RE: command-line editing speed

RE: command-line editing speed

[Norman.Azadian]

You're right, cranking up BAUD from 300 to 230400 does improve
responsiveness significantly.  Thanks for the tip.

NHA
---
Norman.Azadian@Swisscom.com	+41 31 342 8129
MC-TO-MIT-SWM-TSW			Swisscom Mobile AG

-----Original Message-----
From: Bart Schaefer [mailto:schaefer@brasslantern.com] 
Sent: Freitag, 9. Mai 2003 20:24
To: zsh-users@sunsite.dk
Subject: Re: command-line editing speed


On May 9, 10:02am, <Norman.Azadian@swisscom.com> wrote:
}
} problem is that, even on a machine with load average 0.02, when I lean
} on the h or l key to move the cursor quickly, nothing happens for a }
while and then it jumps 20 or 30 characters.

That sounds exactly like the $BAUD setting issue that I also described.

Re: command-line editing speed

[Bart Schaefer]

On May 9, 10:02am, <Norman.Azadian@swisscom.com> wrote:
}
} problem is that, even on a machine with load average 0.02, when I lean
} on the h or l key to move the cursor quickly, nothing happens for a
} while and then it jumps 20 or 30 characters.

That sounds exactly like the $BAUD setting issue that I also described.

RE: command-line editing speed

[Norman.Azadian]

I don't think the problem has anything to do with the repeat rate.
Testing with ksh I always get the performance that I expect.  The
problem is that, even on a machine with load average 0.02, when I lean
on the h or l key to move the cursor quickly, nothing happens for a
while and then it jumps 20 or 30 characters.

NHA
---
Norman.Azadian@Swisscom.com	+41 31 342 8129
MC-TO-MIT-SWM-TSW			Swisscom Mobile AG

-----Original Message-----
From: Bart Schaefer [mailto:schaefer@brasslantern.com] 
Sent: Donnerstag, 8. Mai 2003 20:46
To: zsh-users@sunsite.dk
Subject: Re: command-line editing speed


On May 8, 12:11pm, <Norman.Azadian@swisscom.com> wrote:
} 
} I'm running version 3.1.5 ($-=3567HRXZhiklms)  on a relatively
powerful } HP machine.  Recently we've been noticing that command-line
editing } seems to have become much slower.

For quite some time now I've been noticing that my keyboard "driver"
loses its key-repeat-rate setting.  That is, have to use the console
"kbdrate" command to restore it, it has nothing to do with the X11 key
repeat setting.  This doesn't happen in any regular way that I can track
down, and I sometimes go for weeks without it happening at all.

However, your problem is more likely related to something else I've
noticed:  ioctl()/stty doesn't always report the right simulated "baud
rate" for X terminal emulators.  ZLE adjusts its display behavior based
on the reported baud rate.  I've had to put code into my .zshrc to force
the BAUD variable to have a reasonable setting regardless of what stty
claimed it to be.

Re: command-line editing speed

[Bart Schaefer]

On May 8, 12:11pm, <Norman.Azadian@swisscom.com> wrote:
} 
} I'm running version 3.1.5 ($-=3567HRXZhiklms)  on a relatively powerful
} HP machine.  Recently we've been noticing that command-line editing
} seems to have become much slower.

For quite some time now I've been noticing that my keyboard "driver"
loses its key-repeat-rate setting.  That is, have to use the console
"kbdrate" command to restore it, it has nothing to do with the X11 key
repeat setting.  This doesn't happen in any regular way that I can track
down, and I sometimes go for weeks without it happening at all.

However, your problem is more likely related to something else I've
noticed:  ioctl()/stty doesn't always report the right simulated "baud
rate" for X terminal emulators.  ZLE adjusts its display behavior based
on the reported baud rate.  I've had to put code into my .zshrc to force
the BAUD variable to have a reasonable setting regardless of what stty
claimed it to be.

command-line editing speed

[Norman.Azadian]

[-- Attachment #1.1: Type: text/plain, Size: 878 bytes --]

I'm running version 3.1.5 ($-=3567HRXZhiklms)  on a relatively powerful
HP machine.  Recently we've been noticing that command-line editing
seems to have become much slower.  Obviously this isn't due to changing
the zsh.  Curiously, it seems to be worse on our new machine, which is
faster and has much fewer processes running and much lower load average
(<0.5).  On both machines, it is now considerably slower than ksh, and I
think this did not use to be the case.  To test this, I simply use the
automatic repeat to key in about 60 chars, then hit escape and try going
back and forth with the automatic repeat of h and l.

Any clues how it could have gotten slower?  Why is the new machine
slower than the old?  Would a newer version of ZSH be faster or slower?

NHA
---
Norman.Azadian@Swisscom.com	+41 31 342 8129
MC-TO-MIT-SWM-TSW			Swisscom Mobile AG