* handle premature EOF in the PBM decoder
* BufferedReadStreamExtensions: remove the 'Try' prefix
* count EOF hits in BufferedReadStream
* use EofHitCounter in pbm tests
* Naming convention tweaks
---------
Co-authored-by: James Jackson-South <james_south@hotmail.com>
* Add webp decoder option BackgroundColorHandling to decide howto handle the background color in the ANIM chunk
* Add test for Issue 2528
* Update src/ImageSharp/Formats/Webp/BackgroundColorHandling.cs
Co-authored-by: James Jackson-South <james_south@hotmail.com>
* Fix path to test file
* Another attempt fixing the file path: path should be lower case
* Revert fa4febf7
---------
Co-authored-by: James Jackson-South <james_south@hotmail.com>
* Handle EOF in bit reader when data is bad.
* Allow parallel processing of multi-megapixel image
* Stream seek can exceed the length of a stream
* Try triggering on release branches
* Update JpegBitReader.cs
* Skin on Win .NET 6
* All Win OS is an issue
* Address feedback
* add validation to CanIterateWithoutIntOverflow
---------
Co-authored-by: antonfirsov <antonfir@gmail.com>
Hm, I remembered that movemask isn't the fastest, and ptest (TestZ in .NET-terms) is faster but current benchmarks didn't prove this, also Intel's instruction table didn't show any benefit in terms of latency or throughput.
Thus simplified that check.
Also I fixed a little bug with PreviouslySeenPixels array
It's weird, but I don't see any problems with encoding and the tests are ok, I wrote the memory stream to files and they look the same, but the hashes and bytes aren't the same
This is very weird