>>77
Damn it. There's no ">>" to match in the URL.
This is irrelevant. It is the bug hunt I'm after.
All right, provide any firm statement that names a string that fails this way.
No, this is not an answer. Name the exact quote attempt in this thread that failed with >>65. The only quote attempts after >>65 was posted, that are not fully links, are ">>64-67" and ">>1,10-20,30" from >>67. They both stop being links at the hyphen. Do these fail with >>65 and "irregex compilation in a running instance of SchemeBBS"? Yes or no. This is programming, not religion. I need any firm assertion that I can then verify myself as being true or false.
It's actually quite easy to check in the REPL that the regex in >>65 will fail at the first hyphen, as you noted:
]{0,2})|(([1-9][0-9]{0,2})-([1-9][0-9]{0,2})))){0,11}" "11,20,25,34-80" "fails")
;Value 13: "fails-80"
1 ]=> (irregex-replace "[1-9][0-9]{0,2}(-[1-9][0-9]{0,2})?(,[1-9][0-9]{0,2}(-[1-9][0-9]{0,2})?){0,11}" "11,20,25,34-80" "works")
;Value 14: "works"The bug is in irregex not in your regex. There is an inconsistency between irregex-match and irregex-replace or something I don't get. Anyway, rewriting the regex seems to circumvent it.