Jonathan Behrens
Jonathan Behrens
Yes. After this lands, `image` will switch to only enabling the rayon feature on `exr` when its own rayon feature is enabled
Something along these lines was previously attempted in #377. It wasn't able to get a performance improvement at the time, but some of the scaffolding around accumulating multiple paeth rows...
Want to submit a PR?
Figuring out the proper fix is a big part of the work of fuzzing. In this case, the surrounding code is computing a sum and then checking whether it exactly...
I took a look at the build timings for this crate, and one thing that stood out was that about 3/4 of total compilation time was consumed by only a...
This split likely wouldn't decrease the number of dependencies by very much. The reason is that generally we use the same crate for both encoding and decoding. And even within...
Thanks for sharing those compile times! 18 seconds is certainly alarming, especially gives how few features are included. On my machine, that feature combination takes closer to 3 seconds for...
With #2470 and #2468 merged, compile times should be much improved. In particular, if you don't use image encoders then the time spent compiling them should be substantially reduced.
I'm supportive of this investigation, though we should make sure to benchmark carefully and to double check that the same performance gains aren't possible the current interfaces. [I poked around...
So far we've avoided ever dealing with `&[Rgb]` or similar slices, so I'm not thrilled about changing that. For now using unsafe casts seems fine, but I'd want to see...