In D21046#461591, @tcanabrava wrote:

can you put a screenshoot of the before / after ? I can't really read math.

• Add blank line to come back as original

• Remove unnecessary headers

• Correct the use of QString removing std::move

In D20398#446407, @pino wrote:

Which warnings do the two function trigger?
Also, adding { ... } breaks the code style.

• Fix code style

This is the warning I get
https://clang.llvm.org/extra/clang-tidy/checks/readability-static-accessed-through-instance.html

In D20335#446223, @tcanabrava wrote:

true, but then the instance is there anyway. what we gain from this?

• Fix code style

In D20334#445058, @pino wrote:

The API documentation of kpEffectInvert::applyEffect() says:

// Inverts the colours of each pixel in the given image.
// These functions differ from QImage::invertPixels() in the following ways:
//
// 1. for 8-bit images, it inverts the colours of the Colour Table
//    (this means that you would get visually similar results to inversion
//     at higher bit depths - rather than a "random-looking" inversion
//     depending on the contents of the Colour Table)
// 2. never inverts the Alpha Buffer

Did you check that the change does not invalidate these? (especially point 1)

Yes, I tested with all kind of images, from 1 to 32 bits. Everything works the same as before.
The changes affect only if all channels are selected, the conversion is faster than before in this case.

In D20187#441919, @aacid wrote:

@patrickelectric @andreagenor will one of you commti this?

In D20187#441890, @cfeck wrote:

André, see https://secure.phabricator.com/book/phabricator/article/arcanist_new_project/

Why this file was added?
Please describe your intention with these patches.

In D20036#438624, @patrickelectric wrote:

Do you have an estimation of how fast is the code now with openmp ?

• Add OpenMP as a dependency

• Added OpenMP to optimize loops and exchange std::pow to multiply

I know it's just a test and I think irrelevant to land in the master.

I don't think is necessary to revert. I'm working to modernize and clean up the code and this commit solves the problem at this time. In the future, I will create a better solution and maybe complete remove these logs.
I think these logs was made just to debug in some development phase but is not required anymore.

In D19813#432369, @pino wrote:

The commit message is too generic, and it does not actually say what was the problem.
Also, the patch is completely unrelated to what the commit message (generically) says.

remove conflicts

In D18741#409312, @broulik wrote:

This breaks the font tool, can no longer change font style or font size

• Remove code conflicts

• Remove the default case and leave transparent one

• Adjust the indentation to tabular style

Yes, it works fine. Just to remove warnings.

• Correct the code indentation