[Bug 1366207] [NEW] qgl.h typedefs GLdouble to float instead of double

[Saikrishna Arcot]

Public bug reported:

In qgl.h, for the OpenGL ES 2.0, GLdouble is typedef'ed to float. This
causes a compile error in the armhf architcture in downstream programs
including OpenSceneGraph and other programs. Recently, Qt committed a
patch (https://codereview.qt-project.org/#/c/93984/1//ALL) for Qt 5 to
change the typedef from float to double, since (according to the commit
message), the typedef was from S60 and wasn't being used anywhere.

I've backported the patch to Qt 4, but I wasn't able to do a full build
for armhf; the build was taking 12+ hours for me. However, it does seem
to have compiled the OpenGL source code files and a few examples related
to this successfully. Therefore, I wasn't fully able to verify that this
compiles successfully.

This is intended to be a bugfix-only update; however, since the ABI is
effectively being changed, this change might warrant some discussion
here. The expectation for downstream packages on armhf is that more
packages can be compiled, since some packages have this typedef error,
but some packages may also fail to compile if they rely on the fact that
this was a float. I'm expecting that the number for the latter case is
low to 0, since this is a widening conversion.

I've attached the raw patch file. This will need to be added to
debian/patches.

** Affects: qt4-x11 (Ubuntu)
     Importance: Undecided
         Status: Won't Fix

** Patch added: "qopengl-typedef-fix.patch"
   https://bugs.launchpad.net/bugs/1366207/+attachment/4196694/+files/qopengl-typedef-fix.patch

-- 
You received this bug notification because you are a member of Kubuntu
Bugs, which is subscribed to qt4-x11 in Ubuntu.
https://bugs.launchpad.net/bugs/1366207

Title:
  qgl.h typedefs GLdouble to float instead of double

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/qt4-x11/+bug/1366207/+subscriptions