Re: ircII & more fixes.

David Mosberger-Tang (davidm@azstarnet.com)
Mon, 26 Feb 1996 08:18:27 -0700


>>>>> On Mon, 26 Feb 96 11:19:54 +0100, "Fred Roy" <froy@gr.osf.org> said:

>>>>> On Sun, 25 Feb 1996 05:08:56 +0100 (MET), Mikael Nykvist
>>>>> <viper@ludd.luth.se> says:

Fred> Mikael >The more patch is just a "quickhack" which
Mikael> replaces termio->termios, TCGETA->TCGETS,TCSETA->TCSETS...

Mikael> Howdy! As I promised one week ago, I should clean up my
Mikael> patches and make 'em availble.. Ofcourse, maybe the best
Mikael> solution still is a kernel solution...

>>>>> On the same day, David Mosberger-Tang <davidm@AZStarNet.com> replies:

David> No, since there is nothing wrong with the kernel, I don't
David> think it will change w.r.t. termio handling. Best thing to
David> do is probably to use termios and termios only. (See my
David> earlier mail and Dave Miller's follow up as to why the kernel
David> isn't broken.)

Fred> Sorry to insist David, there is probably no need for kernel
Fred> changes, but perhaps some in system header files: if the
Fred> modern (e.g. RedHat) Linux/Alpha <sys/ioctl.h> still includes
Fred> <termios.h>, it does not fit with a source which includes both
Fred> <sys/ioctl.h> and <termio.h> ...

Oh, yes, include file changes will happen. That's what I'm working on
right now in GNU libc. I'm not planning on making any changes to the
existing 4.6.27-based libc (unless there is a compelling reason to do
so before GNU libc is ready). There is basically one include file
that still needs some work but after that, a first release of GNU libc
for the Alpha should be in the mail shortly. The termio vs. termios
problem due to sys/ioctl.h should be gone then. There will be many
other improvements as well (not the least a sqrt() that performs
decently and profiling that *really* works and ...).

--david