[Buildroot] [PATCH] ffmpeg: fix static linking build failure when using libavutil

Giulio Benetti giulio.benetti at micronovasrl.com
Wed Oct 17 10:36:44 UTC 2018


Hello,

Il 16/10/2018 21:31, Giulio Benetti ha scritto:
> Hello Thomas,
> 
> Il 12/10/2018 22:54, Giulio Benetti ha scritto:
>> Hello,
>>
>> Il 12/10/2018 21:50, Thomas Petazzoni ha scritto:
>>> Hello,
>>>
>>> On Fri, 12 Oct 2018 19:38:14 +0200, Giulio Benetti wrote:
>>>
>>>>>>>> Indeed, when we're dynamic linking, there is no need to link consumers
>>>>>>>> of libavutil with libdrm. It is only when we're statically linking that
>>>>>>>> it should be done. And that's exactly what Libs vs. Libs.private is for.
>>>>>>>
>>>>>>> You've explained me it very well, but by now FFmpeg upstreamed my patch:
>>>>>>> https://git.ffmpeg.org/gitweb/ffmpeg.git/commit/c50dc77ac708e98d02da7c422a6b9cbf9f565aa5
>>>>>>>
>>>>>>> So, do I rework and resubmit to them or can I re-create a patch for BR
>>>>>>> with what you've pointed me above?
>>>> I've missed one thing to ask you about ffmpeg:
>>>> Can I still use this hack:
>>>> https://git.ffmpeg.org/gitweb/ffmpeg.git/blobdiff/bc2301429e9c779237e43acf913331af018211f2..c50dc77ac708e98d02da7c422a6b9cbf9f565aa5:/configure
>>>>
>>>> to obtain if -ldrm is used and then add it to Libs.private like this:
>>>> Libs.private: -ldrm
>>>>
>>>> as you've pointed me.
>>>>
>>>> Or there is something that makes it simpler and better?
>>>
>>> I'm not sure what "hack" you're talking about here. My only point it is
>>> only needed to link "indirect" libraries when static linking, and
>>> therefore such "indirect" libraries should be listed in Libs.private
>>> and not Libs in pkg-config files.
> 
> I've worked this out, but I have a question:
> Since FFMpeg upstreamed my previous wrong patch:
> https://git.ffmpeg.org/gitweb/ffmpeg.git/commit/c50dc77ac708e98d02da7c422a6b9cbf9f565aa5
> 
> now I would submit another patch to FFMpeg to correct Libs vs
> Libs.private problem.
> 
> Here is the new patch:
> https://github.com/giuliobenetti/FFmpeg/commit/d682f418b7912b9d9e1818aefa61a51a2868fb20
> 
> So:
> do I have to submit to Buildroot a patch containing both patches?
> Or only one that merge both patches together even if it's not upstreamable?

I think I've answered myself, I need to:
- send patch to Revert "c50dc77ac708e98d02da7c422a6b9cbf9f565aa5"
- send new patch that is the squash of these 2 patches(1 already 
upstream and 1 that fixes the problem):
 
https://github.com/giuliobenetti/FFmpeg/commit/c50dc77ac708e98d02da7c422a6b9cbf9f565aa5
 
https://github.com/giuliobenetti/FFmpeg/commit/d682f418b7912b9d9e1818aefa61a51a2868fb20

-- 
Giulio Benetti
CTO

MICRONOVA SRL
Sede: Via A. Niedda 3 - 35010 Vigonza (PD)
Tel. 049/8931563 - Fax 049/8931346
Cod.Fiscale - P.IVA 02663420285
Capitale Sociale € 26.000 i.v.
Iscritta al Reg. Imprese di Padova N. 02663420285
Numero R.E.A. 258642


More information about the buildroot mailing list