SO DOES THIS MEAN EVERY SINGLE DESKTOP ENVIRONMENT IN #WAYLAND THAT USES #LCMS2 IS BROKEN IN THE SAME WAY

I would suggest #lcms2 to not optimize by default but use a flag to enable optimization, because programs that uses lcms2 doesn't usually have an extra checkbox for "use accurate conversion", and by default you get wrong result (which is most of the time because it involves conversion from sRGB to and from other formats)

Took QUITE SOME TIME I finally found out, that: If you have two color profiles for #lcms2 input and output, _AND_ if both of the color profiles have a gamma curve, _AND_ if one of them is sRGB, then lcms2 will optimize it into pow(2.2) instead of proper sRGB transformation...

So I guess I finally know why I need to do two step transformation to get a correct result :/

I don't understand why some print shops ask for sRGB source documents *without* providing a gamut limiting icc profile so you can see what the hell you are doing.

I get taking in sRGB files for non-vector printing. But without that gamut limiter you're guessing what kind of gamut narrowing will happen so you should still have access to an icc profile.

At least, in my opinion as a self-made expert 😅

#inkscape #printing #cmyk #lcms2 #icc

#Poppler, the #FreeDesktop #PDF library used by #Evince / #Okular / #GNOME #Papers & pretty much every floss PDF reader app on #Linux, has always been slow at searching text. As in, "3 to 5 minutes to search through 5000 pages"…

Beyond the fact that it's single-threaded, there's an issue that has remained mysterious for 11 years (possibly more): "Why is TextOutputDev slow?"

Today I profiled the hell out of it.
Result: it's #LCMS2 ❓
Why is color management involved?! 🤯

https://gitlab.freedesktop.org/poppler/poppler/-/issues/1472