Ok, I understand the linear name now. Thanks for that.
It brings up another issue: the 'addition' layer mode in sRGB is very wrong.
'Addition' layer mode adds the pixels values. A test I did is easy to verify:
I fill a layer with 50% gray, then duplicate the layer, and on the top layer use 'addition' layer mode.
In legacy 'addition' layer mode 50%+50%=100%, so the result is white.
But in the new default 'addition' mode the result is not pure white, because the 50% gray is not really 50% of the linear light value.
What Gimp 2.10 does differently than some programs is even in 8bit sRGB it removes the gamma before figuring out the correct value and then re-applying the gamma.
Krita, for example, doesn't even care about linear color space unless the bit depth is greater than 8 bits.
It brings up another issue: the 'addition' layer mode in sRGB is very wrong.
'Addition' layer mode adds the pixels values. A test I did is easy to verify:
I fill a layer with 50% gray, then duplicate the layer, and on the top layer use 'addition' layer mode.
In legacy 'addition' layer mode 50%+50%=100%, so the result is white.
But in the new default 'addition' mode the result is not pure white, because the 50% gray is not really 50% of the linear light value.
What Gimp 2.10 does differently than some programs is even in 8bit sRGB it removes the gamma before figuring out the correct value and then re-applying the gamma.
Krita, for example, doesn't even care about linear color space unless the bit depth is greater than 8 bits.