Open
Description
When we read from stdin, bytes are consumed, however if we are just probing content (e.g. the image subsystem figuring out if something is an image), this will garble the state for the next one.
So if we could buffer these and provide some internal re-wind mechanism, that would be good. Worst case, we've to buffer everything to disk first, but if the API allows enough access to provide data in different ways, we can buffer the first chunks, then later just connect the existing stream. To be figured out.
Metadata
Metadata
Assignees
Labels
No labels