Mike Gran <spk...@yahoo.com> writes:

> Hmmm.  I should have been watching the clock.  I've left master in
> an odd state w.r.t. strings and chars.  The storage is available for
> non-8-bit strings and chars, but, nothing can be done with them.

I don't think it's a serious issue.

Just a small comment:

--8<---------------cut here---------------start------------->8---
scheme@(guile-user)> (%string-dump "foo")
$1 = ((string . "foo") (start . 0) (length . 3) (shared . #f) (stringbuf . 
#<unknown-type (0x30227 . 0x6f6f66) @ 0x7f1c59c99b00>) (stringbuf-chars . 
"foo") (stringbuf-length . 3) (stringbuf-shared . #f) (stringbuf-inline . #t) 
(stringbuf-wide . #f))
--8<---------------cut here---------------end--------------->8---

The stringbuf object cannot be printed (as shown above), has no public
type predicate, etc., so I think we'd better not return it to Scheme
code.  Can you change that in `%string-dump'?

Note for later: it'd be great if we had micro-benchmarks to compare the
former and the new string implementations (on Latin-1 strings, that is),
and to compare Latin-1 and wide strings in the new implementation.

Thanks,
Ludo'.



Reply via email to