public inbox for isar-users@googlegroups.com
 help / color / mirror / Atom feed
From: "Schmidt, Adriaan" <adriaan.schmidt@siemens.com>
To: Anton Mikanovich <amikan@ilbers.de>,
	"isar-users@googlegroups.com" <isar-users@googlegroups.com>
Cc: Baurzhan Ismagulov <ibr@ilbers.de>,
	"henning.schild@siemens.com" <henning.schild@siemens.com>
Subject: RE: [PATCH v6 13/13] sstate: add test case
Date: Fri, 19 Nov 2021 10:33:39 +0000	[thread overview]
Message-ID: <AS4PR10MB5318D0D7C63DF6A05B341160ED9C9@AS4PR10MB5318.EURPRD10.PROD.OUTLOOK.COM> (raw)
In-Reply-To: <af2063a6-b9b0-8af3-0a03-b62da4231596@ilbers.de>

Anton Mikanovich, 17. November 2021 17:52:
> 17.11.2021 11:21, Schmidt, Adriaan wrote:
> > Anton Mikanovich, 16. November 2021 17:47:
> >> 11.11.2021 09:47, Adriaan Schmidt wrote:
> >>> This also adds a parameter `sstate` to the build function, which is
> >>> only passed as `1` for the actual sstate test. None of the other tests
> >>> uses sstate caching.
> >>>
> >> Thanks for adding test case, but if you need to add it to fast/full in
> >> should be moved to the beginning of the build_test.py (after repro
> >> maybe) to not cleanup images needed for start_vm testing. Executing it
> >> at the end will also remove all the logs of common tests.
> >> Maybe we can not remove tmp but just set any empty SSTATE_DIR?
> > The whole testing of the sstate-cache is about building with a "warm" cache
> (populated SSTATE_DIR), but without any previous artifacts in tmp. Also,
> having logs from previous builds will break the check for executed tasks,
> which reads log.task_order files.
> >
> > One solution I see is to use a dedicated build dir for that test. This
> could also be a general option, to make it explicit which test cases depend
> on each other (use the same build dir) and which are independent (have a
> separate build dir for each test). That would mean to have the build dir not
> as an Avocado parameter, but have each test case set it individually.
> >
> >> Changing testsuite API should probably be moved to the separate patch.
> > If we're changing API, I would also propose to use keyword arguments
> instead of positional ones, and setting of defaults. This would make things
> much more readable, e.g., turning:
> > build_dir, bb_args = self.prep('Sstate', '', 0, 1, 1)
> > self.perform_build_test(image_target, 0, 1, None)
> > into:
> > build_dir, bb_args = self.prep('Sstate', sstate=True, debsrc_cache=True)
> > self.perform_build_test(image_target, sstate=True)
> >
> >
> > I'm happy to propose something, either before or (which would probably make
> Henning happy) after merging the sstate feature.
> >
> > Adriaan
> >
> Using separate build dirs was already in TODO for the next avocado
> improvements as requirement for test cases parallelization.
> If talking about testing API extension we can try to pass dict of
> key-value params as an argument. It will allow us to add any custom
> params later.

I have something that addresses parameter-passing and almost runs.
Will post a refactoring proposal soon.

> Everything besides test case looks good, so we can merge the first 12
> patches separately and move testing into separate thread.

I agree. But only merge patches 1..10, and leave everything test-related for another thread.

Adriaan


  reply	other threads:[~2021-11-19 10:33 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-11  6:47 [PATCH v6 00/13] Add sstate-cache Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 01/13] oe imports in central location Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 02/13] images: create deploy dir Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 03/13] rootfs: recursively depend on packages Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 04/13] base: remove unneeded "before do_build" task dependencies Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 05/13] dpkg: add explicit dependency to isar-apt Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 06/13] meta: add sstate feature from oe Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 07/13] sstate: configure Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 08/13] sstate: add caching to isar-bootstrap Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 09/13] sstate: add caching to rootfs Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 10/13] sstate: add caching to debian packages Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 11/13] test: pass absolute path for build_dir Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 12/13] test: make bitbake_args a list Adriaan Schmidt
2021-11-11  6:47 ` [PATCH v6 13/13] sstate: add test case Adriaan Schmidt
2021-11-16 16:46   ` Anton Mikanovich
2021-11-16 22:47     ` Henning Schild
2021-11-17  8:21     ` Schmidt, Adriaan
2021-11-17 16:51       ` Anton Mikanovich
2021-11-19 10:33         ` Schmidt, Adriaan [this message]
2021-12-02  9:19 ` [PATCH v6 00/13] Add sstate-cache Anton Mikanovich

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=AS4PR10MB5318D0D7C63DF6A05B341160ED9C9@AS4PR10MB5318.EURPRD10.PROD.OUTLOOK.COM \
    --to=adriaan.schmidt@siemens.com \
    --cc=amikan@ilbers.de \
    --cc=henning.schild@siemens.com \
    --cc=ibr@ilbers.de \
    --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