Identifier may not start with -, so that --ABORT-- may not be mistaken as an identifier. And although we state that HEADERNAME are like identifiers followed by :, the given regex allows them to start with -. This raises the question of whether --ABORT--: is a HEADERNAME, or if is --ABORT-- followed by :.
I suggest we simply fix the HEADERNAME definition to disallow a leading -. This is what is implemented in Spot 1.99a and jhoafparser 1.0.1 already. (But maybe there is a development version of jhoafparser that recognizes -: as a HEADERNAME by now, because this is one case in the test-suite developed by Joachim.)
Yes, I think that allowing - at the start of the regex for HEADERNAME is just a typo. We should simply remove this. It's just a corner case that's tricky to implement in the parser and serves no purpose.
Identifier may not start with
-
, so that--ABORT--
may not be mistaken as an identifier. And although we state that HEADERNAME are like identifiers followed by:
, the given regex allows them to start with-
. This raises the question of whether--ABORT--:
is a HEADERNAME, or if is--ABORT--
followed by:
.I suggest we simply fix the HEADERNAME definition to disallow a leading
-
. This is what is implemented in Spot 1.99a and jhoafparser 1.0.1 already. (But maybe there is a development version of jhoafparser that recognizes-:
as a HEADERNAME by now, because this is one case in the test-suite developed by Joachim.)