RELAXNG Compact Syntax and character escapes

["Jeff Rafter" <lists@jeffrafter.com>]

I am working on a RNC tokenizer and have reached a stumbling block. In
section 2.4 of the specification the parser must handle character escape
sequences of the form \x{n}. Easy enough. But the note in the section (and
grammar) specifies that there may be more than one "x" character. In the
grammar this is indefinite-- and this is presumably correct as the
reversible transformation may be compounded upon itself (some of the
confusion may stem from the odd text in the note "...the transformation
replaces adds an extra...").

Obviously however, there are some pathological cases here. I am wondering if
there are real world limits that others are imposing (e.g. max 2 xx).
Additionally, are there limits that other developers are placing on the
length of the hexadecimal string? I have written a chunking parser for the
decoding and am trying to handle the end of buffer cases in a sane fashion.

I tried to see if there were any errata on the subject and found a broken
link.

Thanks,
Jeff Rafter