Re: [Texmacs-dev] Stable release schedule?

[David Allouche]

On Fri, Dec 12, 2003 at 01:06:16PM +0100, Joris van der Hoeven wrote:
> > In that process I have been making a few minor improvements for my own
> > convenience and I have been systematically fixing some bugs which have
> > been reported for years and never fixed and other problems which caused
> > the software not to match the documentation.
> 
> That is good too, but please don't forget to make progress
> in the documentation of the TeXmacs primitives in parallel :^)

Not yet started this. I agree I should just go on and read the existing
doc so I can start writing primitives documentation soon. I did not
expect that working in a sequential order would involve that much
bugfixing.

> > The fixes to the keyboard wildcarding and variants system are especially
> > important, but they are also non-trivial and could use an extra public
> > release for testing.
> 
> Unstable fixes will be postponed until just after the next stable release.
> With the new release policy, they can easily be incorporated into
> subsequent releases. So if you or I are not very sure about the fix,
> then they will be postponed. Nevertheless, there will still be a few
> unstable releases, so the earlier you send me your patches, the earlier
> that they will be incorporated.

I am pretty confident in the keyboard wilcard fixes. I have been using
experimental fixes (widthdrawn) which actually had problems until I
eventually understood how the whole thing was supposed to work.

> > Clearly, all of those fixes should get into the next
> > stable release. I think it would be a good thing to wait until I have
> > finished the current review before releasing 1.0.3.
> 
> The 1.0.3 version is planned for januari. I do not plan to delay
> this moment. In any case, not all minor or even major bugs will
> be fixed by then anyway. But the more we fix, the better...

Like january 1st or january 31st?

> > Since Savannah is currently down, I am holding those patches back, but I
> > can post them on the mailing list.
> 
> Yes, that is fine. You may also put them on your website and
> just send us a message; that makes it easier for me to download them.

I regularly post my patches on my arch archive, on my website.

If you are still reluctant to even _give a try_ to arch, you can
retrieve the actual changeset files. They are tarballs containing (among
other things) simple patches for modified files and full-text copies for
new and removed files. So you can use them just with tar and patch.

I just finished checking patch application and updating to 1.0.2.10. I
have not yet actually tested that the patches for 1.0.2.9 actually
_work_ with 1.0.2.10, but I do not see why they would not.

The changesets for my current branch can be found there:

  
http://ddaa.net/%7barchives%7d/2003b-texmacs/texmacs/texmacs--ddaa/texmacs--ddaa--2.10/

Note that the base-0 revision contains nothing interesting. Just look
into the patch-N directories. Please do no pick random patches for which
I have not asked for a merge. I may back-out some experimental changes,
if you have already applied them, that would be extra unecessary work
for you.

The current mergeable patches are:

patch-1
    missing conversion file
patch-2
    tmhtml: entities for non-ASCII chars (2)
patch-3
    edit: make the hybrid return key extensible
patch-5
    S-F5-" text shortcut for ASCII double straight quote
patch-6
    fix in-hungarian? mode predicate
patch-7
    footer message for complete hybrid commands
patch-9
    keyboard wildcard backtracking
patch-10
    multi-esc for hyper
patch-11
    variant simplification with unused modifiers
patch-12
    set multiple cell formats of a selection at once

patch-4 and patch-8 are just administrative stuff for Arch.

BTW, patch-1 is a missing file from a patch that you failed to apply
properly _twice_. You should be a bit more careful.

As a final note, thanks for properly separating each applied patch in a
different CVS commit. That makes it _much_ easier for me to test the
patches were correctly applied. The missing file just slipped in the
only commit which grouped several patches. If you had not grouped the
patches you would have noticed that nothing had been changed.

-- 
                                                            -- ddaa