Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!uunet!auspex!guy
From: guy@auspex.auspex.com (Guy Harris)
Newsgroups: comp.unix.wizards
Subject: Re: New (GNU) kernels--what I think
Keywords: UNIX progress; controversy
Message-ID: <1806@auspex.auspex.com>
Date: 10 Jun 89 20:00:06 GMT
References: <2501@gandalf.UUCP> <1724@auspex.auspex.com> <2513@gandalf.UUCP>
Reply-To: guy@auspex.auspex.com (Guy Harris)
Organization: Auspex Systems, Santa Clara
Lines: 26


 >> What do you mean "by name"?  Does "name" mean the name of the open file,
 >> or the name of the "who" (which means a process) that has the file open?
 >> The former isn't quite as easy as you seem to think....
 >I'm talking about "by filename" (relative to the current host).
 >Given the current structure of the UNIX kernel, and filesystem code,
 >  it is neither easy, nor cheap.  But the fact is that other systems do
 >  this sort of thing, apparently cheaply.

"Apparently" cheaply, or truly cheaply?  There is extra goop to carry
around; I don't know how worthwhile it is to carry it all around.

>What I was trying to convey was "whatever seems reasonable within the
>  context of the current national character set".

"Current" national character set?  I'd rather not have the file system
code know what the current national character set is, and leave that up
to some higher-level software.

>That approach is only reasonable if the underlying driver allows
>  character-at-a-time processing at user (library??) level to be
>  done cheaply. Traditionally, this has been a problem, which
>  is why there is user-interface stuff in the tty driver.

Well, maybe.  Try snarfing Kenneth Almquist's "atty" program from
"comp.sources.unix" and see whether it's a serious problem or not.