On Thu, 4 Nov 1999, Mike Smith wrote:
Sysctl is faster than kstat once you have performed the name-oid
lookup. There is basically nothing that kstat can do that sysctl can't
do better and faster, apart from lookup-by-name.
Except for dynamic registration right?
No, Peter fixed that
On Thu, 04 Nov 1999 23:05:30 +0100, Jos Backus wrote:
This patch adds a ``-x'' flag to ftpd, which instructs ftpd to obtain
an exclusive lock on files it commits to disk as a result of a store
operation. This way it becomes easy to tell whether a download has
finished, in case the file
So fstat(1) doesn't show you that the file is opened to ftpd?
No, it does indeed show that.
You really have to lock the files to help you with this problem?
It seems a more natural solution to me than grepping for ftpd in fstat's
output regarding the file. Also, I think that approach
On Fri, 05 Nov 1999 11:15:18 +0100, Jos Backus wrote:
So fstat(1) doesn't show you that the file is opened to ftpd?
No, it does indeed show that.
Then use fstat. :-)
It seems a more natural solution to me than grepping for ftpd in fstat's
output regarding the file.
I think you've
On Fri, Nov 05, 1999 at 12:18:21PM +0200, Sheldon Hearn wrote:
Then use fstat. :-)
OK, OK :)
I think you've developed a complex solution to a more simply solved
problem. UNIX offers you lots of little tools for good reason. Adding
functionality to ftpd that is available through other
On Fri, 05 Nov 1999 11:39:29 +0100, Jos Backus wrote:
Scanning the directory for new files, as the aforementioned script does. If
you have more than one script doing this at the same time, both may conclude
that a given file is ``available'' and try to act upon it.
Then it's your _script_
On Fri, Nov 05, 1999 at 12:58:42PM +0200, Sheldon Hearn wrote:
Then it's your _script_ that should do careful locking to avoid tripping
up over itself, surely? :-)
Yeah, in fact it does, it uses lockf ;-p
Let me take a step back. I'm not saying that what you're doing to ftpd
is wrong.
On Fri, 05 Nov 1999 12:09:18 +0100, Jos Backus wrote:
I'm in the anti-bloat camp, and I agree with this sentiment.
What would be more interesting, I think, is investigating the use of
locking by default. One wonders what it'd break, and how we'd work
around it. ;-)
Ciao,
Sheldon.
To
Mike Smith wrote:
On Thu, 4 Nov 1999, Mike Smith wrote:
Sysctl is faster than kstat once you have performed the name-oid
lookup. There is basically nothing that kstat can do that sysctl can't
do better and faster, apart from lookup-by-name.
Except for dynamic registration
On Wed, 3 Nov 1999, John Polstra wrote:
In article [EMAIL PROTECTED],
Christopher Sedore [EMAIL PROTECTED] wrote:
[...snip info on aio stuff...]
I hope to try 1000 descriptors soon.
That's great news! So have you gotten rid of some of these absurdly
low fixed limits?
I'm a Sound Blaster Live! user and will be glad to
be an Alpha tester for a driver if someone or some group is working on porting
the driver from Linux, since it is now opensource.
Greetings,
Tomer.
Jason tells me you have my card which exhibited exactly this symptom with
a pc164 two years ago, though it worked in an x86 box.
Isn't that the the one now without a BIOS? That's the old 1020 PCI card,
which Qlogic made maybe 500? It's the only one I have. Did you want
it back?
-matt
On Fri, 5 Nov 1999, Tomer Weller wrote:
I'm a Sound Blaster Live! user and will be glad to be an Alpha tester for a driver
if someone or some group is working on porting the driver from Linux, since it is now
opensource.
Greetings, Tomer.
Maybe let people know at
13 matches
Mail list logo