Re: [OE-core] [PATCH 2/2] convert-overrides.py: handle few more cases of overrides

["Denys Dmytriyenko" <denis@denix.org>]

On Wed, Aug 04, 2021 at 06:04:27PM +0100, Richard Purdie wrote:
> On Wed, 2021-08-04 at 12:07 -0400, Denys Dmytriyenko wrote:
> > On Wed, Aug 04, 2021 at 04:29:45PM +0200, Martin Jansa wrote:
> > > Yes, I see the same in meta-oe and meta-webosose.
> > > 
> > > It's easy to spot false translations in incremental conversion where the
> > > diff is relatively small, but doing initial conversion on big layer with
> > > this change already included might be more painful.
> > > 
> > > On Wed, Aug 4, 2021 at 12:31 PM Richard Purdie <
> > > richard.purdie@linuxfoundation.org> wrote:
> > > 
> > > > On Tue, 2021-08-03 at 21:28 -0400, Denys Dmytriyenko wrote:
> > > > > Add task-configure and few more supported values of TARGET_OS override.
> > > > > 
> > > > > Signed-off-by: Denys Dmytriyenko <denis@denix.org>
> > > > > ---
> > > > >  scripts/contrib/convert-overrides.py | 5 +++--
> > > > >  1 file changed, 3 insertions(+), 2 deletions(-)
> > > > > 
> > > > > diff --git a/scripts/contrib/convert-overrides.py
> > > > b/scripts/contrib/convert-overrides.py
> > > > > index 6505b18..23f63a2 100755
> > > > > --- a/scripts/contrib/convert-overrides.py
> > > > > +++ b/scripts/contrib/convert-overrides.py
> > > > > @@ -36,9 +36,10 @@ vars = vars + ["mipsarch", "x86-x32", "mips16e",
> > > > "microblaze", "e5500-64b", "mip
> > > > >  vars = vars + ["class-native", "class-target", "class-cross-canadian",
> > > > "class-cross", "class-devupstream"]
> > > > >  vars = vars + ["tune-",  "pn-", "forcevariable"]
> > > > >  vars = vars + ["libc-musl", "libc-glibc",
> > > > "libc-newlib","libc-baremetal"]
> > > > > -vars = vars + ["task-compile", "task-install", "task-clean",
> > > > "task-image-qa", "task-rm_work", "task-image-complete", "task-populate-sdk"]
> > > > > +vars = vars + ["task-configure", "task-compile", "task-install",
> > > > "task-clean", "task-image-qa", "task-rm_work", "task-image-complete",
> > > > "task-populate-sdk"]
> > > > >  vars = vars + ["toolchain-clang", "mydistro", "nios2", "sdkmingw32",
> > > > "overrideone", "overridetwo"]
> > > > >  vars = vars + ["linux-gnux32", "linux-muslx32", "linux-gnun32",
> > > > "mingw32", "poky", "darwin", "linuxstdbase"]
> > > > > +vars = vars + ["linux", "linux-gnueabi", "elf", "eabi"]
> > > > 
> > > > I have reservations about linux and elf being here since this does lead to
> > > > a lot of false translations for OE-Core. Since OE-Core is now translated I
> > > > guess that isn't the driving factor so I'll probably take this but it is
> > > > worth
> > > > being aware of.
> > > > 
> > > > I did just have to retest a translation of core so I now need to filter all
> > > > these out the diff...
> > 
> > Should I re-submit with 'linux' and 'elf' removed?
> 
> I think given OE-Core isn't alone with the issue, that may be wise. I can tweak 
> the patch if you want?

Richard,

Whichever is easier for you.

-- 
Denys