Re: [PATCH][master-next] webkitgtk: add PACKAGECONFIG gobject-introspection

[Khem Raj <raj.khem@gmail.com>]

On Thu, 2020-02-20 at 14:11 -0800, Andre McCurdy wrote:
> On Thu, Feb 20, 2020 at 2:06 PM Khem Raj <raj.khem@gmail.com> wrote:
> > On Thu, 2020-02-20 at 14:00 -0800, Andre McCurdy wrote:
> > > On Thu, Feb 20, 2020 at 12:33 PM Khem Raj <raj.khem@gmail.com>
> > > wrote:
> > > > On Thu, Feb 20, 2020 at 12:24 PM Andreas Müller <
> > > > schnitzeltony@gmail.com> wrote:
> > > > > On Thu, Feb 20, 2020 at 1:45 PM Andreas Müller <
> > > > > schnitzeltony@gmail.com> wrote:
> > > > > > On Thu, Feb 20, 2020 at 11:07 AM Alexander Kanavin
> > > > > > <alex.kanavin@gmail.com> wrote:
> > > > > > > You need to at the same time drop the introspection
> > > > > > > enablement logic from EXTRA_OECMAKE I think?
> > > > > > > This line:
> > > > > > > http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-sato/webkit/webkitgtk_2.26.4.bb?h=master-next#n70
> > > > > > > 
> > > > > > > And for consistency, it would be good to do the same for
> > > > > > > gtk-
> > > > > > > doc.
> > > > > > 
> > > > > > This is odd: I fell over this because there was no gir
> > > > > > created.
> > > > > > Did
> > > > > > not notice it was few lines below. Need to check that -
> > > > > > maybe I
> > > > > > am a
> > > > > > bit off-sync to master or something is wrong elsewhere.
> > > > > > 
> > > > > > Will sync to master and send V2 then.
> > > > > > 
> > > > > Finally had time to look into again. In the recipe there is
> > > > > 
> > > > > GI_DATA_ENABLED_armv7a = "False"
> > > > > GI_DATA_ENABLED_armv7ve = "False"
> > > > > 
> > > > > => My patch is wrong it overrides above accidentally - thanks
> > > > > Alex!
> > > > > 
> > > > > Since webkit builds gir fine (my machine is armv7ve) that
> > > > > does
> > > > > not
> > > > > seem necessary any more. I would like to send out a patch
> > > > > removing at
> > > > > least armv7ve - shall I take the risk to remove armv7a -
> > > > > maybe as
> > > > > two
> > > > > patches?
> > > 
> > > These two should be equivalent for all practical purposes and so
> > > should be updated together. ie remove them both, one patch.
> > 
> > it will run user mode qemu underneath, which is not same for both,
> > are
> > you sure libraries and apps it will run are identical ?
> 
> Binaries built for armv7ve will contain HW integer division
> instructions. Binaries build for armv7a will not.
> 
> That should have no impact on gobject-introspection.

if it executes correctly. The reason why this was done in first place
is that introspection app inside qemu either segfaulted or got hung.

> 
> There's a much bigger risk in letting these two build variants drift
> apart than there is from updating them together but only testing one.

qemu has failed in mysterious ways with slight differences in ISAs
Not oppposed to doing it, All I am asking for is to try it before
making the change.