[Gnu-arch-users] tla1.1 plans

[Tom Lord]

    > From: Karel Gardas <address@hidden>

    > now when inode signature optimization is in place and also other important
    > features, have you considered to do real tla1.1 release soon? i.e. after
    > intensive testing of course...

    > Or do you plan to write some other new features for 1.1?

    > [....]

    > PS: I think you can simply reply to arch-devel, since I guess other users
    > might also be interested in your future plans.


The original agenda for 1.1, aside from simple bug fixes, was:


        *) the inode signature optimization

        *) =tagging-method generalization

        *) fancier support for "partial commits"


The agenda changed in a few ways as development proceeded.
First, a new item was added:


        *) CLI improvements

           This was added because (a) it was needed and (b) several
           people started contributing such improvements.


And this item has changed:

        *) =tagging-method generalization

           I am beginning to think that the "full plan" in src/tla/=TODO
           is excessive -- the recently added `untagged-source'
           directive may be quite sufficient.   Therefore, I'm leaning
           towards not implementing the full plan for 1.1.


Other than design-note scribbles, there has been no progress on this
item:

        *) fancier support for "partial commits"


Something that sorely needs to be done suggests adding a new item:

        *) update the tutorial, front to back




So here is what I am thinking about 1.1:

a) "Sometime in October" is when 1.1 should happen, barring surprises.

b) (Mostly) clearing the issue database is one goal for that release.
   I don't see any problems on the horizon about that.

c) I'm wondering how people are feeling about CLI changes.
   Contributions in this area have slowed down now, which makes 
   me think we're in pretty good shape.

   Perhaps it would be reasonable to freeze out incompatible CLI changes 
   in the 1.2 series?

d) I have to make a decision about partial commits -- can I do
   something reasonable in this area in time for an October 1.1.
   I'll keep the list informed about that.

Putting all this together, the summary is:

        *) 1.1 in October

        *) (Mostly) clear the bug database

        *) Revise the tutorial

        *) Maybe add fancier partial commits


Looking ahead to 1.2, no decisions have been made, but some
off-the-top-of-my-head thoughts are:


        *) 1.2 in January or February (knock on wood)

        *) Further performance hacks as convenient

        *) Towards the "overarchish", meaning, closer integration 
           or initiation of some subset of:

                a) inclusion of Perspective with the project
                b) inclusion of ViewArch with the project
                c) inclusion of Emacsen-modes with the project
                d) perhaps an arch-friendly bug tracker
                e) perhaps `tla grab' on steroids
                f) inclusion/improvement of tools for "mirroring 
                   an upstream CVS" or "bidirectional exchange
                   with upstream CVS"
                g) perhaps email/usenet changeset distribution
                h) perhaps changeset factoring tools

        *) perhaps a delta-compressed alternative to revision
           libraries

        *) perhaps finer-grain mirroring (e.g., I got patch-N from
           that archive but don't have patch-(N-1) -- but let me
           start a mirror anyway)

        *) scripting and hook work to kill pristine trees

        *) perhaps demonstration mirrors of some non-arch projects


Meanwhile, I still seem to be ramping up on my tiny Scheme-ish/Lisp
project.   So, lemme both put in a plug for that (when it comes around
I'll be louder about it) and mention that it may turn out to be useful
for some of the overarchish work, if the timing works out.

-t