public inbox for isar-users@googlegroups.com
 help / color / mirror / Atom feed
From: Jan Kiszka <jan.kiszka@web.de>
To: Claudius Heine <claudius.heine.ext@siemens.com>,
	isar-users@googlegroups.com
Cc: Claudius Heine <ch@denx.de>
Subject: Re: [PATCH v4 1/1] meta: added do_transform_template task as templating system and switch
Date: Mon, 25 Mar 2019 12:08:45 +0100	[thread overview]
Message-ID: <10421a3e-15cf-df0a-12a7-485db9ee10db@web.de> (raw)
In-Reply-To: <57786819-c53a-98ed-91f3-0c9adf2ddb7f@web.de>

On 25.03.19 09:44, Jan Kiszka wrote:
> On 25.03.19 09:21, Claudius Heine wrote:
>> Hi,
>>
>> On 25/03/2019 09.13, Jan Kiszka wrote:
>>> On 25.03.19 08:53, Claudius Heine wrote:
>>>> Hi Jan,
>>>>
>>>> On Sun, 2019-03-24 at 19:28 +0100, Jan Kiszka wrote:
>>>>> On 14.02.19 12:28, [ext] claudius.heine.ext@siemens.com wrote:
>>>>>> From: Claudius Heine <ch@denx.de>
>>>>>>
>>>>>> The new introduced variables are TEMPLATE_FILES and TEMPLATE_VARS.
>>>>>> TEMPLATE_FILES is a list of files that should be processed in the
>>>>>> `do_transform_template` task and TEMPLATE_VARS is a list of bitbake
>>>>>> variable names that should be substituted in the template file.
>>>>>>
>>>>>
>>>>> Bad news: This patch introduced a regression into the release (sorry,
>>>>> didn't
>>>>> manage to test earlier). It destroys the reuse of self-built
>>>>> identical packages
>>>>> for different targets. Namely, jailhouse-images now tries to build
>>>>> the shared
>>>>> (per arch) kernel multiple times, and that fails miserably because
>>>>> there is a
>>>>> common folder for all of them. The issue disappears when removing the
>>>>> do_transform_template task from the chain.
>>>>
>>>> Hmm, have you looked at the resulting 'TEMPLATE_VARS' variable? All
>>>> variables in there need to be the same in order for it to have the same
>>>> stamp. If that is not the case then maybe overwriting it instead of
>>>> appending it might work.
>>>>
>>>> These are the defaults set in 'template.bbclass':
>>>>
>>>>      TEMPLATE_VARS ?= "PN PV DESCRIPTION HOMEPAGE MAINTAINER KERNEL_NAME
>>>>                        MACHINE DISTRO_ARCH"
>>>>
>>>> If I understand you correctly then your recipe should not be dependant
>>>> on the MACHINE? So either remove MACHINE 'TEMPLATE_VARS_remove =
>>>> "MACHINE"' or just set a 'TEMPLATE_VARS' that you really need in your
>>>> template files.
>>>
>>> OK, might be that "MACHINE" is the problem here. Will check. If that turns out
>>> to be true, it is highly counter-intuitive and should be fixed, at least for
>>> nop-jobs.
>>
>> Well sane defaults is always WIP :)
>>
>> I think I added this because the u-boot recipe in meta needed it. It seems
>> that u-boot is special here.
>
> OK, clearing TEMPLATE_VARS seems to address the issue (still seeing others, but
> that might be a local topic).
>
> Given these side effects, I'm in favor of removing MACHINE from the defaults.
> The other vars look harmless, at least so far.
>

Actually, we have to: Even standard packages like isar-cfg-localepurge start to
fail building when there are multiple targets of the same arch pulling them in.
They share the same workdir, thus break when trying to be built multiple times.

Will send a patch.

Jan

  reply	other threads:[~2019-03-25 11:08 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-14 11:28 [PATCH v4 0/1] Template system claudius.heine.ext
2019-02-14 11:28 ` [PATCH v4 1/1] meta: added do_transform_template task as templating system and switch claudius.heine.ext
2019-03-24 18:28   ` Jan Kiszka
2019-03-25  7:53     ` Claudius Heine
2019-03-25  8:13       ` Jan Kiszka
2019-03-25  8:21         ` Claudius Heine
2019-03-25  8:44           ` Jan Kiszka
2019-03-25 11:08             ` Jan Kiszka [this message]
2019-02-15 14:45 ` [PATCH v4 0/1] Template system Maxim Yu. Osipov

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=10421a3e-15cf-df0a-12a7-485db9ee10db@web.de \
    --to=jan.kiszka@web.de \
    --cc=ch@denx.de \
    --cc=claudius.heine.ext@siemens.com \
    --cc=isar-users@googlegroups.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox