Aurelia Molzer

Regarding SIMD also take a look at https://github.com/okaneco/safe_unaligned_simd as on-stable but more unsafe alternative to std::simd. Some previous experiments with `std::simd` showed that it left quite a bit of performance...

I tried to reproduce and with the current version the ICC file is extract as expected (and bit-by-bit identical to the input as it should).

A custom wrapper sounds fine, it just introduces one additional requirement to me: We should be able to change the interface for hooks to update those bounds, in a _compatible_...

@fintelia With #2372 merged, is there anything outstanding in the issue?

The biggest concern being solved with an extra crate is the licensing / static dependency tree situation. However, I can also see users expecting formats to 'just be available' when...

Addressing the use-case of substituting the builtin decoder with a personal one is quite valid. I'd be glad to approve on the PR and registry design if the registered hooks...

I think it's great that hooks can be specified in terms of patterns now. But that is somehow at odds with use cases and evolvability. > That's because the [ImageReader::make_decoder](https://github.com/image-rs/image/blob/f95d04058a9981bee4989643f1ee0430694e417e/src/image_reader/image_reader_type.rs#L139)...

> Just to clarify, it sounds like you're looking for a code change for the *2_without_alpha functions so that they are defaults on the trait. Is that accurate? Would you...

This is still a breaking change so in light of the issue with packaging, a release of `0.25.2` before this is merged to the main branch for `0.26` then. But...

I found another reason to support this option: There are two separate concerns at play here. The decoder part, and the image manipulation part. For web, such as in `servo`,...