Re: [Gnu-arch-users] tla-pqm 0.2

[Colin Walters]

On Fri, 2003-10-17 at 13:56, Joshua Haberman wrote:

> If you are developing on your own branch (one not specifically meant for
> a particular experimental change), aren't you going to want to
> star-merge from the canonical branch frequently anyway?  

Definitely.

> You don't want
> to be developing on a tree that is missing longstanding changes from
> other contributors, do you?  As I see it, doing a star-merge on a
> development branch is not part of the process of submitting one or more
> changesets, it's a periodic operation to perform on personal branches.

Er...we're not talking about automating merges into the personal branch,
we're talking about automating them *from* the personal branch.

In other words, I meant that you do the star-merge into the central
branch on the client, and then send the resulting changeset to the
server.  This is instead of just sending a merge request to the server,
which would do the star-merge itself.

$ tla get address@hidden/hello-world--mainline--1.0 hello-world-example
$ tla get address@hidden/hello-world--mainline--1.0 hello-world-mine
$ cd hello-world-example
$ tla star-merge -o ,merge address@hidden/hello-world--mainline--1.0
$ tar czvf ,merge.tar.gz
$ echo 'changeset' | gpg --clearsign | mail -s 'fix up some bugs' 
--attachment=,merge.tar.gz address@hidden

All of the above could be implemented easily as a shell script.

> Also think of contributors who don't have permission to write to the
> canonical branch.  They should be able to submit merge requests or
> changesets and have the people who do have write access approve or deny
> them.

Yes...I think the way to implement this though would be to have the
people who can approve changesets run their own tla-pqm instances.  Then
instead of submitting merge requests to example-dev, the sub-satellite
developers submit them to the main developers.

signature.asc
Description: This is a digitally signed message part