openbsd
27 years agoRemove old -P flag since it doesn't do anything and make -P mean
millert [Sun, 6 Apr 1997 06:26:23 +0000 (06:26 +0000)]
Remove old -P flag since it doesn't do anything and make -P mean
"preserve leading /" like GNU tar (no longer use -S for this).

27 years agosupport cd9660 disklabel spoofing; also fix block offset confusion
deraadt [Sun, 6 Apr 1997 06:13:03 +0000 (06:13 +0000)]
support cd9660 disklabel spoofing; also fix block offset confusion

27 years agoAllow creation of archives with block sizes up to 64512 bytes
millert [Sun, 6 Apr 1997 06:11:10 +0000 (06:11 +0000)]
Allow creation of archives with block sizes up to 64512 bytes
but complain about non-portability if > 32256 bytes.
Brought up by George Robbins <grr@shandakor.tharsis.com>

27 years agocorrect blocksize confusion
deraadt [Sun, 6 Apr 1997 06:09:25 +0000 (06:09 +0000)]
correct blocksize confusion

27 years agocorrect labelsector confusion
deraadt [Sun, 6 Apr 1997 06:04:26 +0000 (06:04 +0000)]
correct labelsector confusion

27 years agosupport cd9660 disklabel spoofing; also fix block offset confusion
deraadt [Sun, 6 Apr 1997 06:03:44 +0000 (06:03 +0000)]
support cd9660 disklabel spoofing; also fix block offset confusion

27 years agoEradicate #include <sys/*> from machine/*.h. Prompting from Gene.
briggs [Sun, 6 Apr 1997 02:59:52 +0000 (02:59 +0000)]
Eradicate #include <sys/*> from machine/*.h.  Prompting from Gene.

27 years agoAttempt to kludge around the fact that we can not yet get the correct MAC
briggs [Sun, 6 Apr 1997 01:02:13 +0000 (01:02 +0000)]
Attempt to kludge around the fact that we can not yet get the correct MAC
address for the SONIC on PB5xx series.  This uses the bus_space_bad_addr
function that I added earlier.

27 years ago32bit corrections
mickey [Sat, 5 Apr 1997 23:49:07 +0000 (23:49 +0000)]
32bit corrections

27 years agoadd quirk table entry for the Ultima A6000C
kstailey [Sat, 5 Apr 1997 23:34:18 +0000 (23:34 +0000)]
add quirk table entry for the Ultima A6000C

27 years agodevboot is no longer in boot.c
mickey [Sat, 5 Apr 1997 23:27:35 +0000 (23:27 +0000)]
devboot is no longer in boot.c
missed from last commit

27 years agofix NULL ptr deref in ss_set_window()
kstailey [Sat, 5 Apr 1997 22:59:13 +0000 (22:59 +0000)]
fix NULL ptr deref in ss_set_window()

27 years agoFix bug introduced with "exit(0) if no files" behavior change.
millert [Sat, 5 Apr 1997 22:51:53 +0000 (22:51 +0000)]
Fix bug introduced with "exit(0) if no files" behavior change.

27 years agoStrip leading '/' of pathnames (only in tar mode). -S option turns
millert [Sat, 5 Apr 1997 22:36:12 +0000 (22:36 +0000)]
Strip leading '/' of pathnames (only in tar mode).  -S option turns
this off like GNU tar.

27 years agosettimeofday(2) restruction moved to secure level 2.
millert [Sat, 5 Apr 1997 22:06:10 +0000 (22:06 +0000)]
settimeofday(2) restruction moved to secure level 2.

27 years agocorrect DEV_BSIZE vs lp->d_secsize confusion; spotted by hte late night icb gang...
deraadt [Sat, 5 Apr 1997 21:56:02 +0000 (21:56 +0000)]
correct DEV_BSIZE vs lp->d_secsize confusion; spotted by hte late night icb gang. Other ports need fixing still

27 years agordroot entry for ramdisk kernel.
mickey [Sat, 5 Apr 1997 21:48:28 +0000 (21:48 +0000)]
rdroot entry for ramdisk kernel.
import from hp300 port

27 years agonew exec stuff
mickey [Sat, 5 Apr 1997 21:29:49 +0000 (21:29 +0000)]
new exec stuff

27 years agodo not up wiring count for page tables (especially those actually
flipk [Sat, 5 Apr 1997 21:24:48 +0000 (21:24 +0000)]
do not up wiring count for page tables (especially those actually
owned by the kernel pmap). Chuck Cranor (chuck@maria.wustl.edu) and Theo.

27 years agoUse in_addr_t not u_long and use in_port_t in some cases instead
millert [Sat, 5 Apr 1997 21:13:05 +0000 (21:13 +0000)]
Use in_addr_t not u_long and use in_port_t in some cases instead
of u_short or u_int16_t.

27 years agoUse in_addr_t not u_long.
millert [Sat, 5 Apr 1997 20:58:28 +0000 (20:58 +0000)]
Use in_addr_t not u_long.

27 years agoremoved conditional code
kstailey [Sat, 5 Apr 1997 20:48:58 +0000 (20:48 +0000)]
removed conditional code

27 years agosendsig and sys_sigreturn are in m68k area now
kstailey [Sat, 5 Apr 1997 20:44:25 +0000 (20:44 +0000)]
sendsig and sys_sigreturn are in m68k area now

27 years agofpu_type -> fputype
kstailey [Sat, 5 Apr 1997 20:32:49 +0000 (20:32 +0000)]
fpu_type -> fputype

27 years agofpu_type -> fputype
kstailey [Sat, 5 Apr 1997 20:22:01 +0000 (20:22 +0000)]
fpu_type -> fputype

27 years agoif TERM=dumb disable progress and editing for emacs ange-ftp
kstailey [Sat, 5 Apr 1997 19:53:10 +0000 (19:53 +0000)]
if TERM=dumb disable progress and editing for emacs ange-ftp

27 years agoput gdt & idt in the separate file.
mickey [Sat, 5 Apr 1997 18:56:14 +0000 (18:56 +0000)]
put gdt & idt in the separate file.
print trapno in dump_regs.
no more NO_IDTR.

27 years agoImplement a bus_space_bad_addr() function to aid in some device probing with
briggs [Sat, 5 Apr 1997 16:22:06 +0000 (16:22 +0000)]
Implement a bus_space_bad_addr() function to aid in some device probing with
the bus_space universe.

27 years agoMove the PB140/145 info to it's own case. A couple of the entry points
briggs [Sat, 5 Apr 1997 16:19:29 +0000 (16:19 +0000)]
Move the PB140/145 info to it's own case.  A couple of the entry points
are different from other cases.

27 years agoMRG-based ADB support for the PowerBook series. Contributed by
briggs [Sat, 5 Apr 1997 15:29:10 +0000 (15:29 +0000)]
MRG-based ADB support for the PowerBook series.  Contributed by
hamada@next.etron.kanazawa-it.ac.jp (Takashi Hamada)

27 years agoDocument that negative offsets will cause EINVAL as well.
tholo [Sat, 5 Apr 1997 05:07:58 +0000 (05:07 +0000)]
Document that negative offsets will cause EINVAL as well.

27 years agoCheck for overflow; from FreeBSD
tholo [Sat, 5 Apr 1997 05:05:44 +0000 (05:05 +0000)]
Check for overflow; from FreeBSD

27 years agoPatch from BSDI (via Keith Bostic, via NetBSD):
tholo [Sat, 5 Apr 1997 04:37:40 +0000 (04:37 +0000)]
Patch from BSDI (via Keith Bostic, via NetBSD):
>NFS doesn't support chflags; ignore errors unless there's reason
>to believe we're losing bits.  (Note, this still won't be right
>if the server supports flags and we were trying to *remove* flags
>on a file that we copied, i.e., that we didn't create.)

27 years agoso strlen.c compiles again...
millert [Sat, 5 Apr 1997 04:26:54 +0000 (04:26 +0000)]
so strlen.c compiles again...

27 years agoFrom NetBSD PR/3405: Krister Walfridsson <cato@ulysses.df.lth.se>: Various
tholo [Sat, 5 Apr 1997 04:18:45 +0000 (04:18 +0000)]
From NetBSD PR/3405: Krister Walfridsson <cato@ulysses.df.lth.se>: Various
typos; since we compile with POSIX libm, we should not xref matherr(3), and
describe the errno settings.

27 years agoFix -DNO_MMALLOC (needs to be in MMALLOC_CFLAGS not MMALLOC_DISABLE).
millert [Sat, 5 Apr 1997 02:42:24 +0000 (02:42 +0000)]
Fix -DNO_MMALLOC (needs to be in MMALLOC_CFLAGS not MMALLOC_DISABLE).
Now gdb will link.

27 years agoRLIMIT_NPROC; mike.long@analog.com
deraadt [Fri, 4 Apr 1997 22:59:46 +0000 (22:59 +0000)]
RLIMIT_NPROC; mike.long@analog.com

27 years agoclarify
deraadt [Fri, 4 Apr 1997 22:20:00 +0000 (22:20 +0000)]
clarify

27 years agoinit allocated memory.
mickey [Fri, 4 Apr 1997 21:52:40 +0000 (21:52 +0000)]
init allocated memory.

27 years agoImprovements based on commets by George Robbins <grr@shandakor.tharsis.com>
millert [Fri, 4 Apr 1997 20:54:50 +0000 (20:54 +0000)]
Improvements based on commets by George Robbins <grr@shandakor.tharsis.com>
 - ``tar cf'' will now just exit instead of waiting for a file list
   on stdin.  The old behavior should be re-enabled with an option.
 - Better wording in tar man page for -o and -w

27 years agowith small quantities of beer and pizza i produce fewer bugs. really.
deraadt [Fri, 4 Apr 1997 19:56:35 +0000 (19:56 +0000)]
with small quantities of beer and pizza i produce fewer bugs. really.

27 years agoif the user changes an option, spit out a warning telling them they need
deraadt [Fri, 4 Apr 1997 19:54:29 +0000 (19:54 +0000)]
if the user changes an option, spit out a warning telling them they need
to make clean. i expect i will get compliments for this change, but
i'd really rather receive beer and pizza.

27 years agowhen mktemp() is hard to fix, use 10 X
deraadt [Fri, 4 Apr 1997 18:41:25 +0000 (18:41 +0000)]
when mktemp() is hard to fix, use 10 X

27 years agocheck MID field like strip(1) does; PR#149, mickey; fix by me
deraadt [Fri, 4 Apr 1997 18:27:07 +0000 (18:27 +0000)]
check MID field like strip(1) does; PR#149, mickey; fix by me

27 years agosort & add a few
deraadt [Fri, 4 Apr 1997 17:50:34 +0000 (17:50 +0000)]
sort & add a few

27 years agodescribe 2nd utmp record better; gwing@primenet.com.au
deraadt [Fri, 4 Apr 1997 17:35:07 +0000 (17:35 +0000)]
describe 2nd utmp record better; gwing@primenet.com.au

27 years agoNuke.
weingart [Fri, 4 Apr 1997 17:24:47 +0000 (17:24 +0000)]
Nuke.

27 years agoReplace exec_i386.S with these two files. Startprog.S is a
weingart [Fri, 4 Apr 1997 17:23:29 +0000 (17:23 +0000)]
Replace exec_i386.S with these two files.  Startprog.S is a
stopgap measure, to be removed once I get locore.s done.

27 years agodo not whack pid file if run -d; hwr@pilhuhn.de
deraadt [Fri, 4 Apr 1997 16:57:01 +0000 (16:57 +0000)]
do not whack pid file if run -d; hwr@pilhuhn.de

27 years ago#include <string.h> since we use string functions...
millert [Fri, 4 Apr 1997 16:51:37 +0000 (16:51 +0000)]
#include <string.h> since we use string functions...

27 years agoseek(2) is not needed
deraadt [Fri, 4 Apr 1997 16:45:25 +0000 (16:45 +0000)]
seek(2) is not needed

27 years agofixed bcopy/memcpy conflict by using private mem* routines.
mickey [Fri, 4 Apr 1997 16:17:02 +0000 (16:17 +0000)]
fixed bcopy/memcpy conflict by using private mem* routines.
changed generic pointer type to (void *)
and, md5 is in libkern, since more than one component of the kernel uses it.
what would be better place? well, imagine some lkm that uses md5 someday...

27 years agoadd 'machine check' trap for Pentium and PPro cpus.
mickey [Fri, 4 Apr 1997 16:14:07 +0000 (16:14 +0000)]
add 'machine check' trap for Pentium and PPro cpus.
no special handler for it. i guess there will be rare cases when
you receive it, and then you should repair your cpu, since
it happens on internal cpu error.

27 years agoremove copy.s as it is in M.I. m68k files now
kstailey [Fri, 4 Apr 1997 14:56:57 +0000 (14:56 +0000)]
remove copy.s as it is in M.I. m68k files now

27 years agoDo not set DCR_SYNC for the C/Q[68]xx systems.
briggs [Fri, 4 Apr 1997 14:48:56 +0000 (14:48 +0000)]
Do not set DCR_SYNC for the C/Q[68]xx systems.
Thanks to Steve Brown <sbrown@best.com> for working with me to find this.

27 years agochanges required to get gcc with the new g77 0.5.20 compiled smoothly
graichen [Fri, 4 Apr 1997 13:25:39 +0000 (13:25 +0000)]
changes required to get gcc with the new g77 0.5.20 compiled smoothly
on OpenBSD

27 years agosync g77 to version 0.5.20 - i hope i got everything right because there
graichen [Fri, 4 Apr 1997 13:18:15 +0000 (13:18 +0000)]
sync g77 to version 0.5.20 - i hope i got everything right because there
is no patch from 0.5.19 to 0.5.20 - so i did it by diffing two gcc trees
looking carefully at the results

what does the new g77 give us:
* now it completely works on the alpha (64bit)
* faster
* less bugs :-)

27 years agoimport of libU77 - part of sync to g77-0.5.20
graichen [Fri, 4 Apr 1997 13:11:09 +0000 (13:11 +0000)]
import of libU77 - part of sync to g77-0.5.20

27 years agoblow away source routed sockets
deraadt [Fri, 4 Apr 1997 09:07:05 +0000 (09:07 +0000)]
blow away source routed sockets

27 years agodo source routing check correctly; this could use a test
deraadt [Fri, 4 Apr 1997 09:06:28 +0000 (09:06 +0000)]
do source routing check correctly; this could use a test

27 years agowtmp repairs; netbsd
deraadt [Fri, 4 Apr 1997 08:59:38 +0000 (08:59 +0000)]
wtmp repairs; netbsd

27 years agomention mr.tarkovskey
mickey [Fri, 4 Apr 1997 05:54:26 +0000 (05:54 +0000)]
mention mr.tarkovskey

27 years agodump regs w/ NMI (int 2)
mickey [Fri, 4 Apr 1997 04:47:44 +0000 (04:47 +0000)]
dump regs w/ NMI (int 2)
more debugging is on the way

27 years agoFix NTOHS and HTONS macros.
millert [Fri, 4 Apr 1997 03:17:06 +0000 (03:17 +0000)]
Fix NTOHS and HTONS macros.

27 years agoDefine in_addr_t and in_port_t everywhere but use u_intXX_t for
millert [Fri, 4 Apr 1997 03:05:28 +0000 (03:05 +0000)]
Define in_addr_t and in_port_t everywhere but use u_intXX_t for
htonl and friends.

27 years agowe are a 4.4bsd-style system; correct pathnames
deraadt [Fri, 4 Apr 1997 02:02:58 +0000 (02:02 +0000)]
we are a 4.4bsd-style system; correct pathnames

27 years agoadd STRIPDISC line; kleink@netbsd.org
deraadt [Fri, 4 Apr 1997 01:47:29 +0000 (01:47 +0000)]
add STRIPDISC line; kleink@netbsd.org

27 years agos/STRIPIPDISC/STRIPDISC/; kleink@netbsd.org
deraadt [Fri, 4 Apr 1997 01:46:03 +0000 (01:46 +0000)]
s/STRIPIPDISC/STRIPDISC/; kleink@netbsd.org

27 years agopretty
deraadt [Thu, 3 Apr 1997 21:01:07 +0000 (21:01 +0000)]
pretty

27 years agoAvoid confusion on running /etc/daily on daylight-savings change days.
gene [Thu, 3 Apr 1997 20:12:07 +0000 (20:12 +0000)]
Avoid confusion on running /etc/daily on daylight-savings change days.

27 years agofix my last fix. warnx, not warn!
mickey [Thu, 3 Apr 1997 19:33:50 +0000 (19:33 +0000)]
fix my last fix. warnx, not warn!

27 years agomore about Gogol (;
mickey [Thu, 3 Apr 1997 19:20:11 +0000 (19:20 +0000)]
more about Gogol (;

27 years agotabify
kstailey [Thu, 3 Apr 1997 19:03:09 +0000 (19:03 +0000)]
tabify

27 years agotake 2
kstailey [Thu, 3 Apr 1997 17:39:38 +0000 (17:39 +0000)]
take 2

27 years agomake it match what is in magdir/magic
kstailey [Thu, 3 Apr 1997 17:35:47 +0000 (17:35 +0000)]
make it match what is in magdir/magic

27 years ago>Number: 3412
kstailey [Thu, 3 Apr 1997 15:46:36 +0000 (15:46 +0000)]
>Number:         3412
>Category:       port-alpha
>Synopsis:       incorrect use of long crashes netbsd/alpha
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Mar 29 23:50:01 1997
>Last-Modified:
>Originator:     Ross Harvey
>Organization:
Avalon Computer Systems, Inc.
>Release:        NetBSD-current
>Environment:
System: NetBSD epsilon.ghs.com 1.2D NetBSD 1.2D (e) #5: Sat Mar 29 22:32:33 PST 1997 ross@epsilon.ghs.com:/bsd/ross/e alpha

>Description:
Adding "pseudo-device ipfilter 1" to netbsd/alpha will
cause it to crash when ifconfig turns on the interface.

Ipfilter code in netinet/fil.c uses pointer arithmetic to
effectively make structure-to-scalar casts. It uses long when
it really wants an int32 and gets alignment faults that crash
netbsd/alpha. General 64 bit problem, not exactly alpha.

>How-To-Repeat:
Add pseudo-device ipfilter 1. Boot your alpha. 2. Try to come
up multiuser. The first packet or so will halt everything.

>Fix:
Apply patch...

27 years agoQuote CC in args to make
briggs [Thu, 3 Apr 1997 13:01:40 +0000 (13:01 +0000)]
Quote CC in args to make

27 years agohokchar -> _hokchar, and make it non-static. Use it in getnetnamadr.c as
downsj [Thu, 3 Apr 1997 08:33:03 +0000 (08:33 +0000)]
hokchar -> _hokchar, and make it non-static.  Use it in getnetnamadr.c as
well.

27 years agoanother lookup[] loop accident
deraadt [Thu, 3 Apr 1997 07:54:01 +0000 (07:54 +0000)]
another lookup[] loop accident

27 years agoFull implementation of lookups support, plus some bug fixes. Still no YP.
downsj [Thu, 3 Apr 1997 07:31:55 +0000 (07:31 +0000)]
Full implementation of lookups support, plus some bug fixes.  Still no YP.

27 years agoan -> a; soil@quick.net
deraadt [Thu, 3 Apr 1997 07:22:31 +0000 (07:22 +0000)]
an -> a; soil@quick.net

27 years agoFrom FreeBSD 1.7 by markm. The log message:
imp [Thu, 3 Apr 1997 07:07:36 +0000 (07:07 +0000)]
From FreeBSD 1.7 by markm.  The log message:
Priorities were broken. If there was an Index: line and ***/--- lines
with valid names, the ***/---names were taken first.
this broke eg:
Index: foo/Makefile
==========
RCS <blah>
Retrieving <blah>
diff <blah>
*** Makefile <blah>
--- Makefile <blah>

By trying to patch the Makefile in the _curent_ directory, rather than
the one in the foo/ directory.

27 years agoBUGS section no longer applies, the new resolver implements network
downsj [Thu, 3 Apr 1997 06:54:58 +0000 (06:54 +0000)]
BUGS section no longer applies, the new resolver implements network
lookups with DNS.

27 years agouse MATCH()
deraadt [Thu, 3 Apr 1997 05:53:46 +0000 (05:53 +0000)]
use MATCH()

27 years agoif dns lookup fails, still honour _res.lookups[]
deraadt [Thu, 3 Apr 1997 05:53:23 +0000 (05:53 +0000)]
if dns lookup fails, still honour _res.lookups[]

27 years agoensure host_addr[] remains nicely aligned; pr#145, grr@shandakor.tharsis.com
deraadt [Thu, 3 Apr 1997 05:52:35 +0000 (05:52 +0000)]
ensure host_addr[] remains nicely aligned; pr#145, grr@shandakor.tharsis.com

27 years ago10 X's for mktemp() and mkstemp().
millert [Thu, 3 Apr 1997 05:31:37 +0000 (05:31 +0000)]
10 X's for mktemp() and mkstemp().

27 years agoPR#141 from David Leonard <d@occult.fnarg.net.au> with some small mods by
briggs [Thu, 3 Apr 1997 03:53:25 +0000 (03:53 +0000)]
PR#141 from David Leonard <d@occult.fnarg.net.au> with some small mods by
me.
- Provide option (enabled in GENERIC and GENERICSBC) to disable ADB
  configuration when serial console is enabled.  Old behavior allows
  booting even if ADB probe hangs (Apple had several ways to access
  ADB).
- Configure ite0 differently--don't assume that it's only extant if it
  is the console.

27 years agoa start a making getnetbyaddr() respect "lookup" keyword in resolv.conf
kstailey [Thu, 3 Apr 1997 02:15:16 +0000 (02:15 +0000)]
a start a making getnetbyaddr() respect "lookup" keyword in resolv.conf

27 years agoadd htonl()
kstailey [Thu, 3 Apr 1997 02:12:57 +0000 (02:12 +0000)]
add htonl()

27 years agoper wishes of deraadt add ufs to default fs's to search.
michaels [Thu, 3 Apr 1997 00:57:49 +0000 (00:57 +0000)]
per wishes of deraadt add ufs to default fs's to search.

27 years agoremove corrupted original tag
michaels [Thu, 3 Apr 1997 00:57:37 +0000 (00:57 +0000)]
remove corrupted original tag

27 years agoUse new makeinfo -P dir option to prepend gcc/f to the include path.
kstailey [Wed, 2 Apr 1997 22:57:27 +0000 (22:57 +0000)]
Use new makeinfo -P dir option to prepend gcc/f to the include path.

You must rebuild and install the new makeinfo before building gcc!

This corrects the gcc/install.texi gcc/f/install.texi name collision.

27 years agoAdd "-P dir" option to prepend a dir to the include path.
kstailey [Wed, 2 Apr 1997 22:55:36 +0000 (22:55 +0000)]
Add "-P dir" option to prepend a dir to the include path.
This is driven by g77 using install.texi which collides with gcc install.texi.
This has been submitted to bug-texinfo@prep.ai.mit.edu.

*** You must rebuild/reinstall makeinfo before building gcc! ***

27 years agoBetter implementation of buf_space_*_raw_multi_* that deals well
niklas [Wed, 2 Apr 1997 22:08:07 +0000 (22:08 +0000)]
Better implementation of buf_space_*_raw_multi_* that deals well
with unaligned buffers

27 years agoFix end boundary of mapping for regions ending at offset 1 in a page.
niklas [Wed, 2 Apr 1997 22:02:55 +0000 (22:02 +0000)]
Fix end boundary of mapping for regions ending at offset 1 in a page.
From NetBSD PR#3441, Matthias Drochner <drochner@zelz26.zel.kfa-juelich.de>

27 years agoRestructured, as well as added a set of host-to-little/big-endian-16/32-bit
niklas [Wed, 2 Apr 1997 20:40:47 +0000 (20:40 +0000)]
Restructured, as well as added a set of host-to-little/big-endian-16/32-bit
utility macros (and their opposites), named like htole16 etc.

27 years agouse FDPART(), limit drives to 2 to avoid dev_t problems; grr@shandakor.tharsis.com
deraadt [Wed, 2 Apr 1997 18:28:48 +0000 (18:28 +0000)]
use FDPART(), limit drives to 2 to avoid dev_t problems; grr@shandakor.tharsis.com

27 years agowarn about MAXPARTITIONS != 16, grr@shandakor.tharsis.com
deraadt [Wed, 2 Apr 1997 18:28:09 +0000 (18:28 +0000)]
warn about MAXPARTITIONS != 16, grr@shandakor.tharsis.com

27 years agodoc /etc/yp/DOMAINNAME server list file better
deraadt [Wed, 2 Apr 1997 18:05:47 +0000 (18:05 +0000)]
doc /etc/yp/DOMAINNAME server list file better