Re: Some small problems I found
On Sunday, 28 Sep, Nikhil Nair wrote:
>
> Also, I didn't follow the proper installation procedure (i.e. using
> dinstall), but did a lot of it by hand. That may account for some of the
> problems.
>
I think it accounts for the absence of much more problems you'd
have encountered should you go with dinstall...
> - pc164 kernel is broken. Pasi, please either remove it from your ftp
> site, or put a README about it. It locked up when I tried to configure my
> network device (using `ifconfig eth0 ...'), while it did actually
> recognise the device (a DE450 using the de4x5 driver). So I couldn't say
> where in the kernel the problem was.
We must rebuild all those kernels anyway.
>
> - Some packages were obviously playing up. man-db 2.3.10-39 segfaults.
I'm running man-db 2.3.10-35 and it's OK (except for some rare unaligned
traps). Either the problem was introduced between 35 and 39, or
segfaults are coming from somewhere else. Which executable is causing
problems?
> Makedev is broken, but I happen to have an old MAKEDEV shell-script - it
> was from my old Debian system, it may be 2 or more years old (presumably
> they switched to a binary since then). If anyone wants to see this, let
> me know (or maybe better to ask the maintainer).
Good point. I never understood why they rewrote MAKEDEV. Actually,
there are 3 questions I wish I had answers to:
- What was wrong with the shell script?
- Was it really necessary to employ a LALR(1) parser to parse, hmm,
/proc/devices?
- Even then, why should it be a proprietary parser, what's wrong with
yacc?
(I should have asked maintainer or upstream author, though.)
I think the old MAKEDEV script is exactly what we need, if it works OK.
I wouldn't say current makedev is really broken, after all it works, but I
doubt that it is feasible to fix unaligned access (which is, I suppose,
caused by that bloody proprietary parser).
>
> Sorry, that's all I can remember for the moment ...
>
Thanks for the bug reports. By the way, we have a nice alpha-port
bug tracking page at beezer.
Cheers,
Nikita
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-alpha-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: