All of lore.kernel.org
 help / color / mirror / Atom feed
From: Gary Thomas <gary@mlbassoc.com>
To: Otavio Salvador <otavio@ossystems.com.br>
Cc: "meta-freescale@yoctoproject.org" <meta-freescale@yoctoproject.org>
Subject: Re: [meta-fsl-arm][PATCH] u-boot-fslc: Add tag to git SRC_URI
Date: Mon, 09 Dec 2013 10:31:48 -0700	[thread overview]
Message-ID: <52A5FE84.8000203@mlbassoc.com> (raw)
In-Reply-To: <CAP9ODKpbgt=-+YNmtSEA+BiHFdCnRcd+Squacs6adfJcTs=yZA@mail.gmail.com>

On 2013-12-09 10:10, Otavio Salvador wrote:
> On Mon, Dec 9, 2013 at 3:05 PM, Gary Thomas <gary@mlbassoc.com> wrote:
>> On 2013-12-09 10:00, Otavio Salvador wrote:
>>>
>>> On Mon, Dec 9, 2013 at 2:59 PM, Gary Thomas <gary@mlbassoc.com> wrote:
>>>>
>>>> On 2013-12-09 09:55, Otavio Salvador wrote:
>>>>>
>>>>>
>>>>> On Mon, Dec 9, 2013 at 2:53 PM, Gary Thomas <gary@mlbassoc.com> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>> Do you have an use example where you'd need it?
>>>>>>>>
>>>>>>> I think he means to do something like this:
>>>>>>>
>>>>>>> GITTAG ??= "patches-2013.10"
>>>>>>> SRC_URI = "git://github.com/Freescale/u-boot-imx.git;tag=${GITTAG}"
>>>>>>>
>>>>>>> This way he can override GITTAG in his .bbappend, correct?
>>>>>>
>>>>>>
>>>>>>
>>>>>> Correct.  The ??= isn't even necessary since the .bbappend can
>>>>>> always override it.
>>>>>
>>>>>
>>>>>
>>>>> In this case you're using an old version of the bootloader in your
>>>>> internal BSP, right? I'd expect you to add an .bb file for this
>>>>> version instead and use PREFERRED_VERSION to use it.
>>>>
>>>>
>>>>
>>>> Why should I do that when .bbappend works perfectly well?  I can also
>>>> see this as a case when new boards are added or old ones are still
>>>> around and they get updated on different schedules.
>>>>
>>>> Also, what does it hurt to be flexible?
>>>
>>>
>>> In this case you'd have a 2013.10 recipe building/installing a 2013.04
>>> version for example, this is misleading and confusing for someone
>>> using the BSP.
>>>
>>
>> You do what you want - I just hope that I don't end up having
>> to clone all of meta-fsl-arm* just because you fear a little
>> flexibility :-(
>
> I think you got me wrong.
>
> The discussion here is to try to identify needs and try to come with
> clean solution which works most people while keep it clean and simple
> to understand as possible.
>
> I think it is quite confusing when someone seems a bbappend which
> changes a recipe for an old version and this does not match the recipe
> name. I am fully aware this is /possible/ to do but this does not mean
> this is advisable or desired.
>
> I think that John's proposed solution is going to address both concerns well.
>

Here's what I want to be able to support: I'm building my
own targets (actual hardware) and I start from some version
of u-boot-fslc, linux-boundary, etc.  My local working copies
of that are based on a particular version and I want to still
use the same recipes from meta-fsl-arm* even if they are updated
for more recent versions.  This is even more important for the
kernel.  I have boards that are based on boundary-imx_3.0.35_4.0.0
but the meta-fsl-arm* recipes are now based on boundary-imx_3.0.35_4.1.0.
Note that the recipe linux-boundary_3.0.35.bb didn't change names
or even version numbers when the branch was changed. Without the ability
to override the branch in my .bbappend, I can't use the meta-fsl-arm*
recipe at all and would end up having to clone it which is really
bad/unacceptable.

I would not be using this in a way that the version being built was
different than the recipe version implied.  If, for example, meta-fsl-arm*
replaces u-boot-fslc_2013.10.bb with u-boot-fslc_2013.12.bb (and dropped
the older u-boot-fslc_2013.10.bb), just like  you I would not find it
acceptable to use .bbappend to go back to the 2013.10 branch/version.
In that case, I would clone the 2013.10 recipe into my own layer(s).
I just try to use the main layers as much as possible.

-- 
----------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------


  reply	other threads:[~2013-12-09 17:31 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-12-09 16:10 [meta-fsl-arm][PATCH] u-boot-fslc: Add tag to git SRC_URI John Weber
2013-12-09 16:26 ` Gary Thomas
2013-12-09 16:28   ` John Weber
2013-12-09 16:35     ` Gary Thomas
2013-12-09 16:35       ` Otavio Salvador
2013-12-09 16:45         ` John Weber
2013-12-09 16:53           ` Gary Thomas
2013-12-09 16:55             ` Otavio Salvador
2013-12-09 16:59               ` Gary Thomas
2013-12-09 17:00                 ` Otavio Salvador
2013-12-09 17:03                   ` John Weber
2013-12-09 17:05                     ` Otavio Salvador
2013-12-09 17:05                   ` Gary Thomas
2013-12-09 17:10                     ` Otavio Salvador
2013-12-09 17:31                       ` Gary Thomas [this message]
2013-12-09 17:58                         ` Otavio Salvador
2013-12-09 16:38       ` John Weber

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=52A5FE84.8000203@mlbassoc.com \
    --to=gary@mlbassoc.com \
    --cc=meta-freescale@yoctoproject.org \
    --cc=otavio@ossystems.com.br \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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.