Topic: Are handles ever recycled?

[FireFox]

Hello, I'm wondering whether it's safe to use handles as unique identifiers? Specifically for BASS_ChannelSetSync. As sync callbacks as processed and removed (via the BASS_SYNC_ONETIME flag), is there a chance that a future call to BASS_ChannelSetSync could ever return a handle that's the same as one that was previously processed? And if so if that is a realistic chance and should therefore avoid using handles as global identifiers? Thank you

[MB_SOFT]

i guess 32 bit handles starts from -2147483648 and end to 2147483647 so in my opinion it's not realistic to reach 2147483647

[FireFox]

Sure, but I'm not sure whether it's a coincidence that new handles always seems to increment or whether they can reset back to -2147483648 way earlier if those initial handles are freed/disposed early.

[Ian @ un4seen]

All sample/channel/DSP/FX/sync handles are currently globally unique, with each new handle being 1 higher than the previously allocated handle. Those handles are 32-bit with the highest bit always set, which means they won't start being recycled until over 2 billion (2^31) have been created. Since BASS 2.4.17, they'll now actually start being recycled after 1 billion (2^30) have been created, because bit 30 is always set for samples and their channels and unset for all other channel types (streams/mods/recordings).

It isn't really necessary for sync handles to be globally unique because BASS_ChannelRemoveSync requires a channel handle too, so it is possible that they could be changed to only be unique per-channel. So it might not be safe to assumes that sync handles will always be globally unique, if that is what you were asking?