Date: Sun, 27 Jun 1999 12:09:28 -0700 From: Doug <Doug@gorean.org> To: Randall Hopper <aa8vb@ipass.net> Cc: stable@freebsd.org Subject: Re: 3.2-RELEASE Feedback Message-ID: <377676E8.360985EA@gorean.org> References: <19990626184311.A5290@ipass.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Thank you for your very detailed report. We definitely need feedback like
this.
Randall Hopper wrote:
> 2) Where's the tcpdump port/package -- missing?
Hmmm.. you want to use a tcpdump other than the one that comes with the
system in /usr/sbin/tcpdump? I don't remember ever seeing a tcpdump port,
but that's not really my area of expertise, so I might have missed it.
> 3) If I switch the console to 43 or 60 line mode with vidcontrol, "man" and
> other full-screen progs don't automatically pick up the correct number
> of lines like they did in 3.0-RELEASE. They assume 25-line mode.
>
> vidcontrol < /dev/ttyv0 -f 8x8 cp866-8x8
> vidcontrol < /dev/ttyv0 132x43
> vidcontrol < /dev/ttyv1 132x43
> vidcontrol < /dev/ttyv2 132x43
>
> I have to fiddle with "stty rows" and hand the console the right number.
This is usually a function of your pager, and/or your shell. If you use
bash, try setting 'shopt -s checkwinsize' and see if that helps. If you are
still using 'more' as the default pager, try installing less from the ports
and then do, 'export PAGER=/usr/local/bin/less' (or the appropriate *csh
setenv command) and see if that helps. If neither do, you should send in a
PR for this since (TMK) you shouldn't have to jump through those hoops. Do
'man send-pr' for more details if needed.
> 7) wcd0 became acd0 in the kernel config file, but we still mount wcd0c. (?)
> Is this an oversight? It's in the default fstab for example.
More of a compatability issue. We like to rename things for no good
reason, then keep the old names lying about to confuse people. Welcome to
FreeBSD. :)
> Also, the kernel config and build doesn't fail with the wcd0 device in
> it; but the CDROM doesn't work after booting on that kernel.
Hmmm.. are you sure that you have included the controller that the CD is
on in your kernel config? These are the wdc[01] devices. Also, do you have
the ATAPI options in there properly?
> 8) (Nit) Boot-up probes spelling error:
> "bt_isa_probe: Probe failled for card at 0x330"
This is something you should definitely send a PR for. It will get fixed
10 times faster than just pointing it out here. Also, if you supply a patch
(thereby indicating that you have found the exact location and have saved
people looking at the PR that much more work) it will almost certainly be
committed. If you need a hint on how to find it, grep is your friend.
> 9) rc.firewall - if firewall_type is a file, it feeds each line in it to
> ipfw.
>
> BUT ipfw doesn't toss comments, so you can't document your firewall
> commands, and this IMO isn't a good option.
This has been kicked around many times, and each time there is a "good"
reason not to "fix" this. So, the consensus is, if you want to do it that
way, go right ahead. I don't agree with the consensus, but there are some
things around here not worth beating your head against.
> 10) There was no ports_update package in the packages dir when I downloaded.
> I found a reference to this in DejaNews, and grabbed the /usr/ports/Mk
> tree from ftp.freebsd.org. It worked, but what should users really be
> doing?
What has been happening of late is that in order to follow the ports you
kind of had to follow -stable. Some of the recent changes, like including
the mk stuff in the /usr/ports tree have helped this situation. I think
that after the next release it will be even easier because it seems like
the rest of the bogons have shaken out of this, but for now what users
"should" do is kind of up in the air. I personally encourage people not to
update ports unless they also update the system at least to a recent
version of -stable.
> 11) When copying directory trees between two IDE drives on the PCI
> bus, the system really dogs down -- emacs takes 15 seconds to come up
> when normally it takes 1 -- didn't do that on 3.0-RELEASE.
I would raise this issue in a seperate e-mail. It sounds to me like
something worthy of attention.
> 13) 3.2-RELEASE still locks the system hard when a file system is mismounted,
> just as 3.0-RELEASE did:
>
> # mount /dev/wd4s1 /mnt
> - fails with incorrect superblock (it's a FAT partition)
> # mount -t msdos /dev/wd4s1 /mnt
> - locks the system hard
Also something that deserves a seperate e-mail. Is it possible for you to
add DDB to your kernel and then drop into the debugger when this happens
and do a backtrace? That would probably be very helpful to the developers
of the disk system. I'd say that you should raise 11) and 13) on the
-hackers list.
> 14) I repartitioned and labeled a slice on wd4, and each time Fdisk
> succeeded and Label failed the first time ("Invalid Argument"). Then,
> after I exited sysinstall and restarted it, Label and Newfs worked --
> strange. I repeated this procedure twice -- same results.
The sysinstall interface to those programs is notoriously tempermental.
The command line versions are rather difficult to learn, but more
consistent in their behavior.
> 15) /opt/bin/xanim.2.70.1 +Ze +B +CF4 +Av10 -geometry +100+100 fox.mov
> ld.so failed: Can't find shared library "libgnumalloc.so.2.0"
>
> Something missing from the compat libs. I pulled it off my 3.0-R
> partition on another disk.
I believe this is fixed in -stable, I have that file and I didn't do
anything special to get it.
Once again, thank you for your detailed feedback. I hope the above is
helpful.
Doug
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?377676E8.360985EA>