Thank you so much for your help! Unfortunately, it didn't work. I'm running FM v7.2, and it "claims" to support unicode, but it now seems evident that it does not support entering unicode as a bullet in an autonumber format. Please correct me if I'm wrong.
My goal was to use the right-pointing triangle as a bullet. That character can be found in Times New Roman as unicode 25BA.
Instead, I decided to use the right-pointing triangle found in the ZapfDingbat BT font. Its hex code is \xe4.
Hex code can be used as a bullet in an autonumber format in v7.2.
Why don't I just upgrade to the latest version of FM? Well, v7.2 has always met my needs.
> \uNNNN, where NNNN is the 4-digit unicode value.
4 hexadecimal digit, I presume.
That would cover the first 65536 code points in Unicode, but Unicode is an open-ended encoding, and is already up to 110,181 glyphs.
In the FM versions that support \u notation in dialogs (which might not include 7.2), is there any provision for denoting glyphs with code points of over 16 bits?
As for failure mode, I don't understand what you mean by that. However, I can supply a symptom when the unicode is entered.
When I type \u25ba in the autonumber format textbox, while using a character format based on Times New Roman, the string "\u25ba" appears as the bullet. No symbol or image appears on the page, only the string \u25ba. So that makes me think unicode is not supported as an autonumber format in v7.2.
Do you have any thoughts about that?
> ... the string "\u25ba" appears as the bullet.
That pretty clearly implies that FM7.2 does not support \u unicode notation in at least that dialog.
You might also try just pasting the actual Unicode character from a character map applet, or try UTF-8 notation (what FM uses in MIF, I think it's \e##).
The "code points over 16 bits" question was addressed to Arnis, who has more experience with FM Unicode behavior than I do. The \u#### notation only supports a code point range of \u0000 to \uFFFF, or 0 to 65535. Unicode is an extensible standard that supports vastly more than 64K glyphs (it's already double that today). Denoting the high-order code points would require using another notation with a scope larger than 16 bits.