Could a few people look at how the Arabic unicode is rendered in View tables?
It is not clear if the baseline being wrong is new or has been there… but unnoticed. And if it is only Fedora or also occurs on other OSes. They seem correct in the Person Editor dialog and titlebar but not in the Person Views.
Rendering was the same after uninstalling Themes addon and restarting Gramps.
After noticing that the glyphs seemed to be cropped at the bottom, did a search for how the baseline is determined. From “Readability in Arabic-Persian Scripts”, the Baseline section had some outlier glyphs with extreme descenders and ascenders.
To have adjacent rows with arabic, these sample glyphs have been pasted as a new person in the example.gramps tree.
It is indeed cropped.
Can you please also add some Indo-Arabic numerals? ( ٠١٢٣٤٥٦٧٨٩ )
Maybe also the Persian numerals (pretty similar except for some digits) - ۰۱۲۳۴۵۶۷۸۹
This is Dark Themed KDE powered by Manjaro with the same data:
Doesn’t seem to be a problem on Windows. I tried with several different fonts, but never got any difference. In fact, the Arabic letters never even changed, so I suspect that they were always being rendered in the same font (Gtk font substitution to cover all the necessary Unicode characters).
I suspect that the font providing the Arabic chars on your system has an issue.
I reloaded the Themes addon and restarted Gramps (the line was still shifted downward and cropped)
The default Themes settings were :
Adwaita, deselected Dark Variant, Cantarell 11pt font, deselected Toolbar Text
The baseline reset properly by simply changing Themes
and switching back to Adwaita did not show the problem again. Weird
I did notice that “Restore to defaults” reset everything but “Toolbar Text”
Is there another way to block Noto Nastaliq Urdu inside Gramps or it’s too complicated?
BTW I’m familiar with this font and the font in the original screenshot is definitely not Noto Nastaliq Urdu, is it only because of the line height or actual usage?
