[PySide] Keeping GUI responsive
public at enkore.de
Fri Dec 6 17:45:35 CET 2019
You still get multiple threads, which is still useful. Note that many
operations do unlock the GIL and give you parallelism, for example
performing I/O. So using QThreads in Python to implement parallel I/O or
asynchronous I/O is very much a reasonable thing to do.
It just won't help with speeding up CPU-limited Python code.
Am 06.12.19 um 17:33 schrieb Jason H:
> What those links say and what is actually happening here are two different things.
> It looks like the only way he can do this is to write the thread code in C++, and export that through Shiboken to Python. So the lines:
> for idx, val in enumerate(self.dataset):
> self.dataset[idx] = val * 50
> Have to:
> 1) Operate on Qt data (QMap?)
> 2) Have the QRunnable implemented in C++
> 3) Export the class to Python,
> 4) Have Python start the thread
> I find myself asking why QThread exits at all n PySide or at least not have documentation saying that using the class(es) as one would in C++ will not have the same behavior as in C++.
More information about the PySide