On Tue, Jul 10, 2018 at 10:41:16AM -0300, Eduardo Lima (Etrunko) wrote:
> On 10/07/18 07:31, Frediano Ziglio wrote:
> >>
> >> On Tue, Jul 10, 2018 at 04:51:35AM -0400, Frediano Ziglio wrote:
> >>>>
> >>>> On Tue, Jul 10, 2018 at 07:21:50AM +0100, Frediano Ziglio wrote:
> >>>>> test-overflow was doing a specific test on demarshalling code.
> >>>>> Joining the 2 tests also allows to remove the dependency from the main
> >>>>> protocol allowing to run the test independently from generation
> >>>>> setting.
> >>>>
> >>>>
> >>>>
> >>>>> This is useful with Meson allowing to not generate all code.
> >>>>
> >>>> Fwiw, I don't understand the rationale here.
> >>>>
> >>>
> >>> The "if spice_common_generate_code == 'all'" code in meson.build,
> >>> basically the test was only possible if the code compiled everything
> >>
> >> But why is this a problem? :) Because we only need to generate the code
> >> for tests/*.proto, and generating more is a waste of resources? Or is
> >> this a problem for other reasons?
> >>
> >> Christophe
> >>
> > 
> > Oh... I remember Eduardo wanting to not compile with 
> > spice_common_generate_code == 'all'
> > by default so to rnu the tests you would need to change the defaults.
> > 
> 
> Just a small optimization when building either spice server or
> spice-gtk, we only generate the specific marshallers/demarshallers for
> that given case. If building spice-common itself it will generate code
> for both.

Ok, thanks, would be nice if this was clearer in the commit log (ie just
copy that instead of "This is useful with Meson allowing to not generate
all code", which explains what is being done, not *why* it's being done)

Christophe

Attachment: signature.asc
Description: PGP signature

_______________________________________________
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/spice-devel

Reply via email to