Color picker Lab readouts depend on the output profile (and are almost always wrong))

[ellelstone]

Version: 5.6
Branch: releases
Commit: 5a4ed73
Commit date: 2019-04-20
Compiler: x86_64-pc-linux-gnu-gcc 7.3.0
Processor: Intel(R)\ Core(TM)\ i7-4790K\ CPU\ @\ 4.00GHz
System: Linux
Bit depth: 64 bits
Gtkmm: V3.24.1
Lensfun: V0.3.2.0
Build type: Gentoo
Build flags: -march=haswell -O2 -pipe -std=c++11 -Werror=unused-label -Wall -Wuninitialized -Wno-deprecated-declarations -Wno-unused-result -fopenmp -Werror=unknown-pragmas -ftree-vectorize
Link flags: -Wl,-O1 -Wl,--as-needed
OpenMP support: yes
MMAP support: ON

Linux gentoo 4.9.34-gentoo #1 SMP Fri Aug 4 18:25:06 EDT 2017 x86_64 Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz GenuineIntel GNU/Linux

Here's the problem:

In RT-5.6 from gentoo portage, and also RT-locallab updated and built last week, the color picker Lab readouts depend on the output profile, producing noticeably variable and wrong results especially when choosing linear TRC output profiles. Eg a reading of around L=14 that really should be around L=45.

For example, after opening an sRGB image (gimp built-in sRGB) the only selected output profile in RT that shows correct results is RTv4_sRGB . RT_sRGB (a V2 profile?) is slightly off, as is using the input profile as the output profile. Similarly with images in any other RGB color space.

Here's an example pp3 and image file - change the output profile in RT and watch the Lab values change when color-picking:
rt-lab-readout-bug.zip

My apologies, I'm not going to supply a screenshot or a video. This bug is easy to reproduce. In case there is any doubt as to what part of the user interface I'm referring to, the Lab readouts are in the leftside panel. The place to change the output profile is in the right side "Color" panel.