Re: [PATCH v2 0/7] expand list of optional libraries

[David Marchand <david.marchand@redhat.com>]

On Fri, Jun 23, 2023 at 12:25 PM David Marchand
<david.marchand@redhat.com> wrote:
>
> On Thu, Jun 22, 2023 at 5:13 PM Bruce Richardson
> <bruce.richardson@intel.com> wrote:
> >
> > DPDK still has many libraries which cannot be disabled as part of a
> > build. With the ongoing work to make it easier to only build a subset
> > of the libraries in DPDK, we can also work to expand the list of
> > libraries which can be enabled/disabled as desired.
> >
> > This patch addresses a number of the "low-hanging fruit" libraries,
> > where only the unit test builds need minor changes to support
> > making the library optional. The rest of the build system is already
> > well set up for selective disabling of libraries.
> >
> > For better support of enabling components, especially those more
> > integrated into DPDK unit tests, rework of the test meson.build file
> > is likely needed. For example, it could probably be better rewritten
> > to use a dictionary of files and the dependencies of each file, and
> > the unit test commands each provides. However, such rework is a
> > significant effort, and outside the scope of this patchset.
> >
> > V2: fix checkpatch issues, since checkpatch doesn't like empty commit
> >     messages (even if the title is pretty self-explanatory!)
> >
> > Bruce Richardson (7):
> >   build: make most device classes optional
> >   build: make membership library optional
> >   build: make bpf library optional
> >   build: make efd library optional
> >   build: make distributor library optional
> >   build: make fragmentation library optional
> >   build: make reorder library optional
> >
> >  app/test/meson.build | 81 +++++++++++++++++++++++++++-----------------
> >  lib/meson.build      | 13 +++++++
> >  2 files changed, 63 insertions(+), 31 deletions(-)
>
> There is a trivial conflict on the main repo (after lib/graph change),
> so the CI did not run this v2 revision.
>
> The v1 looked fine but there was a strange error in Intel report:
> http://mails.dpdk.org/archives/test-report/2023-June/415672.html

Hum, we did not receive the ovsrobot report on v1 (not sure why), but
looking at it manually, we have one failure too:
https://github.com/ovsrobot/dpdk/actions/runs/5346924872/jobs/9694636520#step:18:219

This is because we ask for the l3fwd example but its dependencies are disabled.


-- 
David Marchand