Sean Gillies

@dugalh since those are closely related, let's keep them together here until we decide that they should break out. I'd love to know more about your use cases and what...

@dugalh thanks for the info! I'm going to think harder about the zipfile case. Indeed, it's the zipfile that's the container, not the directory containing the zipfile. For the threading...

@dugalh I can reproduce the problem with building overviews, with both `io.open` and `fsspec`. It seems, from logging, to occur around writing. This is the last log message before the...

@rouault are there implications for VSI plugins in a multithreaded situation that I have missed? Such as locking or mutual exclusion?

@rouault the GIL must be acquired for these callbacks because they involve Python objects. The most common GIL deadlocks I read about on the internet involve threads that are destroyed...

In #3115 I have added tests of multi-threaded reprojection and TIFF decompression. I observe no deadlocks in those cases. This suggests to me that the overview deadlock is a special...

@aloi214 Thanks for the report! I have a few questions: * Is fastapi required to produce the increasing memory? What do you see if you just loop over create_mask() without...

@groutr I haven't either. In theory, it might work, but only if every input had the exact same rotation. And in my experience, collections with rotated data have some variation....

@bkavlak thanks for reporting this! That example doesn't look right to me. My understanding is that it's best to pass `rpc_dem` with a path to a DEM, or `rpc_height` with...

@DFEvans thanks for the report and thank you @dugalh for taking a look. This will be fixed in version 1.4.4.