Mail Archives: djgpp/2001/08/14/00:11:31
> Someone did mention that it could be fixed by getting development
> source, and then recompiling... has this worked? Anyone got it
> going?
We're making progress. It depends on the specific bug, and the
package built with the workarounds.
> Maybe (I know you folks are busy) the DJGPP people could provide us
> with a fix. Even if its a temporary one, until the next release
> fixes it better.
Well, it would be nice, but we must *have* a fix first, and we're
actively working on finding a fix that works (1) correctly and (2) on
all known platforms djgpp supports. A number of patches have been
added to the CVS respository, but it's dangerous to make a new release
until we test it enough, and for most of the bugs we'd have to rebuild
many packages also (gcc, rhide, etc). It's not a simple thing. The
only other thing we could provide is a source patch, and you can get
those from CVS or the mail archives already.
> The fact is, Win9x is dead now, it will be Win XP next time around.
> Win2000 is at SP2 and has been around for 12-18months now. Maybe
> its time to support it? Maybe its time to stop complaining about
> "bugs" in Microsoft stuff, and just provide a godamn workaround.
> Complaining about the fact that its Microsofts "fault" because its
> their bug doesnt fix it.
Please be a little more understanding and less sarcastic. These bugs
are subtle and nasty and we *are* working on them. They *are* bugs[*]
in MS's software, and we're putting a lot of effort into working
around them. It's not easy. Charles Sandmann has even come out of
"retirement" to help us with some of them.
Remember that MS does *NOT* want legacy dos programs to live. They'd
prefer non-MS programs die too. We get no support from them, and each
release has even worse support for DOS apps. Eventually they'll drop
DOS support completely, and there will be nothing we will be able to
do at that point to support DJGPP on it.
[*] Well, I guess if you never document anything, nothing can
technically be a "bug". But each release acts differently in
subtle ways, and, to me, inconsistency is itself a bug in a public
(albeit unpublished) API.
- Raw text -