unofficial mirror of guix-devel@gnu.org 
 help / color / mirror / code / Atom feed
* Ungrafting
@ 2016-05-01 20:20 Leo Famulari
  2016-05-02 13:34 ` Ungrafting Mark H Weaver
  0 siblings, 1 reply; 4+ messages in thread
From: Leo Famulari @ 2016-05-01 20:20 UTC (permalink / raw)
  To: guix-devel

Hi all,

When committing a bug fix with a graft, I think it would be a good idea
to follow up on some other branch with a commit that makes the same
change without a graft.

Core-updates was suggested on IRC. This would mean that after each graft
commit, master would need to be merged into core-updates, and then the
"ungrafting" patch could be applied.

Thoughts?

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Ungrafting
  2016-05-01 20:20 Ungrafting Leo Famulari
@ 2016-05-02 13:34 ` Mark H Weaver
  2016-05-02 15:54   ` Ungrafting Ludovic Courtès
  2016-05-02 18:39   ` Ungrafting Leo Famulari
  0 siblings, 2 replies; 4+ messages in thread
From: Mark H Weaver @ 2016-05-02 13:34 UTC (permalink / raw)
  To: Leo Famulari; +Cc: guix-devel

Leo Famulari <leo@famulari.name> writes:

> When committing a bug fix with a graft, I think it would be a good idea
> to follow up on some other branch with a commit that makes the same
> change without a graft.
>
> Core-updates was suggested on IRC. This would mean that after each graft
> commit, master would need to be merged into core-updates, and then the
> "ungrafting" patch could be applied.

Merging those two will be awkward.  In my experience, the result of git
automatically merging these two commits is to update the main package
*and* to graft it.  For this reason, I think it's preferable for the
ungrafted commit to be on top of the grafted one, i.e. it should remove
the graft and update the origin package in a single commit.

In practice, this means that after applying the graft to master, master
should be merged into core-updates before applying the ungrafting commit
to core-updates.

What do you think?

      Mark

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Ungrafting
  2016-05-02 13:34 ` Ungrafting Mark H Weaver
@ 2016-05-02 15:54   ` Ludovic Courtès
  2016-05-02 18:39   ` Ungrafting Leo Famulari
  1 sibling, 0 replies; 4+ messages in thread
From: Ludovic Courtès @ 2016-05-02 15:54 UTC (permalink / raw)
  To: Mark H Weaver; +Cc: guix-devel

Mark H Weaver <mhw@netris.org> skribis:

> Leo Famulari <leo@famulari.name> writes:
>
>> When committing a bug fix with a graft, I think it would be a good idea
>> to follow up on some other branch with a commit that makes the same
>> change without a graft.
>>
>> Core-updates was suggested on IRC. This would mean that after each graft
>> commit, master would need to be merged into core-updates, and then the
>> "ungrafting" patch could be applied.
>
> Merging those two will be awkward.  In my experience, the result of git
> automatically merging these two commits is to update the main package
> *and* to graft it.

Good point.

> For this reason, I think it's preferable for the ungrafted commit to
> be on top of the grafted one, i.e. it should remove the graft and
> update the origin package in a single commit.
>
> In practice, this means that after applying the graft to master, master
> should be merged into core-updates before applying the ungrafting commit
> to core-updates.
>
> What do you think?

Sounds like a good workflow, yes.

Regarding builds, it’s really a scheduling problem.  We have periodic
update branches (gnome-updates, python-updates, core-updates), and
Something could tell you whether a change is eligible for one of these
branches, so we could batch related changes together.

Ludo’.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Ungrafting
  2016-05-02 13:34 ` Ungrafting Mark H Weaver
  2016-05-02 15:54   ` Ungrafting Ludovic Courtès
@ 2016-05-02 18:39   ` Leo Famulari
  1 sibling, 0 replies; 4+ messages in thread
From: Leo Famulari @ 2016-05-02 18:39 UTC (permalink / raw)
  To: Mark H Weaver; +Cc: guix-devel

On Mon, May 02, 2016 at 09:34:34AM -0400, Mark H Weaver wrote:
> Leo Famulari <leo@famulari.name> writes:
> > Core-updates was suggested on IRC. This would mean that after each graft
> > commit, master would need to be merged into core-updates, and then the
> > "ungrafting" patch could be applied.

> In practice, this means that after applying the graft to master, master
> should be merged into core-updates before applying the ungrafting commit
> to core-updates.
> 
> What do you think?

I'm having a hard time seeing the difference between our suggestions.
Can you clarify yours?

To give more detail, here is what I was suggesting:

1) Apply a graft to master.
2) Merge master into core-updates.
3) Apply a commit to core-updates that "un-does" the graft and applies
the bug fix without a graft.

Apologies if my original email was unclear.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2016-05-02 18:40 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-05-01 20:20 Ungrafting Leo Famulari
2016-05-02 13:34 ` Ungrafting Mark H Weaver
2016-05-02 15:54   ` Ungrafting Ludovic Courtès
2016-05-02 18:39   ` Ungrafting Leo Famulari

Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/guix.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).