[BusyBox] insmod on old kernels

[Tom Cameron]

Vladimir and Erik,
	I am using BusyBox in a Firewalling system that I sell to customers
(the software is free, but the service of me setting up the system and
maintaining it is what I charge for).  I _need_ BusyBox and uClibc to work
quickly, reliably, and _constantly_.  Frankly, if something is applied to
the CVS version of either project, and there's no explanation telling me
that it makes something more stable, more efficient, or that it in some way
makes the project better...I will _not_ use it.  I will wait until I hear
that it has been tested and that people like it.  I completely understand
where Erik is coming from.  You can't just change a feature "because".  You
should have justification for doing so.  Also, when someone disagrees with
you, or questions your ideas/findings, don't get insulted by it.  We're all
geeks here, and we all have different ways of doing the same thing...that's
just the nature of the game.  I think that Vlad has contributed some much
desired functionality, and I look forward to what else he may come up with,
but if there's no reasoning behind it, I simply won't use it...I _can't_ use
it...there's too much risk at this point.

--
Thomas Cameron
Network Technician / Operations Specialist
St. Mary's Bank

> -----Original Message-----
> From:	Erik Andersen [SMTP:andersen at lineo.com]
> Sent:	Tuesday, February 27, 2001 2:50 PM
> To:	Vladimir N. Oleynik
> Cc:	BusyBox
> Subject:	Re: [BusyBox] insmod on old kernels
> 
> On Tue Feb 27, 2001 at 09:32:54PM +0300, Vladimir N. Oleynik wrote:
> > 
> > You see old version this test.
> > New version have date 8 Feb.
> > >From my ftp log:
> > Fri Feb 23 13:15:30 2001 1 slag.lineo.com 
> > /home/ftp/pub/my/bb/new/bsearch_tst.c.gz
> > Is not you?
> 
> slag.lineo.com is my work box, not Mark's.
> 
> I only had a few minutes that day to look at your patch.  I read over your
> test, ran it, and was impressed with the results.  When I tried to apply
> this
> code to busybox to test how will it would work there, it did not find the
> correct applet.  See below.
> 
> > This version not have error in bsearch2, and very addition faster.
> 
> Here is your find_applet_by_name3 linear search adjusted for busybox 
> with debugging code added.  Try pasting this into busybox utility.c 
> and run it.  Is it fast?  yes.  Does it work?  No. 
> 
> struct BB_applet *find_applet_by_name(const char *name)
> {
>     struct BB_applet *applet;
> 
>     for(applet = applets; applet < &applets[NUM_APPLETS]; applet++)
>         if(*applet->name == *name && strcmp(applet->name+1, name+1)) {
>             fprintf(stderr, "matched '%s'\n", applet->name);
>             return applet;
>         }
>     return 0;
> }
> 
>     [andersen at slag busybox]$ ./busybox pwd
>     matched 'basename'
>     pwd
>     [andersen at slag busybox]$ ./busybox grep
>     matched 'basename'
>     grep
>     [andersen at slag busybox]$ ./busybox this_does_not_work
>     matched 'basename'
>     this_does_not_work
> 
> I got this far and decided that I did not trust your test results since
> the
> functions you were testing do not work properly.  Then I got busy and
> forgot
> about it.  BTW, here is a linear search that should actually work:
> 
>     struct BB_applet *find_applet_by_name(const char *name)
>     {
> 	for(applet_using = (struct BB_applet *)applets; 
> 			    applet_using < &applets[NUM_APPLETS];
> 		applet_using++) {
> 	    if(strcmp(applet_using->name, name)==0) {
> 		return applet_using;
> 	    }
> 	}
> 	return 0;
>     }
> 
> 
> > Its bad mantaining!
> 
> I'm willing to look at further test results and consider patches.  I am
> not
> willing to apply code just because _you_ think it is good.  If you want
> code
> applied to busybox, or any other open source project, you must play by the
> same
> rules everyone else lives by.  Be polite, post simple patches, and be
> prepared
> to defend your patches or repost them.
> 
>  -Erik
> 
> --
> Erik B. Andersen   email:  andersen at lineo.com
> --This message was written using 73% post-consumer electrons--
> 
> 
> _______________________________________________
> busybox mailing list
> busybox at busybox.net
> http://busybox.net/mailman/listinfo/busybox