So what they did was try to force SDL to not dynamically load Xlib, because 
they were concerned SDL wouldn't have X11 as a package dependency if so 
(although that seems like something that should be listed as a dependency at 
the package level, and in a land where Wayland exists, is now incorrect 
behavior anyhow), and the "just to be sure" was to attempt to disable _all_ 
dynamic loading, which is also incorrect.

Also, it was for SDL 1.2, which was a totally different beast.   :)

One could argue that this makes sense for a distro package, because they
are guaranteeing that the dependencies will exist...a lot of SDL's
dynamic loading mojo is to let it make decisions about what targets to
choose on an arbitrary system, but in a day when both Wayland and X11
(and OpenGL and Vulkan...) may or may not exist on a given system, this
functionality is needed now.

Also, bugs in SDL's configure script aside, --disable-loadso will break
SDL_LoadObject()'s functionality, so it's not a good idea in any case.

I'm not well-versed in how Debian packaging works, but it seems to me
that it should list X11 as an explicit dependency of the package, and
then let SDL dlopen it at runtime (and recover gracefully if it can't).
I wonder if Debian can say "this needs the X11 headers when building,
but doesn't need the X11 package at runtime"...

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1740517

Title:
  SDL2 2.0.6 isn't compiled with Vulkan support

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/libsdl2/+bug/1740517/+subscriptions

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to