On Wed, Nov 16, 2005 at 01:29:21AM -0500, Will Coleda wrote:
> All but one test are passing again in tcl. The failing test boils
> down to this PIR:
> [...]
>
> So, are all these ways of preparing arguments to Glob incorrect? (and
> if so, what's the right way?), or does this behavior point to a bug?
It's a bug, now fixed in r10020. When switching to the new escape
opcode I forgot that we sometimes have "\u" sequences in the escaped
string (Data::Escape didn't produce these).
The answer to your other question ("how should arguments be prepared
for Glob") is that PGE is intended to be able to process strings and
patterns of any charset and encoding; one shouldn't need to do any
escaping or processing of arguments in order for them to work.
Pm
