On 19 Nov 2012, at 22:16 , David Kastrup wrote: "Christopher R. Maden" < address@hidden> writes: On 11/18/2012 05:26 PM, David Kastrup wrote:
At the current point of time, the rule is that "alphabetic" is a-z,
A-Z, and _any_ non-ASCII character. This is a bit excessive, but
short of a reliable "is a letter" test, this was easiest to
implement.
Getting off-topic for -user, but... Most programming languages that
can process Unicode text have class tests that should simplify
letter-ness. (The superscript characters ¹²³ etc. do not have this
property.) Guile has the char-alphabetic? predicate; see <URL:
https://www.gnu.org/software/guile/manual/html_node/Characters.html >,
at least as of 2.0.
We are not using Guile 2.0 yet, and I doubt people would be happy about an undiscussed change once we do.
I perfectly understand the concepts of code points and all the problems associated with it, have been a systems programmer part of my life. And there is a big difference in processing unicode characters as data on which the program does its job and a symbol. But towards what is accepted as "characters" to be to compose a variable name (or macro, or identifier) should be agreed upon. I don't mind either way, but what is stated and documented should in theory be followed.
My suggestions: 1. Document quick-and-dirty that although these characters do seem to work, usage is at your own risk, at least at this moment in the current (and upcoming) versions, and 2. Let's put it on the wishlist to properly support these, test it extensively! and upgrade the doc.s correspondingly.
And indeed avoid this kind of changes at once because when something breaks down people will be unhappy i.m.h.o. -- David Kastrup _______________________________________________ lilypond-user mailing list address@hiddenhttps://lists.gnu.org/mailman/listinfo/lilypond-user
|