All of lore.kernel.org
 help / color / mirror / Atom feed
* bitbake-setup feedback - long subpath + hardcoding machine and distro
@ 2025-11-12  2:53 ChenQi
  2025-11-12  6:09 ` Alexander Kanavin
  0 siblings, 1 reply; 3+ messages in thread
From: ChenQi @ 2025-11-12  2:53 UTC (permalink / raw)
  To: openembedded-core@lists.openembedded.org, Alexander Kanavin
  Cc: Richard Purdie

Hi All,

I'd like to provide one feedback on bitbake-setup.

The subpath is unnecessarily long, e.g., 
poky-master-poky-distro_poky-machine_qemux86-64. And it's hardcoding 
machine and distro values.

Such design might be reasonable for testing infrastructure, but it's not 
good user experience for developers.

For example, what if I want to to build things for qemuarm64? I'm using 
layers under the a subpath with the name of 
poky-master-poky-distro_poky-machine_qemux86-64 but I'm building for 
qemuarm64. This feels a little weird.

Regards,
Qi



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: bitbake-setup feedback - long subpath + hardcoding machine and distro
  2025-11-12  2:53 bitbake-setup feedback - long subpath + hardcoding machine and distro ChenQi
@ 2025-11-12  6:09 ` Alexander Kanavin
  2025-11-12  6:28   ` ChenQi
  0 siblings, 1 reply; 3+ messages in thread
From: Alexander Kanavin @ 2025-11-12  6:09 UTC (permalink / raw)
  To: ChenQi; +Cc: openembedded-core@lists.openembedded.org, Richard Purdie

On Wed, 12 Nov 2025 at 03:54, ChenQi <Qi.Chen@windriver.com> wrote:
>
> The subpath is unnecessarily long, e.g.,
> poky-master-poky-distro_poky-machine_qemux86-64. And it's hardcoding
> machine and distro values.
> Such design might be reasonable for testing infrastructure, but it's not
> good user experience for developers.
>
> For example, what if I want to to build things for qemuarm64? I'm using
> layers under the a subpath with the name of
> poky-master-poky-distro_poky-machine_qemux86-64 but I'm building for
> qemuarm64. This feels a little weird.

The directory name is hardcoding the choices made during
'bitbake-setup init'. The reason is that if you want to run 'init'
again and make different choices, it would be able to create a
different directory, and not error out because the one it wants to
create already exists.

There's a patch for bitbake-setup to give you a chance to write in
something shorter interactively. You can also pass in a shorter choice
with --setup-dir option. You can also use --skip-selection to omit the
selection of distro or machine, or anything else that is asked at
'init', and then you will also get a shorter directory. But the
default needs to ensure different setup directories don't clash, and
so the choices are hardcoded in them. If there's something else we
could do, please tell so.

Alex


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: bitbake-setup feedback - long subpath + hardcoding machine and distro
  2025-11-12  6:09 ` Alexander Kanavin
@ 2025-11-12  6:28   ` ChenQi
  0 siblings, 0 replies; 3+ messages in thread
From: ChenQi @ 2025-11-12  6:28 UTC (permalink / raw)
  To: Alexander Kanavin
  Cc: openembedded-core@lists.openembedded.org, Richard Purdie

On 11/12/25 14:09, Alexander Kanavin wrote:
> On Wed, 12 Nov 2025 at 03:54, ChenQi <Qi.Chen@windriver.com> wrote:
>> The subpath is unnecessarily long, e.g.,
>> poky-master-poky-distro_poky-machine_qemux86-64. And it's hardcoding
>> machine and distro values.
>> Such design might be reasonable for testing infrastructure, but it's not
>> good user experience for developers.
>>
>> For example, what if I want to to build things for qemuarm64? I'm using
>> layers under the a subpath with the name of
>> poky-master-poky-distro_poky-machine_qemux86-64 but I'm building for
>> qemuarm64. This feels a little weird.
> The directory name is hardcoding the choices made during
> 'bitbake-setup init'. The reason is that if you want to run 'init'
> again and make different choices, it would be able to create a
> different directory, and not error out because the one it wants to
> create already exists.
>
> There's a patch for bitbake-setup to give you a chance to write in
> something shorter interactively. You can also pass in a shorter choice
> with --setup-dir option. You can also use --skip-selection to omit the
> selection of distro or machine, or anything else that is asked at
> 'init', and then you will also get a shorter directory. But the
> default needs to ensure different setup directories don't clash, and
> so the choices are hardcoded in them. If there's something else we
> could do, please tell so.
>
> Alex

Hi Alex,

Thanks a lot for your help. The '--setup-dir-name' satisfies what I need.

Regards,
Qi



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2025-11-12  6:28 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-11-12  2:53 bitbake-setup feedback - long subpath + hardcoding machine and distro ChenQi
2025-11-12  6:09 ` Alexander Kanavin
2025-11-12  6:28   ` ChenQi

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.