diff mbox series

[meta-oe] edid-decode: allow to build native variant

Message ID 20240414125805.2253456-1-f_l_k@t-online.de
State Accepted
Headers show
Series [meta-oe] edid-decode: allow to build native variant | expand

Commit Message

Markus Volk April 14, 2024, 12:58 p.m. UTC
This is a requirement for libdisplay-info and fixes:
Program edid-decode found: NO

If libdisplay-info can provide edid-decode-native it looks like:
libdisplay-info| Program edid-decode found: YES (/home/flk/poky/build-hypr/tmp/work/corei7-64-poky-linux/gamescope/3.14.3/recipe-sysroot-native/usr/bin/edid-decode)
libdisplay-info| Program ./edid-decode-check.sh found: YES (/home/flk/poky/build-hypr/tmp/work/corei7-64-poky-linux/gamescope/3.14.3/git/subprojects/libdisplay-info/test/./edid-decode-check.sh)
libdisplay-info| Program ./edid-print-check.sh found: YES (/home/flk/poky/build-hypr/tmp/work/corei7-64-poky-linux/gamescope/3.14.3/git/subprojects/libdisplay-info/test/./edid-print-check.sh)
libdisplay-info| Program ./edid-decode-diff.sh found: YES (/home/flk/poky/build-hypr/tmp/work/corei7-64-poky-linux/gamescope/3.14.3/git/subprojects/libdisplay-info/test/./edid-decode-diff.sh)

Signed-off-by: Markus Volk <f_l_k@t-online.de>
---
 meta-oe/recipes-support/edid-decode/edid-decode_git.bb | 2 ++
 1 file changed, 2 insertions(+)

Comments

Ross Burton April 15, 2024, 11:07 a.m. UTC | #1
On 14 Apr 2024, at 13:58, Markus Volk via lists.openembedded.org <f_l_k=t-online.de@lists.openembedded.org> wrote:
> 
> This is a requirement for libdisplay-info and fixes:
> Program edid-decode found: NO

Surely that’s a bug in the meson.build?  It doesn’t want to find a native binary.

Ross
Markus Volk April 15, 2024, 11:32 a.m. UTC | #2
On Mon, Apr 15 2024 at 11:07:45 AM +00:00:00, Ross Burton 
<Ross.Burton@arm.com> wrote:
> Surely that’s a bug in the meson.build?  It doesn’t want to find 
> a native binary.

Its added here, I think mostly for testing purpose:

<https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L1>
<https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L10>
<https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L59>
<https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L66>

If edid-decode is added a s a native dependency, an additional binary 
will be installed '/usr/bin/di-edid-decode'
Ross Burton April 15, 2024, 11:41 a.m. UTC | #3
> On 15 Apr 2024, at 12:32, Markus Volk <f_l_k@t-online.de> wrote:
> 
> On Mon, Apr 15 2024 at 11:07:45 AM +00:00:00, Ross Burton <Ross.Burton@arm.com> wrote:
>> Surely that´s a bug in the meson.build? It doesn´t want to find a native binary.
> 
> 
> Its added here, I think mostly for testing purpose:
> 
> https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L1
> https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L10
> https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L59
> https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L66
> 
> If edid-decode is added a s a native dependency, an additional binary will be installed '/usr/bin/di-edid-decode’

Are you sure about that?  All of those references are specific to the test suite (which would want native binaries, unless they make the effort to support installable tests).  The di-edid-decode binary is buillt and installed by https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/di-edid-decode/meson.build which doesn’t have any conditionals.

Ross
Markus Volk April 15, 2024, 11:50 a.m. UTC | #4
Thats absolutely true . But still I dont think it is a bug in the 
libdisplay-info build system, because edid-decode is requested for the 
tests and is requested native explicitly.

On Mon, Apr 15 2024 at 11:41:38 AM +00:00:00, Ross Burton 
<Ross.Burton@arm.com> wrote:
> 
> 
>>  On 15 Apr 2024, at 12:32, Markus Volk <f_l_k@t-online.de 
>> <mailto:f_l_k@t-online.de>> wrote:
>> 
>>  On Mon, Apr 15 2024 at 11:07:45 AM +00:00:00, Ross Burton 
>> <Ross.Burton@arm.com <mailto:Ross.Burton@arm.com>> wrote:
>>>  Surely that´s a bug in the meson.build? It doesn´t want to find 
>>> a native binary.
>> 
>> 
>>  Its added here, I think mostly for testing purpose:
>> 
>>  
>> <https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L1>
>>  
>> <https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L10>
>>  
>> <https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L59>
>>  
>> <https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/test/meson.build#L66>
>> 
>>  If edid-decode is added a s a native dependency, an additional 
>> binary will be installed '/usr/bin/di-edid-decode’
> 
> Are you sure about that?  All of those references are specific to the 
> test suite (which would want native binaries, unless they make the 
> effort to support installable tests).  The di-edid-decode binary is 
> buillt and installed by 
> <https://gitlab.freedesktop.org/emersion/libdisplay-info/-/blob/main/di-edid-decode/meson.build> 
> which doesn’t have any conditionals.
> 
> Ross
diff mbox series

Patch

diff --git a/meta-oe/recipes-support/edid-decode/edid-decode_git.bb b/meta-oe/recipes-support/edid-decode/edid-decode_git.bb
index bb8c9de71..6eaec800c 100644
--- a/meta-oe/recipes-support/edid-decode/edid-decode_git.bb
+++ b/meta-oe/recipes-support/edid-decode/edid-decode_git.bb
@@ -14,3 +14,5 @@  S = "${WORKDIR}/git"
 do_install() {
     oe_runmake 'DESTDIR=${D}' install
 }
+
+BBCLASSEXTEND = "native"