Robert Pluim writes: >>>>>> On Tue, 21 Jul 2020 08:40:56 -0400, Amin Bandali said: > > Amin> Hi Eli, Pip, > Amin> Thank you for working on this and fixing it on emacs-27 per the reports. > > Amin> I have not been able to keep a close eye on GNU lists and trackers for a > Amin> while now, but I was wondering if there's been any more progress/updates > Amin> on this issue on master? I sometimes read/write in Persian, and this > Amin> bug is quite an annoyance. :-) Changing the text scale (C-x C-=) and/or > Amin> filling the paragraph (M-q) seems to sometimes help, but not always; and > Amin> the issue remains. > > emacs-27 gets merged to master on a regular basis, and the relevant > commit is present in both: > > $ git branch --contains 30a7ee505a > * emacs-27 > + master > > Are you still seeing issues with the latest master? > > Robert > Ah, great point! This got me suspicious, so I went ahead and actually tried both emacs-27 and master in a few different configurations myself. It seems that with emacs-27, Xft rendering of Arabic/Persian text works just fine, like it used to: . However, using Cairo+HarfBuzz, on both emacs-27 and master the issue is still present: . Excerpts from M-x describe-char RET on a Persian character from Xft and Cairo+HarfBuzz are at and respectively, in case that might be useful somehow. The font used for typesetting the Persian text is Vazir v22.1.0, available from . My .Xresources configuration for the Xft and Cairo+HarfBuzz setups, the former only for emacs-27, and the latter for both emacs-27 and master: Emacs.FontBackend: xft,x Emacs.font: Source Code Pro Medium:size=14 Emacs.FontBackend: ftcrhb,x Emacs.font: Source Code Pro Medium:size=14 Lastly, it might be worth mentioning that if I recall correctly, when using xfthb with emacs-27, I observe the same issue. Which may suggest that perhaps the issue is related to Emacs's HarfBuzz support. Hope this is useful.