I understand Glyphs can generate a new diacritic letter by linking the base letter and the diacritical mark.
In Unicode there are two versions of each diacritical mark, a standalone and a combining version. Ex. the circumflex mark has the 0x02C6 (standalone) and 0x302 (combining) versions, and the grave mark has the 0x060 (standalone) and 0x0300 (combining) versions, which copy Glyphs links to when generating the diacritic letter? Which way is the better practice?
I noticed most of the fonts contain only the standalone copies of the diacritical marks and the combining versions are missing. Is there any use for the standalone versions other than the dynamic linking to diacritic letters?
Why exporting to OTF removes the component linking and dissolves the glyph into outlines, whereas exporting to TTF maintains the linking?
Standalone diaciritics is (unfortunately) very important. If you want to type some accented letters in Czech, Slovak and other languages, you first type the accent (for instance ˇ) and then the letter (n) – you will get ň. And as soon as you click ˇ (Czech or Slovak users have caron on their standard keyboard), ˇ appears in the cursor position. Omitting noncombining caron would be very confusing for users.
I recommend doing the combining ones first, then just generate the non-combining marks, they will automatically be created as compounds.
And regarding your second question: PS-based OTFs are supposed to be all outlines, they use a different technique called subroutinisation. Components are a TT technology. Why do you ask?