EGL/NVidia can sometimes clip the cursor based on its plane regardless of the depth mask. This avoids the issue by forcing the cursor up over the existing plane.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/12183
Makes all canvases obey the same rendering parameter. Also keeps the
canvas selection with the anti-alias selection, which makes logical
sense. Reduce the antialias selection to a single choice rather than
two different choices for different canvases
Glew has the problem that it has to be selected at build time if GLX or
EGL is supported by the library, and this in not encoded in the library
name, nor ABI, nor anything.
Then it's easy to get into the situation that a binary is built but
cannot run because glew supports an API different from the one used by
wxWidgets, or the binary fails to link in the end after all objects are
compiled.
epoxy can support both with the same library avoiding this problem.
epoxy is not initialized explicitly, replaced initialization with
version check where one was not done already.
It seems to be available as vcpkg https://vcpkg.link/ports/libepoxy
There are problems related to GL context switching on Windows which does
not seem to be used in kicad
https://github.com/anholt/libepoxy#known-issues-when-running-on-windows
There is also a problem related to multithreaded rendering on Windows
https://github.com/anholt/libepoxy/pull/265 It's harder to tell if
threading is used for rendering but it does not look like kicad is doing
anything complex enough to warrant using multiple rendering threads.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/20630
Fixes https://gitlab.com/kicad/code/kicad/-/issues/12543
This is the second try at 5326c36a5f. The difference here is that we
have moved GL_CONTEXT_MGR into kicommon first which will hopefully
address some of the Windows linkage issues
The GL context lock needs to be shared across kifaces. Otherwise, we
can end up blocking the lock from one kiface. Unfortunately, I can't
find the issue in GitLab right now for where the footprint viewer shows
a blank screen after opening too many contexts. But that's what this
fixes.
The GL context lock needs to be shared across kifaces. Otherwise, we
can end up blocking the lock from one kiface. Unfortunately, I can't
find the issue in GitLab right now for where the footprint viewer shows
a blank screen after opening too many contexts. But that's what this
fixes.
Having thread pool as its own singleton in the library meant that each
kiface had its own threadpool, leading to many multiples of the threads
being started. Placing a singleton class in PGM_BASE ensures that all
kifaces use the same thread pool.
The singleton class can be extended to provide single instance
guarantee for any element across kifaces
Recommendation is to avoid using the year nomenclature as this
information is already encoded in the git repo. Avoids needing to
repeatly update.
Also updates AUTHORS.txt from current repo with contributor names
This means that an early return or an exception between
a manual Push/Pop (or an omission of the Pop) cannot
corrupt the layer stack.
It also means the GAL doesn't have to maintain its own
stack (with the in-scope GAL_SCOPED_ATTRS taking that role).
Reomve the Push/PopDepth functions, as they're only ever
used in pairs, and doing it manually needs more care.
Fixes https://gitlab.com/kicad/code/kicad/-/issues/16304
Add dedicated UI rendering layers
Switch to screen-space rendering to avoid blurriness
Fix a bug in OpenGL GAL that causes layer
ordering to be broken when using Scale
Fix some issues with VIEW_GROUP layer ordering
