Mateusz Łoskot

### Test 3 Updated to [Test variety of **bit-aligned pixel** configurations: bgr111, bgr121, bgr222, bgr232, bgr233](https://github.com/mloskot/boost-gil-workshop/commit/fe634604d04670bf34022e430e98c8ef30282167) and the results are even more interesting. The checksum is calculated for **regular** RGB8...

I only reproduced this particular issue with MSVC. But, enabling `variant=release` wtth GCC (see #50), leads to failures of `gil/test/channel.cpp` tests which I have not investigated yet.

@chhenning Thanks for this. I think it may confirm my assumption from https://github.com/boostorg/gil/issues/49#issuecomment-375383391 about bug around the locator.

## Running `minimal_chhenning.cpp` test Source of `minimal_chhenning.cpp` from https://github.com/boostorg/gil/issues/49#issuecomment-377725324 ### VS2017 x64 Release  ### clang 5.0 on Ubuntu 18.04 (optimised, with `-fsanitizer=integer`) ``` $ export UBSAN_OPTIONS=print_stacktrace=1 $ clang++-5.0 -std=c++11...

@chhenning I don't want to speculate about a compiler bug, yet. I think it's more reasonable to rather look for eliminating evidence on _what the issue is not caused by_,...

Absolutely no rush, take your time.

> Are there any intentions to add WebP support to Boost GIL? Yes, I think it's a very good idea to support the format. > If yes: Is someone already...

If you want to work on something, you should rather know how to do it, don't you think so?

I'm sorry but no resources for babysitting. If you have very specific issue or question about GIL, then feel free to ask.

It is fine to try format any new code using the `.clang-format` that we are discussing as part of https://github.com/boostorg/gil/pull/596