On Thursday 14 May 2009 18:51:56 Cathey, Jim wrote:
> >What the heck is a "shared tty device"?
>
> More than one process has active R/W access
> to any given TTY.  Shared.  TTY device.
> Interleaved reads are death,

So you won't ever be running two instances of a shell or vi on the same 
terminal.  (And yes, I've had to diagnose this.  I've encountered the 
problem, there just wasn't a _name_ for it because it was a _bug_.)

> but interleaved 
> writes are not entirely unheard-of.

Sure, the "wall" program can pop up randomly or dmesg output winds up crapping 
over your current terminal, or you backgrounded a copy of firefox without 
redirecting both stdout _and_ stderr to /dev/null and then suddenly it 
decides to get all chatty on you at some later date...  Yes, this will screw 
up your display.  So what?

Under certain circumstances the cursor position can change randomly due to 
some completely unrelated process.  Some unrelated process can also send you 
a "kill" signal.  The kernel can do it in an out of memory situation.  What 
possible _relevance_ could this have?

> >Busybox does not compile or run on Unix System 6.
>
> Does talk quite nicely to my VT52, though.

And the last time you actually did that before writing that message was...?  
(On second thought, I don't care.)

> Maybe not for much longer?

First of all, you can export LINES and COLUMNS before running busybox as a 
trivial workaround, so it had no reason to probe.  Trivial workaround, and 
one I already described in this conversation.  (You can also turn off the 
command line editing config option I said this should be attached to, so it 
doesn't care about cursor up and such and pretty much won't use ansi escape 
sequences at all.)

Secondly, the first google hit for "vt52 escape sequence" on google is this:

  http://ascii-table.com/ansi-escape-sequences-vt-100.php

Which says vt100 actually _does_ support the esc[6n "Get cursor position" 
sequence.

Thirdly, you're talking about a terminals that stopped being manufactured 30 
years ago.  I had to fight to keep Red Hat 9 as an accepted build 
environment, and you're talking about hardware that's 3 times older than the 
busybox project itself.  Had the project's current maintainer even even been 
_born_ yet when that stuff was still being sold?  No, I wouldn't personally 
waste effort on supporting that...

> -- Jim

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox

Reply via email to