Re: [Buildroot] [PATCH 1/2] package/ipmitool: backport fixes for registry issues

["Yann E. MORIN" <yann.morin.1998@free.fr>]

Thomas, All,

On 2023-01-14 22:58 +0100, Thomas Petazzoni via buildroot spake thusly:
> On Sat, 14 Jan 2023 15:33:06 -0600
> Vincent Fazio <vfazio@gmail.com> wrote:
> > replies below (i hope, i haven't tried responding to ML emails in gmail)
> It's almost good, except that gmail adds one extra > on the first line
> of each paragraph of your replies.

Still, I haven't seen Vincent's reply on the list...

> > > So in the current Buildroot situation (before this patch), the registry
> > > is always installed, correct?
> > >
> > > Correct, but with a caveat... as of the version bump to 1.8.19, the file  
> > being installed is broken and the tool actually won't work.
> 
> So the current ipmitool in Buildroot is broken. Could perhaps separate
> a "fix" that could be backported from a large improvement that allows
> to not used the registry?

It was I who asked Vincent to add all the patches from upstream with a
single commit in Buildroot.

Indeed, my reasonning was that they all pertain to a single issue: the
package is broken because 1) the registry URL is now incorrect (it
downloads an HTML blob instead of a plain-text description) and 2) the
build is not reproducible (as the registry can change on a whim).

Furthermore, even though the registry is currently installed, it is of
no use because the file is not in the format expected by the tools, so
I think it is accetable that we make installing the registry optional,
and that the option does not default to y (although I would not mind
if it did).

> > All upstream patches have been accepted into mainline. The commit hashes  
> > in the patches are the mainline commits
> 
> Thing is we don't which are "upstream patches". If the patches are
> upstream, you can do:
> 
> Signed-off-by: John Doe <john.doe@foobar.com>
> Upstream: https://github.com/some/where/commit/12456abc

I usually am content with just a git-formatted patch, with the actual
history in proper reasing order:

    From <HASH> <date>
    From: author....
    Date: date...
    Subject: subject...

    Commit log...

    upsteam-sob...
    [john.doe@foobar.com: backport from upstream]
    Signed-off-by: John Doe <john.doe@foobar.com>
    ---
    diff...

thus we can indeed check the commit hash at the beginning of the patch
and check whether it is part of a new upstream release.

If the patch had to be tweaked, then;

    From <LOCAL_HASH> <date>
    From: author....
    Date: date...
    Subject: subject...

    Commit log...

    upsteam-sob...
    [john.doe@foobar.com: backport from upstream HASH]
    Signed-off-by: John Doe <john.doe@foobar.com>
    ---
    diff...

> This makes it very clear that it has been accepted upstream, and will
> help someone who will do the next ipmitool version bump.

But yes, as long as the information is there, that's still OK.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot