[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: WC-NG: How can we help with 1.7-readiness?

From: Greg Stein <gstein_at_gmail.com>
Date: Thu, 11 Feb 2010 23:29:26 -0500

On Thu, Feb 11, 2010 at 13:40, C. Michael Pilato <cmpilato_at_collab.net> wrote:
> Hyrum, I know that you (and Greg) have been asked about this before, but
> I've never quite seen a response that really packages stuff up for the
> newbie.  And yes, for WC-NG purposes, I'm a newbie.  That said, I'm sure I'm
> not the only one.  And I'm sure I'm not alone in wanting to jump in and help
> with the WC-NG work but not knowing where to start poking.  So...

Yeah. We've been coding more than documenting, and that is definitely
a weakness. I think we've got a solid "here is where we're going"
basically settled out between Hyrum and I, after all the hacking and a
few in-person brain-melds that have happened. But, frankly, neither of
us have sat down and committed time to doc that. Of course, part of
the problem is that a lot of this is R&D. The target is fuzzy, so a
lot of Zen comes into play, which defies documenting. We sort of
"know" what the end should look like, but also know that documenting
*that* is most likely a waste of time, given the variability of
results within the R&D process.

> What identifiable tasks remain to get WC-NG into a state of 1.7-readiness?
> Are issues filed to track those things so that casual observers can a) be
> aware of them and b) feel empowered to attack them without stepping on
> somebody's toes?

I see several major items for wc-ng completion:

* remove entry_t usage within libsvn_wc/client
* remove access_t usage withing libsvn_wc/client
* move props into wc.db
* move to single/root wc.db
* switch to new pristine file mgmt

Speaking for myself, I have no toes. Jump around all you want.

Stepping back to a meta- level. I recognize my coding level has
dropped, due to EU-US moves, ASF moves, holidays, and Xbox
distraction. I'm working on resetting my day-to-day schedule.

> To what degree is notes/wc-ng/* up to date with the current thinking?
>
> To what degree is notes/wc-ng/* considered complete?

Unclear.

If somebody feels wordsmithy w.r.t those docs, then I'd love to
voice-conf with that person, and haul in Hyrum and Bert. (and invite
the others devs, but concentrate on wc-ng braindump)

> To what degree is the knowledge in notes/wc-ng/* beneficial for the task at
> hand?  Is there a whole different class of knowledge that's needed either in
> addition to or instead of those design notes in order to adequately address
> the immediate WC-NG incompleteness?

We've seen some great success with "not-greg/hyrum" on eliminating
entry_t/access_t. But the knowledge around doing that is not
quantified. It represents a lot of osmotic knowledge.

> I guess another form of these questions could be:  If you were in charge of
> bossing Subversion devs around, how would you divvy up what needs to be done
> on WC-NG and with what would you equip your charges in order to get the job
> done?

Eliminate entry_t and access_t. Work on pristine migration, since that
has not been tackled by anyone, anytime. Outside of those areas, there
will be some coordination issues (w/ myself, bert, hyrum, and other
tasked peeps).

Cheers,
-g
Received on 2010-02-12 05:30:01 CET

This is an archived mail posted to the Subversion Dev mailing list.