I believe that it is a feature. It is most often used to
have the same change on two different branches of the same
project,
-Martin
--
Employee of Qualcomm Innovation Center, Inc. which is a
member of Code Aurora Forum
I believe that it is a feature. It is most often used toOn Sunday, January 15, 2012 01:21:33 am Matthias Sohn wrote:
> We've got 2 changes for 2 different repositories hosted
> on the same gerrit server having the same change-id
> http://egit.eclipse.org/r/#q,I74a82b31505fadf8378069c5f65
> 54f1033c28f9b,n,z
>
> Is this a bug or a feature ? If it's a feature the
> change-id can't be used anymore to un-ambiguously link
> to a change.
have the same change on two different branches of the same
project,
> 2012/1/16 Martin Fick <mf...@codeaurora.org>
>
> > I believe that it is a feature. It is most often used to
> > have the same change on two different branches of the same
> > project,
>
> ok, this is useful, but what is then the way of unambiguously
> hyper-linking to a single change ?
The change number or the commit SHA-1?
--
Magnus Bäck Opinions are my own and do not necessarily
SW Configuration Manager represent the ones of my employer, etc.
Sony Ericsson
Change number.
The commit SHA1 can be the same on multiple changes also, we
just had that problem internally, for the same branch even!
Imagine two projects with the same file in them with the same bug. Just like if they were on 2 branches in the same project, it makes sense to apply the same fix (change) to both projects,
-Martin
Employee of Qualcomm Innovation Center,Inc. which is a member of Code Aurora Forum
Pravin <pra...@nvidia.com> wrote:
>
>I am still not clear on this. Except the commit message everything
>seems
>to be different. e.g.The two changes affect totally different set of
>files.
>Can the change-Id still be the same?
Since users can assign any changeid to any commit that they want (unlike a git sha1, there is nothing which enforces it to be derived from the code, even if it is a good idea to make it be so originally). But the intent of a changeid is to represent patchsets related to a common logical idea. While it may not be common, I believe that a reasonable case could he made under certain circumstances that the same logical idea applies to multiple projects, it is these cases which I feel justifies allowing the same changeid on different projects.
-Martin
Note also that nothing prevents the same git commit (sha1) from being used in multiple projects, and I don't mean because of collisions.
Since users can assign any changeid to any commit that they want (unlike a git sha1, there is nothing which enforces it to be derived from the code, even if it is a good idea to make it be so originally). But the intent of a changeid is to represent patchsets related to a common logical idea. While it may not be common, I believe that a reasonable case could he made under certain circumstances that the same logical idea applies to multiple projects, it is these cases which I feel justifies allowing the same changeid on different projects.
Pravin <pra...@nvidia.com> wrote:
>
>I am still not clear on this. Except the commit message everything
>seems
>to be different. e.g.The two changes affect totally different set of
>files.
>Can the change-Id still be the same?
-Martin
Note also that nothing prevents the same git commit (sha1) from being used in multiple projects, and I don't mean because of collisions.
> 2012/1/17 Martin Fick <mf...@codeaurora.org>
[...]
> > Note also that nothing prevents the same git commit (sha1) from
> > being used in multiple projects, and I don't mean because of
> > collisions.
>
> yes, though it's very unlikely since it's a hash over the content and
> history
You're talking about a collision, but I believe what Martin referred to
was two projects that share the same history and where you in fact can
reuse the exact same commit. Example:
git push ssh://hostname/project1 HEAD:refs/for/master
git push ssh://hostname/project2 HEAD:refs/for/master
"Magnus Bäck" <magnu...@sonyericsson.com> wrote:
>On Tuesday, January 17, 2012 at 08:13 CET,
> Matthias Sohn <matthi...@googlemail.com> wrote:
>
>> 2012/1/17 Martin Fick <mf...@codeaurora.org>
>
>[...]
>
>> > Note also that nothing prevents the same git commit (sha1) from
>> > being used in multiple projects, and I don't mean because of
>> > collisions.
>>
>> yes, though it's very unlikely since it's a hash over the content and
>> history
>
>You're talking about a collision, but I believe what Martin referred to
>was two projects that share the same history and where you in fact can
>reuse the exact same commit. Example:
>
>git push ssh://hostname/project1 HEAD:refs/for/master
>git push ssh://hostname/project2 HEAD:refs/for/master
Yes, thank-you.