On Thu, Dec 09, 2021 at 08:15:01AM -0500, Aaron Conole wrote:
> Jerin Jacob <jerinjac...@gmail.com> writes:
>
> > On Thu, Dec 9, 2021 at 12:30 AM Jie Zhou <j...@linux.microsoft.com> wrote:
> >>
> >> Even though test_interrupts.c can compile on Windows, skip interrupt
> >> tests for now since majority of eal_interrupt on Windows are stubs.
> >> Will remove the skip after interrupt being fully enabled on Windows.
> >>
> >> Signed-off-by: Jie Zhou <j...@linux.microsoft.com>
> >> Acked-by: Dmitry Kozlyuk <dmitry.kozl...@gmail.com>
> >>
> >> ---
> >> app/test/test_interrupts.c | 10 ++++++++++
> >> 1 file changed, 10 insertions(+)
> >>
> >> diff --git a/app/test/test_interrupts.c b/app/test/test_interrupts.c
> >> index 2a05399f96..eec9b2805b 100644
> >> --- a/app/test/test_interrupts.c
> >> +++ b/app/test/test_interrupts.c
> >> @@ -12,6 +12,15 @@
> >>
> >> #include "test.h"
> >>
> >> +#ifdef RTE_EXEC_ENV_WINDOWS
> >
> > Across the series,
> > Instead of adding conditional compilation everywhere, Why not disable
> > specific file
> > for compilation for windows?
> > Purpose of EAL to abstract the differences in execution environment
> > and application
> > should not know that.
>
> I think this was done because there would be two test lists in the meson
> unit test file. But this is the second comment about these ifdef's, and
> maybe we should revisit that discussion. Is there a different way to
> accomplish not running the tests which are not appropriate for windows
> builds, while not having two overlapping lists of unit tests in the
> meson build file?
>
I'm wondering if a reasonable compromise solution might be to have the
build system expose a usable RTE_EXEC_ENV symbol that can be used in C-code
if statements rather than just in ifdefs. That would allow us to easily add
e.g.
if (RTE_EXEC_ENV == rte_env_linux)
return TEST_SKIPPED;
into each test function needing it. Two lines of C-code is a lot easier to
add and manage than #ifdefs covering the whole file, or alternative lists
in meson.
/Bruce
