Re: NIO charsets attacked by gang of rabid platypi

[Tom Tromey]

>>>>> "Sven" == Sven de Marothy <address@hidden> writes:

Sven> Ok, now that I hopefully got your attention with the subject line,
Sven> I'd like everyone's opinion on whether it's OK yet to get rid of the
Sven> now-unused gnu.java.io character encoding/decoding stuff?

I'm in favor of removing it.  Unused obsolete code should just go
away... IMO it would have been fine to delete it when you checked in
the patch to switch over.  (In cases where were concerned that we
might want to roll back, we can just make a cvs tag before the big
rearrangement...)

Sven> Also, I'd like to know which NIO charset provider is preferred, now that
Sven> we have two in CVS. (pure java and iconv-based)

I would say, default to java and fall back to iconv.  Otherwise, on
Linux platforms, we're unlikely to see any testing of the java
providers at all (since iconv there tends to be pretty featureful).

Furthermore I think, at least initially, it is fine to have this
simply be hard-coded.  If some VM needs a Configuration.java flag or a
system property, we can add it easily enough on demand.

Tom