Opened 2 years ago
Last modified 3 weeks ago
#66259 new defect
libvpx fails to build universal on Apple silicon
Reported by: | Gcenx | Owned by: | |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | ports | Version: | |
Keywords: | Cc: | mascguy (Christopher Nielsen), jbaraban, jpmelko | |
Port: | libvpx |
Description
Not currently at my mac but here’s a log I’d received https://github.com/Gcenx/macports-wine/files/9999499/main.log
During building the x86_64 slice it’s trying to link to arm64 only symbols?
Attachments (1)
Change History (9)
comment:1 Changed 2 years ago by jbaraban
Cc: | jbaraban added |
---|
comment:2 Changed 22 months ago by jbaraban
comment:4 Changed 21 months ago by kaamui
In fact the message was pretty clear : it fails because some of the libs it tries to link were for arm64 only.
In my case, it was libavcodec and libvpx (arm64 only).
You can check it using this command lipo /opt/local/lib/<any_dylib> -info
Then, I only had to uninstall non universal ports and relaunch my initial command :
//some of the followings are probably useless, but in case I misunderstood what happened, here's everything I did sudo port uninstall ffmpeg sudo port uninstall ffmpeg +universal sudo port uninstall libvpx sudo port clean ffmpeg sudo port clean ffmpeg +universal sudo port install ffmpeg +universal
Changed 10 months ago by kencu (Ken)
comment:5 Changed 10 months ago by kencu (Ken)
if the tests are disabled, it will build universal.
% port -v installed libvpx The following ports are currently installed: libvpx @1.13.1_0+universal (active) requested_variants='+universal' platform='darwin 23' archs='arm64 x86_64' date='2024-01-22T22:03:49-0800'
comment:6 Changed 3 weeks ago by ryandesign (Ryan Carsten Schmidt)
Cc: | jpmelko added |
---|
See duplicate #71115 for more info.
comment:7 Changed 3 weeks ago by jpmelko
Starting from scratch (deleting /opt/local), libvpx is installed
titi@Dionysos ~ % port installed libvpx The following ports are currently installed: libvpx @1.13.1_0 (active)
comment:8 Changed 3 weeks ago by ryandesign (Ryan Carsten Schmidt)
Good, that is consistent with existing observations.
Hi - any chance of a fix for this? Thanks