[no subject]

[evanlang@uss.lonestar.org (Evan Langlois)]

Lots of questions:

What are the advantages of screen?  Why should I use it?   I know its
a matter of personal preference, but so far I've found myself using most
of Unix stuff that the list has found useful.   Never thought I'd
admit it, but Unix utils are useful :-)

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
Workable Formatter, found on a.a.u.e.) to format manpages. Moreover, the
unformatted page is required for the whatis database.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Is there a way to compress/uncompress this stuff on the fly?  Many
multiple levels deep?   Maybe tar it so all the man pages can be in
one file?   Disk space is limiting.  With DataLite or whatever (not
Data-Diet, the other one that works at the BIOS level) work with
MiNT on a Minix parition .. sort of a Minix-Stacker?

Anyone got a good compression algorythm they could hack into the Minix
driver :-)  And fsck :-)

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
If the masterfd is closed, the process having opened the slavefd does not
terminate, but takes about all CPU cycles it can, slowing down my Atari
considerably.
Shouldn't it be that Mint sends SIGHUP or SIGTERM to the process(es) having
opened the slavefd when the masterfd is closed?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

I thought that the SIGHUP or SIGPIPE or whatever would be sent when
the last byte was read from the queue.  Maybe the changes to the pipefs
that change off-by-one problem have messed that up, or maybe your
app is ignoring a signal or something?

Or maybe its just broke :-)

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
to ask: Is that officially documented behaviour? ;) It seems like every
process supports only one active timeout. Shouldn't this be fixed?
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Are you saying that Fselect's timeout and the Talarm are the same?
What about the other Timeout functions similar to Talarm (forgot the
names); will these share the same timeout as Talarm?  I wouldn't think
they would be the same, so I would guess they should be fixed.

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
No, MiNT doesn't have the braindead SYSV signals. It has BSD signals,
which don't get zapped when the signal has occured.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

I thought the SIGALRM was an exception since once it is expired, there
is no time remaining.  So, the handler doesn't get reset, but the
alarm would have to be reset.  So, I guess that isn't an exception to the
rule, since you don't have to call Psigalarm again, no?

I've hacked minixfs to use the update daemon, and the tsync and the
update daemon seem to call slightly different code (not sure why this
makes an difference) but I no longer have trouble with my Minix 
partition!   Update works, and addroottimeout doesn't, unless there
has been a change since H5/6 or Minix PL 10.  
 
I still have the problem of errors stating a that zone range error has
occured, Drive D Zone 0 Count 1, Bad Filesystem, Repair with fsck.  Only
fsck doesn't even find the problem (annoying but seems to be harmless).

The other problem is that fsck says Pass 3a Fixing Directories
But then it doesn't say what is wrong with my directories, and never
fixes anything since if I run it again, it tries to fix it again.
These problems came about at different times, but may be related
anyway.
 
Any ideas anyone?

Now that I have update back, I'm happy :-)  No more inode bitmap
errors.

CYA,
Evan Langlois