
problems. The drawback is that, since we're in the middle of a migration,
and URLs are changing, I wouldn't want to push this update until at least
the web migration is complete.

What do you all think? Good idea? Bad? Indifferent?

-- 
Michael J. Pedersen
My IM IDs: Jabber/[email protected], ICQ/103345809, AIM/pedermj022171
          Yahoo/pedermj2002, MSN/[email protected]

-- 
You received this message because you are subscribed to the Google Groups 
"TurboGears Trunk" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/turbogears-trunk?hl=en.

Reply via email to