There is a bug in generating Persian and Urdu sets. To see it, please make a new empty Glyphs file, then right-click on Persian and generate all 32 glyphs. It would make them but the number of the glyphs would show 29/32. The 3 “Yeh-Farsi” init, medi and fina wouldn’t registered or show in Persian set as generated even though they are generated.
The same thing happens for 2 glyphs of “four-persian.urdu” and “seven-persian.urdu” in Urdu set.
I fixed that yesterday. Will be in the next update. But that is mostly a cosmetic problem with the numbers in the sidebar.
Thank you very much Georg! It is cosmetic but annoying. 
Another one that bothers me a lot: I really appreciate if you could fix this one:
In Arabic, when double clicking on a component to open it beside the active glyph, it would change to medi or init version of the glyph and stick to the previous glyph and most of the time, both glyphs would change. It is so counter-productive. For example, I have inserted “dotless beh” inside the “peh”, now I want to change that dotless beh. I wish we could somehow do it by double-clicking the component inside peh without it changing to other version of the glyph. I suggest something like holding Option while double-clicking the component would open the same instance of the component beside it, instead of the attached version of that. Is it possible?
I’m not sure how to fix this. What I usually do in those cases, is to switch to LTR.
Inserting a “ZERO WIDTH NON-JOINER” (U+200C) between those would do it most of the cases. Thank you for your tip of switching to LTR. Is there a shortcut for switching the direction?
Adding the zero width joiners would be quite annoying.