[petsc-dev] close ts-fixes branch?

[Satish Balay]

On Thu, 25 Mar 2010, Kai Germaschewski wrote:

> On Thu, Mar 25, 2010 at 12:00 PM, Lisandro Dalcin <dalcinl at gmail.com> wrote:
> 
> > On 25 March 2010 01:14, Kai Germaschewski <kai.germaschewski at unh.edu>
> > wrote:
> > > Fine with me (I'm the one who introduced it originally).
> > >
> > > My off-topic 2 cents: I hate reply-to headers.
> > >
> >
> > BTW, I think this is an example of bad usage of mercurial branches...
> > You should use a separate clone for these kind of tasks. Mercurial
> > branches would make more sense for release versus development,
> > although even PETSc does not uses this model. IMHO, creating a branch,
> > push some fixes, merge back to default, and close branch is just
> > repository pollution.
> >
> 
> I actually disagree on this. The ts-fixes branch was a separate clone in the
> beginning, actually. I did go through the efforts of separating this into a
> sequence of patches, instead of having one "this one patch fixes all
> approach", and whereas maybe for that particular case it could be considered
> overkill, in general I consider it highly useful to break up changes into
> pieces that can actually be verified by looking at them.

agree to this.

> So the way I see it, there are basically three ways to merge a change like
> this (or any which is a sequence of patches) into the main repo:
> 
> 1) do them in a separate branch, then merge that branch (and close the
> branch, which I didn't but should have (mercurial used to not even support
> this, a major shortcoming...)
> 2) do them as a sequence on the top of head directly. Unfortunately, that
> goes wrong if someone else changes head while working on the sequence of
> patches. In that case, you end up with two branches which will merge with a
> merge changeset. That's effectively the same thing as 1), branching and
> joining, except that the branch doesn't have an explicit name, but I don't
> really see the difference. If you guys don't like the explicit naming, I can
> certainly adapt to that...

generally in-repo-branch and a clone should have similar workflow and
features. the eqivalent of the 'branch-name' for a clone is the 'clone-url'.

For eg: for petsc-dev use use http://petsc.cs.iit.edu/petsc/petsc-dev, and for
petsc-3.1 its http://petsc.cs.iit.edu/petsc/petsc-release-3.1

[whereas if this were implemented as repo-branch - there would be a
branch-tag]

So one can do either in-repo branches or clone branches [with similar
workflow] - so I prefer clone branches.

> 3) Just import the patch (with some manual trickery) into head as a single
> changeset -- for obvious reasons, I don't think this is a good way, and
> history is lost.

Yes this is not good.

> 
> Both 1) and 2) have the disadvantage of introducing potentially lots of
> merge changesets if development goes on for a while while head moves, too,
> and one keeps the branch in sync. git handles this case by doing a "rebase"
> instead of "merge" which keeps one's development clean (but one needs to be
> careful with distributing things along the way, as history gets rewritten.)

Hg also has rebase. With clone repos - one has to be careful as to do
this only with local/unshared changeset. Perhaps this works with
in-branch clones aswell [I've never explored]

> 
> Anyway, this probably shouldn't lead to a git vs hg discussion, but I've
> used mercurial as my main tool for various things for quite a while, and
> recently learned more about git, and for me it solves a lot of the annoying
> little things I kept running into with hg. Including having a gui that let's
> you go through changes before commiting them, and separating them, so you
> don't need to do "can't remember what the heck I changed -- I wish mercurial
> had a frontend like bk" commits.

I use 'qct' for this. Perhaps its not as feature rich as it should be
- but its decent. [it supports git and other tools aswell]

> (By the way, even for hg there's "hg diff" and "hg diff | diffstat"
> which helps...) I used to work on the linux kernel with bk, so it's
> probably no surprise that git seems to support the kind of workflow
> I'm used to, though...

One thing hg lacks is the 'bk revtool' - where its easy to browse
history. Yeah one can browse history on the web interface - but its
not as good as the revtool.

Satish