Re: [PATCH v2] build: increase minimum meson version to 0.53.2

Re: [PATCH v2] build: increase minimum meson version to 0.53.2

[David Marchand]

On Fri, Oct 7, 2022 at 6:24 PM Bruce Richardson
<bruce.richardson@intel.com> wrote:
>
> This patchset bumps the minimum meson version from 0.49.2 to 0.53.2.
> Ideally, the minimum version should be 0.53 without a point release, but
> some DPDK builds (mingw) are broken with 0.53.0 due to issue[1], fixed
> by commit[2] in 0.53.1. Therefore we use the latest point release from
> 0.53 branch i.e. 0.53.2.
>
> Some new features of interest which can now be used in DPDK with this
> new minimum meson version:
>
> * can do header-file checks directly inside find_library calls, rather
>   than needing a separate check.[v0.50].
> * can pass multiple cross-files at the same time when cross-compiling
>   [v0.51].
> * "alias_target" function, to allow use to give better/shorter names
>   for particular build objects [v0.52].
> * auto-generation of clang-format [v0.50] and clang-tidy[v0.52] targets
>   when those tools are present and config dotfiles are present.
>   Similarly ctags and cscope are added as targets when those tools are
>   present [v0.53]
> * meson module for filesystem operations, so meson can now check for the
>   presence of particlar files or directories [v0.53].
> * "summary" function to provide a configuration summary at the end of
>   the meson run [v0.53].
>
> Plus many other features. See [3] for full details of each version.
>
> [1] https://github.com/mesonbuild/meson/issues/6442
> [2] https://github.com/mesonbuild/meson/pull/6457/commits/8e7a7c36b579c11dd389fcb1152a4e8ead8b8782

Nit: we can shorten this link and keep only 12 chars for hash.
https://github.com/mesonbuild/meson/pull/6457/commits/8e7a7c36b579

> [3] https://mesonbuild.com/Release-notes.html
>
> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>

Acked-by: David Marchand <david.marchand@redhat.com>


Looking at the test reports for this patch, focusing on compilation:
- it looks like Intel envs are already using a recent enough meson,
- the change passes fine in GHA now,
- UNH envs are reporting failures, but that's expected since the
images must be refreshed once this patch is merged,

So from this aspect, I think we are all set to merge this patch.


It seems the "intel-Testing" test is in ... berserk mode and spewing
errors for all patches since 10/03.
No clue if we have an issue in the main branch, and this test logs are useless.

I'll wait for Intel PRC to be back and ignore for now.


-- 
David Marchand

RE: [PATCH v2] build: increase minimum meson version to 0.53.2

[Tu, Lijuan]

> -----Original Message-----
> From: David Marchand <david.marchand@redhat.com>
> Sent: Saturday, October 8, 2022 3:13 AM
> To: Richardson, Bruce <bruce.richardson@intel.com>; Thomas Monjalon
> <thomas@monjalon.net>
> Cc: dev@dpdk.org; ci@dpdk.org; Lincoln Lavoie <lylavoie@iol.unh.edu>; Tu,
> Lijuan <lijuan.tu@intel.com>; Chen, Zhaoyan <zhaoyan.chen@intel.com>;
> Mcnamara, John <john.mcnamara@intel.com>
> Subject: Re: [PATCH v2] build: increase minimum meson version to 0.53.2
> 
> On Fri, Oct 7, 2022 at 6:24 PM Bruce Richardson <bruce.richardson@intel.com>
> wrote:
> >
> > This patchset bumps the minimum meson version from 0.49.2 to 0.53.2.
> > Ideally, the minimum version should be 0.53 without a point release,
> > but some DPDK builds (mingw) are broken with 0.53.0 due to issue[1],
> > fixed by commit[2] in 0.53.1. Therefore we use the latest point
> > release from
> > 0.53 branch i.e. 0.53.2.
> >
> > Some new features of interest which can now be used in DPDK with this
> > new minimum meson version:
> >
> > * can do header-file checks directly inside find_library calls, rather
> >   than needing a separate check.[v0.50].
> > * can pass multiple cross-files at the same time when cross-compiling
> >   [v0.51].
> > * "alias_target" function, to allow use to give better/shorter names
> >   for particular build objects [v0.52].
> > * auto-generation of clang-format [v0.50] and clang-tidy[v0.52] targets
> >   when those tools are present and config dotfiles are present.
> >   Similarly ctags and cscope are added as targets when those tools are
> >   present [v0.53]
> > * meson module for filesystem operations, so meson can now check for the
> >   presence of particlar files or directories [v0.53].
> > * "summary" function to provide a configuration summary at the end of
> >   the meson run [v0.53].
> >
> > Plus many other features. See [3] for full details of each version.
> >
> > [1] https://github.com/mesonbuild/meson/issues/6442
> > [2]
> >
> https://github.com/mesonbuild/meson/pull/6457/commits/8e7a7c36b579c11d
> > d389fcb1152a4e8ead8b8782
> 
> Nit: we can shorten this link and keep only 12 chars for hash.
> https://github.com/mesonbuild/meson/pull/6457/commits/8e7a7c36b579
> 
> > [3] https://mesonbuild.com/Release-notes.html
> >
> > Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> 
> Acked-by: David Marchand <david.marchand@redhat.com>
> 
> 
> Looking at the test reports for this patch, focusing on compilation:
> - it looks like Intel envs are already using a recent enough meson,
> - the change passes fine in GHA now,
> - UNH envs are reporting failures, but that's expected since the images must be
> refreshed once this patch is merged,
> 
> So from this aspect, I think we are all set to merge this patch.
> 
> 
> It seems the "intel-Testing" test is in ... berserk mode and spewing errors for all
> patches since 10/03.
> No clue if we have an issue in the main branch, and this test logs are useless.
> 
> I'll wait for Intel PRC to be back and ignore for now.

Yes, the main branch had an issue and has been fixed now. Intel-Testing is back to normal. It works fine.

> 
> 
> --
> David Marchand

Re: [PATCH v2] build: increase minimum meson version to 0.53.2

[David Marchand]

On Mon, Oct 10, 2022 at 7:23 AM Tu, Lijuan <lijuan.tu@intel.com> wrote:
> > It seems the "intel-Testing" test is in ... berserk mode and spewing errors for all
> > patches since 10/03.
> > No clue if we have an issue in the main branch, and this test logs are useless.
> >
> > I'll wait for Intel PRC to be back and ignore for now.
>
> Yes, the main branch had an issue and has been fixed now. Intel-Testing is back to normal. It works fine.

The only info we had was that "something" was broken.
Example: http://mails.dpdk.org/archives/test-report/2022-October/313054.html


When other CI report a failure, executed commands and logs are
available in GHA and UNH dashboard, downloadable as zip files.
So it is relatively easy to figure out if the issue was due to an
infrastructure problem or if it was a test failure.


I don't think Intel CI provide such info. If this is the case, please
document it.
Otherwise, please Intel CI team, work on providing the same kind of info.



Thanks.
-- 
David Marchand

Re: [PATCH v2] build: increase minimum meson version to 0.53.2

[David Marchand]

On Fri, Oct 7, 2022 at 9:13 PM David Marchand <david.marchand@redhat.com> wrote:
>
> On Fri, Oct 7, 2022 at 6:24 PM Bruce Richardson
> <bruce.richardson@intel.com> wrote:
> >
> > This patchset bumps the minimum meson version from 0.49.2 to 0.53.2.
> > Ideally, the minimum version should be 0.53 without a point release, but
> > some DPDK builds (mingw) are broken with 0.53.0 due to issue[1], fixed
> > by commit[2] in 0.53.1. Therefore we use the latest point release from
> > 0.53 branch i.e. 0.53.2.
> >
> > Some new features of interest which can now be used in DPDK with this
> > new minimum meson version:
> >
> > * can do header-file checks directly inside find_library calls, rather
> >   than needing a separate check.[v0.50].
> > * can pass multiple cross-files at the same time when cross-compiling
> >   [v0.51].
> > * "alias_target" function, to allow use to give better/shorter names
> >   for particular build objects [v0.52].
> > * auto-generation of clang-format [v0.50] and clang-tidy[v0.52] targets
> >   when those tools are present and config dotfiles are present.
> >   Similarly ctags and cscope are added as targets when those tools are
> >   present [v0.53]
> > * meson module for filesystem operations, so meson can now check for the
> >   presence of particlar files or directories [v0.53].
> > * "summary" function to provide a configuration summary at the end of
> >   the meson run [v0.53].
> >
> > Plus many other features. See [3] for full details of each version.
> >
> > [1] https://github.com/mesonbuild/meson/issues/6442
> > [2] https://github.com/mesonbuild/meson/pull/6457/commits/8e7a7c36b579c11dd389fcb1152a4e8ead8b8782
> > [3] https://mesonbuild.com/Release-notes.html
> >
> > Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> Acked-by: David Marchand <david.marchand@redhat.com>

Applied, thanks.


-- 
David Marchand

RE: [PATCH v2] build: increase minimum meson version to 0.53.2

[Tu, Lijuan]

> -----Original Message-----
> From: David Marchand <david.marchand@redhat.com>
> Sent: Monday, October 10, 2022 3:48 PM
> To: Tu, Lijuan <lijuan.tu@intel.com>
> Cc: Richardson, Bruce <bruce.richardson@intel.com>; Thomas Monjalon
> <thomas@monjalon.net>; dev@dpdk.org; ci@dpdk.org; Lincoln Lavoie
> <lylavoie@iol.unh.edu>; Chen, Zhaoyan <zhaoyan.chen@intel.com>; Mcnamara,
> John <john.mcnamara@intel.com>
> Subject: Re: [PATCH v2] build: increase minimum meson version to 0.53.2
> 
> On Mon, Oct 10, 2022 at 7:23 AM Tu, Lijuan <lijuan.tu@intel.com> wrote:
> > > It seems the "intel-Testing" test is in ... berserk mode and spewing
> > > errors for all patches since 10/03.
> > > No clue if we have an issue in the main branch, and this test logs are useless.
> > >
> > > I'll wait for Intel PRC to be back and ignore for now.
> >
> > Yes, the main branch had an issue and has been fixed now. Intel-Testing is back
> to normal. It works fine.
> 
> The only info we had was that "something" was broken.
> Example: http://mails.dpdk.org/archives/test-report/2022-
> October/313054.html
> 
> 
> When other CI report a failure, executed commands and logs are available in
> GHA and UNH dashboard, downloadable as zip files.
> So it is relatively easy to figure out if the issue was due to an infrastructure
> problem or if it was a test failure.
> 
> 
> I don't think Intel CI provide such info. If this is the case, please document it.
> Otherwise, please Intel CI team, work on providing the same kind of info.

Intel don't provide such info, and I don't think community provides API or services to support it. 
The https://lab.dpdk.org/results/dashboard/ is for UNH, If dashboard could provide a API for intel, we'd like to provide detail logs.
We need community 's help to set up this, I think we can discus it in CI meeting.

> 
> 
> 
> Thanks.
> --
> David Marchand