Mail Archives: cygwin-apps/2002/01/21/22:22:26
----- Original Message -----
From: "Michael A Chase" <mchase AT ix DOT netcom DOT com>
> > Sure. note that ~io_stream called belongs in setup.log.
>
> It definitely goes there now, but what benefit are we gaining by
having
> low-level progress messages in the permanent log? I thought that was
what
> setup.log.full was for.
If someone other than I adds a new io_stream class, they may (for
whatever reason) not override the destructor. All other functions are
abstract, forcing an implementation, but the destructor cannot be
abstract.
> My original thought was to go through the source files and basically
change
> any log() calls that said 'called' or 'peeked' or similar low-level
progress
> indicators and change them from LOG_TIMESTAMP to either LOG_BABBLE or
> LOG_TIMESTAMP | LOG_BABBLE. If some of those progress messages are
meant to
> be in setup.log, I'll need to know what policy to follow. Just a hint
for
> now and we can finalize the plan when you see some of my patches.
For the peeked/called etc, yes they can be trimmed down, or even
eliminated (ie #ifdef DEBUG) as appropriate. They were my in-development
code debug output, and aren't really needed now :].
> It's going to take me quite a while to get really up to speed, so if
you
> have any repetative work that you haven't found time for, I'd be glad
to
> handle those for you.
Not really. There's a number of small nitpicky things in README that
need doing - new views and the like. There's also the nasty bug that is
affecting corinna, and is most probably heap corruption, but I'm
b****gered if I can figure out where it's occuring.
One useful and fairly trivial hack to get back in touch would be to move
the downloaded files from the site/path to simply path/, to allow
greatest possible hit rate when folk select different mirrors.
Rob
- Raw text -