Date: Tue, 11 Jul 1995 11:04:23 +0200 (MET DST) From: Thomas Graichen <graichen@sirius.physik.fu-berlin.de> To: bugs@freebsd.org Subject: bugs & problems in 2.0.5 Message-ID: <9507110904.AA21396@sirius.physik.fu-berlin.de>
next in thread | raw e-mail | index | archive | help
hello
i've send the following points to hackers but nobody responsed - i think it
was the wrong list for that - these are the things i've noticed to be wrong or
something like in 2.0.5 - please read them to change it to 2.1 - thanks in
advance
-----------------------------------------------------------------
* file don't know much about FreeBSD - see the next lines
graichen@mordillo:~> bc
bc 1.02 (Mar 3, 92) Copyright (C) 1991, 1992 Free Software Foundation, Inc.
This is free software with ABSOLUTELY NO WARRANTY.
For details type `warranty'.
^\Quit (core dumped)
graichen@mordillo:~> file bc.core
bc.core: data
graichen@mordillo:~> cat test.c
void main(void){}
graichen@mordillo:~> gcc -c test.c
graichen@mordillo:~> file test.o
test.o: NetBSD/i386 object file not stripped
graichen@mordillo:~>
i think there should change something until 2.1
-----------------------------------------------------------------
* the ld still has some problems:
graichen@mordillo:~> cat test.c
void main(void){}
graichen@mordillo:~> cc test.c -L/usr/X11R6/lib -lX11
graichen@mordillo:~> ldd a.out
a.out:
-lX11.6 => /usr/X11R6/lib/libX11.so.6.0 (0x801a000)
-lc.2 => /usr/lib/libc.so.2.1 (0x8093000)
graichen@mordillo:~>
ok the problem is mentioned in the manpage
All shared objects presented to ld are marked for run-time loading in the
output file, even if no symbols are needed from them.
but it think it's a real stupid problem because many software writes
more libs via "-l" to it's commandline for cc than needed - normally
it's not a problem - but under FreeBSD you add much more dependencies
(need for special shared libs which may really not be needed) to the
compiled file than nessecary - is it so hard to fix ?
-----------------------------------------------------------------
* how about a newer f2c ?
graichen@mordillo:~> f2c < /dev/null
/* -- translated by f2c (version 19931217).
You must link the resulting object file with the libraries:
-lf2c -lm (in that order)
*/
#include "f2c.h"
i think until now there were many bug eliminated in f2c and his libs -
why not getting a fresh copy from ftp.netlib.org and adapt it to
FreeBSD (don't forget to adapt it to the f77 driver) - if you like i
may do this - therefor anybody with commit priveleges should send me a
mail and i'll do it sometime in the next 2 weeks
-----------------------------------------------------------------
* why not taking tha man-pages or info files about sysv shared mem
etc. from Linux and adapt them to FreeBSD - in the early slackware
there was an info file about it - but i think it's still in there and
as far as i've seen - in the latest man-pages (1.6 i think) are some
man pages about this too - can any sysvipc guru look into it and adapt
them to FreeBSD ?
-----------------------------------------------------------------
* please put the game man-pages and the /etc/dm.conf file into the
game distribution - why they are in the bin part ?
-----------------------------------------------------------------
* how about integrating gated into the FreeBSD basic distribution - i
think there should'nt be lines in netstart like
if [ "X${gated}" = X"YES" -a -r /etc/gated.conf ]; then
echo -n ' gated'; /usr/local/sbin/gated $gatedflags
if there is no gated at the basic system - i think the basic system
should be closed in itself
-----------------------------------------------------------------
* maybe it has changed since 2.0 - i haven't looked into the sources
for 2.0.5 but i think it has'nt changed - savecore should log the
panic info to LOG_INFO not to LOG_ALERT (i mean the line it prints
via syslog before saving the core "savecore: date machine panic due to
..." - you know what i mean ?) - this would make the screen more
readable while saving the core (at the moment savecore starts - prints
some lines of output saves - counts the core to save in kbytes down -
while that the message via syslog comes to the screen then it counts
further at the next line) - another thing is - why is there a newline
after the counting down is finished i think it would be better to overwrite
this line with the next bootup mesages - this value is then always 0 (or i
think it stops displaying at 1024 - maybe) and thus meaningless -> can be
overwritten from the next line
here is a patch against savecore from -current which solves both problems:
.......................................................................
--- savecore.c~ Tue May 30 11:47:00 1995
+++ savecore.c Tue Jul 11 10:56:57 1995
@@ -184,9 +184,9 @@
check_kmem();
if (panicstr)
- syslog(LOG_ALERT, "reboot after panic: %s", panic_mesg);
+ syslog(LOG_INFO, "reboot after panic: %s", panic_mesg);
else
- syslog(LOG_ALERT, "reboot");
+ syslog(LOG_INFO, "reboot");
if ((!get_crashtime() || !check_space()) && !force)
exit(1);
@@ -402,7 +402,6 @@
path, strerror(nw == 0 ? EIO : errno));
err2: syslog(LOG_WARNING,
"WARNING: vmcore may be incomplete");
- (void)printf("\n");
exit(1);
}
}
@@ -651,3 +650,4 @@
(void)syslog(LOG_ERR, "usage: savecore [-cfvz] [-N system] directory");
exit(1);
}
+
.......................................................................
-----------------------------------------------------------------
* some time ago i rewrote the math-manpages to also mention the float
functions (sinf for sin as an example) - how about links from the sin
man-page to sinf - this way you may get the sinf-information also via
man sinf and not only via man sin - "oh there's something about sinf
too"
-----------------------------------------------------------------
* the man-page for yp is total out of date - it says for instance there is no
server etc.
-----------------------------------------------------------------
* is there any reason for the entry with "CCFPU" in /etc/make.conf - i grep'd
through the complete /usr/src/gnu/usr.bin/cc tree and haven't found any ifdef
- and also in the makefile's until there is nothing to find - is this option
still used ?
-----------------------------------------------------------------
the following things i'm not shure about - because i haven't installed
2.0.5 from scratch (i updated from 2.0) - maybe i'm wrong in some points - but
please go on reading:
* ok - i may now read parts of the doc's in german (my native
language) but as far as i know (from testing the alpha-boot floppies)
my keyboard is still handled as a us one (maybe i'm wrong)
-----------------------------------------------------------------
* is /usr/share/man chown'ed right - as i said i install via unpacking
the manpages.* files by hand - and i got bin.bin as owner for this dir
- but catman (... please use "echo /usr/bin/catman | nice -5 ...) runs
su man in the recommended way and can't thus create the
/usr/share/man/cat* entries
-----------------------------------------------------------------
* how is kernel core dumping handled in 2.0.5 (i hope that
i'll never have to do something with it :-) - earlier ther was a
kernel config option DODUMP - this is no longer there (grep through
/usr/src/sys) - and in LINT stands something like
# - Crash dumps will be written to wd0b, if possible. Specifying the
# dump device here is not recommended. Use dumpon(8).
and dumpon(8) says:
Calls to dumpon normally occur in the system multi-user initialization
file /etc/rc, before the savecore(8) program is run.
but ther is no dumpon in /etc/rc - i think this is a bit contradicting
- will it dump by default to my swap-partition - will it by default
not dump ?
-----------------------------------------------------------------
* how about chmod'ing the floppy devices (/dev/*fd?*) g+rw - this way
you may allow some people to use the floppy by adding them to the
group operator (or any other group - floppy for instance if operator
would be to insecure - and changing the floppy devises to group
floppy) ?
-----------------------------------------------------------------
* ok - that's all for now - if something is unclear - send me a mail
t
_______________________________________________________||_____________________
__||
Perfection is reached, not when there is no __|| thomas graichen
longer anything to add, but when there __|| freie universitaet berlin
is no longer anything to take away __|| fachbereich physik
__||
- Antoine de Saint-Exupery - __||
___________________________||____email: graichen@omega.physik.fu-berlin.de____
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9507110904.AA21396>