@UlrikeFischer I checked the current font loader, and the bug is that it uses the last palette color for layers using paletteIndex 0xFFFF (special value that means use text color). So layers that should be black become blue. Your screenshot shows the same bug, these orange glyphs should be black.
@DavidCarlisle @UlrikeFischer Dear friends, I have just made two reports from the user who continues to make misplaced appreciations. This user has the same characteristics as another who wrote the same way and I think has been banned by the site.
@UlrikeFischer The orange ones are layers of larger glyphs, only parts of these glyphs (the dots) need to be colored. The one with no dots are not assigned any color layers at all as it is pointless.
@samcarter Oh, good to know. Then this is not a spam email (I got the same), and I am going to protect our cheese from the mice such that the cutting board has a purpose. ;-)
@UlrikeFischer You can see the expected behavior in this page as well amirifont.org/fatiha-colored.html (should work with latest Edge, Firefox or Chrome browsers). In case you needed more proof about my (and HarfBuzz) interpretation of the spec.
@KhaledHosny I do believe you. In case you saw the bug report in context: I mentioned your name not because I'm sceptic but because I thought Hans would trust your claim more than mine regarding such internals and regarding arabic fonts.
@UlrikeFischer hmm I wonder if that;s how github manages the difference between "member" and "owner" rights (you can have owner rights) what were you changing?
@UlrikeFischer I hope to make a change to kvoptions to use "raw" option lists but then we should probably make a new release (at weekend?), do you have more "pending" changes?
@UlrikeFischer I've had no time to look at the kernel changes (and i want to change it to work better with existing koma code) so it's going to be weekend at earliest but I'll try to give you some advance warning if I'm ready to ship to ctan
@UlrikeFischer OK
@UlrikeFischer luckily I have a test suite in the oberdiek package, i've merged locally and running it now:-)
@DavidCarlisle it looks like a luacolor bug. In the pdf one can see that the color at the end of the text is not reset to black. The black color in the box is inside a q/Q pair.
BT
/F15 9.96264 Tf 1 0 0 1 148.712 657.235 Tm [<001C0023>-28<002B>]TJ
1 0 0 rg 1 0 0 RG
1 0 0 1 148.712 645.28 Tm [<0073007300730073>]TJ
<-------- 0 g 0 G missing
ET
1 0 0 1 148.712 605.643 cm
q <---------box
5.58867 0 0 5.58867 0 0 cm
0 g 0 G
@JosephWright I'm trying to use the RSC package to put DOI links into a paper I'm writing up. Just until I'm ready to send it in, so that I can click links to recheck things easily. However, it isn't putting the DOIs in, even with \usepackage[usedoi=true,linkdoi=true]{rsc} set at the start. Does the package not work anymore or something?