Cameron
1. With or without explicit tokens, the final dequeue order is not guaranteed. Only the ordering between elements enqueued by the same thread (implicit) or token (explicit) is preserved. This...
Depends on whether you have concurrent consumers, but if there's just one, then yes, at that point locking during enqueue with a `ReaderWriterQueue` instead may be better, if you truly...
Yes? How big is the bitmap you're trying to compress? Have you tried a profiler to see where the memory is going? More details, please! :-)
Great, I'm glad the `IDataOutput` version works for you. Yes, in the same gh-IDataOutput branch, there's a `decode` method that accepts an `IDataInput`. It's synchronous and only works on PNGs...
Hmm. Looks like the bitmap really is just too big to load in flash (what is its width/height?). You can get the width and height from the header of the...
Ah. That's a big image :-) When loading a PNG, the bitmap ends up being in memory twice (since there's the raw uncompressed memory from the decode step plus the...
Let me know if there are more unresolved warnings, I don't have the exact same version of GCC locally.
Not sure I understand. Can you provide an example?
I see! This is because POSIX semaphores use absolute deadlines based on CLOCK_REALTIME. I'll see if I can conditionally detect the presence of a glibc 2.30 extension that would work...
Yes. Glibc 2.30+ has a non-standard `sem_clockwait` function that allows CLOCK_MONOTONIC time to be specified instead.