This results in size savings for thirdparty libraries that are only used once, and a size increase for libraries used more than once (eg. OpenSSL). More importantly, it prevents conflicts with other versions of the libraries loaded by other Python modules, such as the version of OpenSSL that the hmac module uses.
We need to be careful to only apply this for packages that are either used once, used in a plug-in module, or if we don't need to pass thirdparty library structures across Panda library boundaries. For example, I haven't done this for Bullet, since the Bullet symbols need to be available through libpandabullet.so due to the fact that pandabullet contains calls to the Bullet libraries in the inline methods.
Fixes#851
These were probably added before makepanda gained the ability to automatically add dependencies of static libraries when linking statically.
They don't really do any harm--the linker will probably optimize these out automatically--but it's cleaner not to add unused dependencies.
It doesn't seem like this setting is actually needed (because mayapath does the necessary set-up), but while it's set, we might as well set it to something more sensible.
Let's show a warning when an option has been removed but its presence would not have affected the build in an important way. For options like --use-awesomium we should still show an error since it invokes behavior we no longer support, but it should be harmless to specify --no-awesomium.
This changes the Assimp include path to point to the
directory containing assimp/ instead of inside assimp/
directly. This is for consistency with how the Assimp
project defines their "include path" and keeps the actual
inclusions themselves unambiguous (since Assimp's headers
have fairly generic filenames).
This is done by adding a PyTargetAdd function, which builds the target into a Python ABI-specific temporary directory, allowing multiple Python versions to be built into the same built dir side-by-side. This could greatly speed up buildbot builds.
It also paves the way for building multiple Python versions in the same makepanda call / installer by changing PyTargetAdd to add one target per enabled Python version.
Since this still takes a long time to build, even with the previous change, it would be better for one CPU to chew on this in the background while the rest of the build continues.
The no-arguments get_text() and set_text() will now return Unicode strings in Python 3, but passing in an encoding will make them return/take bytes objects.
In Python 2, they all take regular strings, but Unicode is also accepted by the no-argument get_text() and set_text().
In the future we probably want to remove most of this interface for Python users, to whom all this is unnecessary since it duplicates functionality already in the standard library.
This reimplements the spinlock on top of std::atomic_flag, which is guaranteed to be lockless. It also inserts the PAUSE (REP NOP) instruction which is strongly recommended to be placed in busy-wait loops by Intel.
This also includes a recursive spinlock implementation.
The spinlock implementation is disabled by default, but can be enabled by adding the --override MUTEX_SPINLOCK=1 flag to makepanda.
Rationale:
1. Per standard, fstream::open takes 2 arguments.
If platforms add a third, they're out of spec.
2. The only platform I could find that takes a file
mask specifically as the third argument is IRIX,
which Panda hasn't targeted in forever.
3. The mask being requested isn't even particularly
interesting - falling back to a platform default
is best.
4. When USE_PANDAFILESTREAM is defined, pfstream
is implemented as PandaFileStream, which doesn't
have a three-argument open() and breaks immediately.
5. makepanda doesn't ever define HAVE_OPEN_MASK
6. It's been broken for so long that, if it were
important to anybody, it would have been fixed by now.
rdb
kamgha
Sam Edwards
Sebastian Hoffmann
Younguk Kim
Mitchell Stokes