RE: [PATCH v7 0/6] use xz and gzip on host (outside chroot)

["MOESSBAUER, FELIX JONATHAN" <felix.moessbauer@siemens.com>]

> -----Original Message-----
> From: Uladzimir Bely <ubely@ilbers.de>
> Sent: Wednesday, October 12, 2022 3:41 PM
> To: Moessbauer, Felix Jonathan (T CED INW-CN)
> <felix.moessbauer@siemens.com>; isar-users@googlegroups.com
> Cc: Schild, Henning (T CED SES-DE) <henning.schild@siemens.com>; Kiszka, Jan
> (T CED) <jan.kiszka@siemens.com>; ibr@ilbers.de; Schmidt, Adriaan (T CED SES-
> DE) <adriaan.schmidt@siemens.com>; Anton Mikanovich <amikan@ilbers.de>
> Subject: Re: [PATCH v7 0/6] use xz and gzip on host (outside chroot)
> 
> In the email from Friday, 7 October 2022 21:53:39 +03 user Anton Mikanovich
> wrote:
> > 27.09.2022 12:51, Felix Moessbauer wrote:
> > > Changes since v6:
> > >
> > > - reorder tests in testsuite to not break vm_boot tests
> > > - rebase against current next
> > >
> > > Changes since v5:
> > >
> > > - before changing WIC deploy logic, revert edc10d93 (p3)
> > > - only reploy requested files (p4)
> > >
> > > Changes since v4:
> > >
> > > - add option WIC_DEPLOY_PARTITIONS = "0" to control if individual
> > > wic partitions are deployed (p3)>
> > >    Note, that the default set to disabled is important.
> > >    This fixes the huge disk consumption when running the CI with WIC
> > > tests
> > >
> > > - add test for WIC_DEPLOY_PARTITIONS logic (p4)
> > > - add logic to handle cross-imaging when distro_arch == host_arch
> > > (similar to buildchroot-(target|host)) (p1)
> > >
> > > Please note, that the additions around WIC_DEPLOY_PARTITIONS are
> > > standalone patches which are only sent as part of this series to fix
> > > the disk space blow up in CI.
> > >
> > > Changes since v3:
> > >
> > > - add test for CROSS=1 wic image including compression
> > > - integrate patch 2 into this series instead of standalone.
> > >
> > > Please note, this series supersedes the "add wic.xz image to default
> > > qemuarm64 target" series but does only requires little more disk
> > > space for the CI tests (compared to ISAR-next). By that, the CI can
> > > still be executed on standard machines.
> > >
> > > Changes since v1/v2:
> > >
> > > instead of switching the compression logic, this patch adds support
> > > to run the whole imager in the buildchroot-host.
> > >
> > > Best regards,
> > > Felix
> > >
> > > Felix Moessbauer (6):
> > >    do not crossbuild SDK for host-arch eq dist-arch
> > >    run imager in buildchroot-host on cross
> > >    Revert "wic: move out all files ending on "direct*""
> > >    add option to control deploy of raw wic partitions
> > >    add test for wic partition deploy logic
> > >    add wic.xz image to qemuarm64-bookworm
> > >
> > >   RECIPE-API-CHANGELOG.md                       |  9 +++++++
> > >   doc/user_manual.md                            | 11 +++++++++
> > >   meta-isar/conf/local.conf.sample              |  3 +++
> > >   .../conf/multiconfig/qemuarm64-bookworm.conf  |  3 +++
> > >   meta/classes/image-tools-extension.bbclass    |  3 ---
> > >   meta/classes/imagetypes_wic.bbclass           | 16 +++++++++----
> > >   meta/classes/sdk.bbclass                      | 14 +++++++++--
> > >   testsuite/cibase.py                           | 10 ++++++++
> > >   testsuite/cibuilder.py                        |  5 +++-
> > >   testsuite/citest.py                           | 24 +++++++++++++++++++
> > >   10 files changed, 88 insertions(+), 10 deletions(-)
> >
> > Applied to next, thanks.
> 
> Now, with "[PATCH v7 5/6] add test for wic partition deploy logic" we lost all
> "nocross" CI logs from "tmp/work" (the test deletes "tmp" directory).
> 
> While our tests are still sometimes depend on each other, the easiest fix would
> be placing "wic" tests before "nocross" ones.

Hi Uladzimir,

Could you please try to implement this minimal patch.
I have a really hard time to run the full testsuite behind our proxies.

The whole CI currently has a lot of inter-test dependencies which make it really hard to add tests for new features.
Ideally, the test order should not matter at all.
In case the build-time is an issue, we should better rely on the sstate cache than on the artifacts of previous tests.

Felix


> 
> --
> Uladzimir Bely
> 
>