Messages in this thread | | | Date | Sat, 19 Sep 1998 23:14:21 -0400 (EDT) | From | <> | Subject | Re: AVL and hash in memory management |
| |
On Sun, 20 Sep 1998, Chris Wedgwood wrote:
> On Sat, Sep 19, 1998 at 10:49:44PM -0400, sfrost@ns.snowman.net wrote: > > > Shell scripts... I'm a sys-admin for a living and write all kinds > > of shell scripts to do things on alot of different platforms (ie: > > Solaris 2.x/SunOS/Linux/others) and it often bugs me how long it > > takes a shell script to run under linux because of the lateny when > > starting sub-shells from the script. > > Linux fork/exec latency is already pretty damn good compared to > solaris, I don't think its every been much slower, if at all.
It's not, and I wasn't saying it was, I suppose I should have made that point more obvious, starting sub-shells under any OS I've come accross so far is slow, would love to see Linux's run alot faster than Solaris's though, would give me another thing to tell management about how Linux is better... :)
> I find shell scripts are slow sometimes - usually because linux uses > bash for /bin/sh where as solaris uses something more lightweight. > (Nice project for someone looking for something - de bloat bash)
That is defintely something else to look at, also I've seen alot of shell scripts that uses '#!/bin/sh', but are really for bash and won't run under the 'sh' found on other systems, would be better if they did '#!/bin/bash', or that 'sh' not just be 'bash', but that's all something for the distributors really...
> > I'm sure there are other applications, just this is something that > > is VERY obvious in my every-day work... > > Oh, as DaveM pointed out, fork/exit isn't uncommon, an he uses the > example of GNU configure scripts. > > When I asked this, I did think about configure scripts and consider > them `not and everyday thing', likewise of interactive fork/exit in > shells.
Hmmm, configure scripts are just about an everyday thing to me, as the admin I get to run around upgrading all the software, after testing it and making sure it doesn't break anything else and other things.
> But I have to concede, that I consider everyday things, and what > other people consider everyday things are not the same, and since > DaveM claims it make a noticeable different when building glibc, it > hard to argue its a good path not worth some optimization then.
I agree, I support a large user base of unix-illiterate people working on unix machines, and what they consider everyday things are very different from what I consider everyday things, even though they also do alot of work on the unix systems, though often they're running shell scripts of programs myself or others in my group have written...
Stephen
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |