What I’m after is getting the glyph that is rendered on the screen. The “ff” appears as two characters in the string (which is normally what you get) but the glyph is only one “character”.
This is particularly nasty when one is rendering non-Latin languages, especially fonts with encoding table problems. Some of them are notorious for rendering fine on Apple’s software (such as pages) and coming out totally wrong in InDesign. The solution usually is either to switch fonts (which solves the problem but changes the font of course), or edit the font’s encoding tables with a font editor (which for some languages, is a lot of work). The thing is the glyphs are all there, so if one can dump the glyphs, that would also do the trick (since they’re all unicode).
Sometimes in desperation, one resorts to opening the glyphs screen and clicking the glyphs, but that is time consuming compared to typing at the keyboard.