strange phenomena on potato
Since I have a working 2.2 kernel (and hence a working potato [hmm, that
sounds strange :-] ), I've been trying a couple of things last night.
- the `man' command seems to sleep about 3-5 seconds before displaying
anything. In fact, `strace' shows that there is an explicit delay of 1
second, implemented by a select() call without any FDs to check for
(this indicates that it's not waiting for anything, it is really an
explicit sleep). My research indicates that this probably happens in
gdbm_close() (in libc?). I don't have the libc sources handy (hi Chris!)
so I couldn't verify this. libc 6.12.1.2-0pre6 and man-db 2.3.10-69i.
- the alignment trap from identd is back (that was fixed, I'm sure).
Ah yes, I had fixed it in 3.07-7.1.slink.1, this is from the changelog
I did then:
* non-maintainer fix reimplemented YET AGAIN (see bug #30366).
Doing a sourceful NMU this time.
* changed the version numbering as otherwise it was considered a
downgrade; there was already a nonmaintainer upload due to the above,
and apparently 3.07-7slink.1 < 3.07-7.1
Looks like time for another sourceful NMU (current version is 3.07-7.2)
- the `file' command sometimes gives "lseek: invalid argument". However,
I can't reproduce this now... `strace' showed that it was passing some
gigantic number as the offset.
Besides these it actually all looks pretty good ;-)
BTW: I haven't been paying that much attention to the build process
lately. I gather there's now an autobuilder? How does one proceed when
wanting to attack the diff list?
Paul Slootman
--
home: paul@wurtel.demon.nl http://www.wurtel.demon.nl/
work: paul@murphy.nl http://www.murphy.nl/
debian: paul@debian.org http://www.debian.org/
isdn4linux: paul@isdn4linux.de http://www.isdn4linux.de/
Reply to: