obnox/wireshark/wip.git
19 years agoIn the final scene of the horror movie, just when you think the monster
guy [Wed, 26 Jul 2000 06:04:34 +0000 (06:04 +0000)]
In the final scene of the horror movie, just when you think the monster
is finally dead, and you're walking away, it springs up again and
attacks.

It appears that the ss990915 version of Alexey Kuznetzov's libpcap patch
has some extra stuff in the per-packet header for some sort of SMP
debugging, and that SuSE Linux 6.3 picked it up.

Thus, even if a libpcap file has the modified magic number, we *still*
have to go through the usual heuristic hell to figure out what type of
file it is.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2164 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoActually delete the top ethereal.spec.in and packaging/solaris/* files this
gerald [Wed, 26 Jul 2000 03:46:02 +0000 (03:46 +0000)]
Actually delete the top ethereal.spec.in and packaging/solaris/* files this
time.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2163 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoSwitch Solaris package build naming to a more generic SVR4. Add RPM and SRPM
gerald [Wed, 26 Jul 2000 03:39:11 +0000 (03:39 +0000)]
Switch Solaris package build naming to a more generic SVR4.  Add RPM and SRPM
package build targets.  Move ethereal.spec(.in) to packaging/rpm.

The spec file is different from Henri's.  We might want to switch to his
for the sake of consistency.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2162 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoWhen rescanning a file, all state information for the frames has
gram [Wed, 26 Jul 2000 03:08:56 +0000 (03:08 +0000)]
When rescanning a file, all state information for the frames has
been deleted. So we have to set fdata->flags.visited to 0 for each frame,
denoting a "fresh" scan.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2161 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoUse unsigned character pointers and arrays rather than signed character
guy [Wed, 26 Jul 2000 00:20:09 +0000 (00:20 +0000)]
Use unsigned character pointers and arrays rather than signed character
pointers and arrays in a number of places, to remove warnings some
compilers give.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2160 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoA pile of DHCP option names, supplied by Jose Pedro Oliveira.
guy [Tue, 25 Jul 2000 21:26:08 +0000 (21:26 +0000)]
A pile of DHCP option names, supplied by Jose Pedro Oliveira.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2159 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoUse the right length when adding the "agent address" entry to the tree
guy [Tue, 25 Jul 2000 17:30:47 +0000 (17:30 +0000)]
Use the right length when adding the "agent address" entry to the tree
for an SNMP trap PDU.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2158 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoFix a comment.
guy [Tue, 25 Jul 2000 07:37:54 +0000 (07:37 +0000)]
Fix a comment.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2157 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoCorrect packet count displayed while capturing with -w option. (there was
gram [Mon, 24 Jul 2000 16:27:34 +0000 (16:27 +0000)]
Correct packet count displayed while capturing with -w option. (there was
an extra increment of cfile.count).

Print a newline to stdout at the end of a capture with "-w" so that the
final packet count isn't obliterated by the shell prompt.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2156 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd a "debian-package" target to Makefile.am to build debian packages.
oabad [Sat, 22 Jul 2000 20:56:13 +0000 (20:56 +0000)]
Add a "debian-package" target to Makefile.am to build debian packages.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2155 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoImproved "autogen.sh" script from Bruce Korb.
guy [Sat, 22 Jul 2000 20:00:29 +0000 (20:00 +0000)]
Improved "autogen.sh" script from Bruce Korb.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2154 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoThe latest version of libpcap is now at www.tcpdump.org, not at LBL.
guy [Sat, 22 Jul 2000 19:28:46 +0000 (19:28 +0000)]
The latest version of libpcap is now at tcpdump.org, not at LBL.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2153 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoSimplify the way the display filter routines get field values from
gram [Sat, 22 Jul 2000 15:58:54 +0000 (15:58 +0000)]
Simplify the way the display filter routines get field values from
the proto tree. Now, proto_get_finfo_ptr_array() can easily be used by
any routine, not just display filter code, to get values from the proto
tree. This might be useful if one were to allow columns in the packet list
to show the value of an arbitrary field.

Fixed a memleak when filtering on a byte arrays.

Fixed erroneous asserts in dfilter-grammar.y, where I used '=' instead of '=='.
They had to do with byte-arrays, too.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2152 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd support for bulding Solaris binary packages. The distribution
gerald [Sat, 22 Jul 2000 04:10:23 +0000 (04:10 +0000)]
Add support for bulding Solaris binary packages.  The distribution
is installed in packaging/solaris.stage, and from there the package
is created.  The checkinstall script depends on GTK+/Glib residing
in /usr/local.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2151 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoGet rid of "Playback" in capture window title.
gram [Fri, 21 Jul 2000 15:56:15 +0000 (15:56 +0000)]
Get rid of "Playback" in capture window title.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2150 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd packet-sctp.c
gram [Fri, 21 Jul 2000 13:18:22 +0000 (13:18 +0000)]
Add packet-sctp.c

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2149 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd SCTP dissector from Tuexen Michael <Michael.Tuexen@icn.siemens.de>
gram [Fri, 21 Jul 2000 12:57:09 +0000 (12:57 +0000)]
Add SCTP dissector from Tuexen Michael <Michael.Tuexen@icn.siemens.de>

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2148 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoChanges from Gerrit Gehnen to make it use tvbuffs and to display the
guy [Fri, 21 Jul 2000 07:51:34 +0000 (07:51 +0000)]
Changes from Gerrit Gehnen to make it use tvbuffs and to display the
response identifier in hex rather than decimal.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2147 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoGive a URL that goes directly to the W2K Server documentation
guy [Fri, 21 Jul 2000 01:40:41 +0000 (01:40 +0000)]
Give a URL that goes directly to the W2K Server documentation
information on WINS and WINS-R records.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2146 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd support for Microsoft's WINS and WINS-R resource records.
guy [Fri, 21 Jul 2000 01:29:04 +0000 (01:29 +0000)]
Add support for Microsoft's WINS and WINS-R resource records.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2145 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIn Tethereal:
guy [Thu, 20 Jul 2000 09:39:43 +0000 (09:39 +0000)]
In Tethereal:

When capturing, report errors trying to create the output file
with "file_open_error_message()".

Make the "for_writing" argument to "file_open_error_message()" a
"gboolean", as it's either TRUE (if the file is being opened for
writing) or FALSE (if it's being opened for reading).

Report EISDIR as "XXX is a directory (folder), not a file.".

When checking whether an "open()" of a capture file succeeded, check
whether "open()" returns a negative number, not whether it returns 0.

In "wtap_open_offline()", if the file to be opened is a directory,
return EISDIR, not WTAP_ERR_NOT_REGULAR_FILE, so that the error message
can say "that's a directory, not a file".

If "wtap_open_offline()" returns WTAP_ERR_NOT_REGULAR_FILE, don't just
say the file is "invalid", say it's a "special file" or socket or some
other weird type of file.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2144 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoUse ESD_TYPE_CRIT for most errors (the model used by various GUIs seems
guy [Thu, 20 Jul 2000 05:10:02 +0000 (05:10 +0000)]
Use ESD_TYPE_CRIT for most errors (the model used by various GUIs seems
to use "warning" dialog boxes only to warn the user "if you do that, bad
things may happen" *and* to offer them the option either to drive on or
quit, so perhaps ESD_TYPE_CRIT should be used for all errors).

However, put "Ethereal: Error" rather than "Ethereal: Critical" in the
title bar, in the hopes that it'll make it clearer that Something Bad
Happened.

If the user specifies that captures should be saved to a user-specified
file rather than a temporary file, report errors trying to create that
file with "file_open_error_message()".

Make the "for_writing" argument to "file_open_error_message()" a
"gboolean", as it's either TRUE (if the file is being opened for
writing) or FALSE (if it's being opened for reading).

Report EISDIR as "XXX is a directory (folder), not a file.".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2143 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoDNSSEC additions/name changes from Jakob Schlyter <jakob@crt.se>.
gerald [Tue, 18 Jul 2000 16:53:51 +0000 (16:53 +0000)]
DNSSEC additions/name changes from Jakob Schlyter <jakob@crt.se>.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2142 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoSupport for dissecting RPCSEC_GSS credentials in ONC RPC, from Dug Song.
guy [Mon, 17 Jul 2000 20:34:00 +0000 (20:34 +0000)]
Support for dissecting RPCSEC_GSS credentials in ONC RPC, from Dug Song.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2141 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agodissect_rpc() listed as an heuristic dissector for tcp and udp.
girlich [Fri, 14 Jul 2000 12:55:58 +0000 (12:55 +0000)]
dissect_rpc() listed as an heuristic dissector for tcp and udp.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2140 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoDirect call to dissect_rpc() removed. This can be made much clearer with
girlich [Fri, 14 Jul 2000 12:54:32 +0000 (12:54 +0000)]
Direct call to dissect_rpc() removed. This can be made much clearer with
an heuristic dissector.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2139 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoHeuristic dissector table added (just like tcp). Direct call to
girlich [Fri, 14 Jul 2000 12:53:00 +0000 (12:53 +0000)]
Heuristic dissector table added (just like tcp). Direct call to
dissect_rpc() could finally disappear.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2138 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoApparently, on systems with glibc 2.2, "inet_aton()" is declared in
guy [Fri, 14 Jul 2000 07:11:53 +0000 (07:11 +0000)]
Apparently, on systems with glibc 2.2, "inet_aton()" is declared in
<arpa/inet.h>, but is, in some fashion, declared differently from the
way we declare it in "inet_v6defs.h", but "inet_ntop()" isn't defined,
so we include "inet_v6defs.h" in "inet_pton.c", which causes
"inet_pton.c" not to compile as we get a collision between the two
declarations.

Move the declaration of "inet_aton()" to "inet_aton.h", define
"NEED_INET_ATON_H" iff we didn't find "inet_aton()" in the system
libraries, and include "inet_aton.h" in the callers of "inet_aton()" iff
"NEED_INET_ATON_H" is defined, so that it doesn't get declared by us if
"inet_aton()" is defined by a system library (which hopefully means it's
declared in <arpa/inet.h> instead).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2137 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoEnhancements/fixed to OSPF from Michael Rozhavsky:
guy [Fri, 14 Jul 2000 03:23:50 +0000 (03:23 +0000)]
Enhancements/fixed to OSPF from Michael Rozhavsky:

1) DD flags in Database Description packets were translated
   incorrectly;

2) Summary-LSA metric was not displayed.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2136 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd an option to show/hide the TCP summary information in the protocol
gram [Thu, 13 Jul 2000 14:16:49 +0000 (14:16 +0000)]
Add an option to show/hide the TCP summary information in the protocol
label in the protocol tree.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2135 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoFile handle dissection works now with the old Linux user-land nfsd too.
girlich [Thu, 13 Jul 2000 13:09:25 +0000 (13:09 +0000)]
File handle dissection works now with the old Linux user-land nfsd too.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2134 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoStart of a NFS file handle dissection. ReliantUNIX and Linux knfsd handles
girlich [Thu, 13 Jul 2000 05:48:50 +0000 (05:48 +0000)]
Start of a NFS file handle dissection. ReliantUNIX and Linux knfsd handles
are correctly dissected.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2133 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIn the "Preferences" dialog box, "Save" sets the current preference
guy [Mon, 10 Jul 2000 09:18:38 +0000 (09:18 +0000)]
In the "Preferences" dialog box, "Save" sets the current preference
values to the ones specified in the dialog box, so it should also
redissect the packets and re-evaluate the display filter if any of them
changed.  (I.e., it did part of an "Apply"; it needs to do all of it.)

"Cancel" also needs to find out if any preferences were changed by the
reversion and redissect and refilter if they did.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2132 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoTreat only ASCII characters as printable in TSAPs; otherwise, as 0xff
guy [Mon, 10 Jul 2000 06:52:29 +0000 (06:52 +0000)]
Treat only ASCII characters as printable in TSAPs; otherwise, as 0xff
and 0xfe, for example, are printable characters in many locales (they're
printable in ISO 8859/x, for example), a TSAP of 0xfffffffefffffffe will
be treated as printable if your locale is one of those, even though the
chances that the TSAP is should be read as a string of y-with-diaresis
and lower-case thorn are pretty slim.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2131 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPut up a progress dialog when changing the time display; it can take a
guy [Sun, 9 Jul 2000 23:22:33 +0000 (23:22 +0000)]
Put up a progress dialog when changing the time display; it can take a
long time on a large capture.

If the user is changing the setting of "Display TOS as DiffServ" *and*
the time stamp format, don't bother doing the time display update -
we'll regenerate the entire display anyway because we changed the
setting of "Display TOS as DiffServ".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2130 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoCheck the options in a BOOTP packet regardless of whether the protocol
guy [Sun, 9 Jul 2000 22:46:53 +0000 (22:46 +0000)]
Check the options in a BOOTP packet regardless of whether the protocol
tree is to be constructed or not; if it has the DHCP option, we want to
mark the packet as DHCP even if we're not constructing the protocol
tree.

Extract the code to process Netware/IP suboptions into a separate
routine.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2129 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIn the "Display Options" dialog box code, separately keep track of the
guy [Sun, 9 Jul 2000 20:59:23 +0000 (20:59 +0000)]
In the "Display Options" dialog box code, separately keep track of the
timestamp type when the dialog box was popped up, so we can revert to it
if the user clicks "Cancel", and the timestamp type as of when we
started *or* as of when we last changed it, so that we can check whether
the user's changed it from its current value and only redisplay the time
fields if they've changed it (as redisplaying the time fields could be
expensive - we have to scan through all the packets and rebuild the
packet list).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2128 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoDocument the "-o" flag, the fact that the "Preferences" dialog box has
guy [Sun, 9 Jul 2000 04:09:09 +0000 (04:09 +0000)]
Document the "-o" flag, the fact that the "Preferences" dialog box has
pages for various protocol preferences, and the existence of the global
and personal preferences files.

We still need to document the values that can be put in the preferences
file.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2127 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoTurn the code of "colorize_packet()" into a static routine that is given
guy [Sun, 9 Jul 2000 03:29:42 +0000 (03:29 +0000)]
Turn the code of "colorize_packet()" into a static routine that is given
a word to use in the progress dialog, and a flag indicating whether the
display filter is to be reevaluated or not, and:

have "colorize_packet()" call that routine with "Colorizing" and
FALSE as those arguments;

have the filtering code call that routine with "Filtering" and
TRUE as those arguments;

add an exported routine to call that routine with "Reprocessing"
and TRUE as those arguments, to use to re-generate the packet
list and to re-filter the packets if a protocol preference has
been changed.

Keep track of whether preferences are changed from their initial value
by a preferences file or a command-line option, or from their previous
value by the "Preferences" dialog box; have "prefs_apply_all()" only
call the "apply" callback for a module if they have.

Call "prefs_apply_all()" after the command-line arguments have been
parsed and after "OK" has been clicked in the "Preferences" dialog box,
to notify modules of preference changes if they've registered a callback
for that.

After "OK" has been clicked in the "Preferences" dialog box, if any
preferences have changed, call the reprocessing routine, as the summary
line for some frames and/or the current display filter's value when
applied to some frames may have changed as a result of a preference
change.  Do the same after "OK" or "Apply" has been clicked in the
"Display Options" dialog box (as it controls a protocol preferences
item.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2126 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd preference for placement of AH payload, at same level or in subtree.
gram [Sat, 8 Jul 2000 10:46:23 +0000 (10:46 +0000)]
Add preference for placement of AH payload, at same level or in subtree.

Move max_len settings in various col_* functions outside of loop.

Add 'writable' flag to col_info. check_col() honors its. dissect_packet()
sets it as TRUE. dissect_ah() optionally sets it to FALSE.

Add col_set_writable() function to set the 'writable' flag. Accepts
frame_data arg just like the rest of the column functions. It checks to
make sure fd->cinfo is not NULL.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2125 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoThe security model values 1 and 2 ("reserved for SNMPv1" and "reserved
guy [Sat, 8 Jul 2000 08:33:29 +0000 (08:33 +0000)]
The security model values 1 and 2 ("reserved for SNMPv1" and "reserved
for SNMPv2") in SNMPv3 packets shouldn't cause the security parameters
to be interpreted as a character string - those values are apparently
not supposed to show up in packets - so we handle those values in the
default case, with the security parameters treated as opaque data.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2124 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoThe BOOTP "magic cookie" should be added in exactly the same fashion
guy [Sat, 8 Jul 2000 07:52:11 +0000 (07:52 +0000)]
The BOOTP "magic cookie" should be added in exactly the same fashion
regardless of whether it's valid or not - extract all four bytes and add
that as the value.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2123 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIf the progress dialog gets a "delete" event, have the handler return
guy [Fri, 7 Jul 2000 23:09:15 +0000 (23:09 +0000)]
If the progress dialog gets a "delete" event, have the handler return
TRUE, meaning "don't delete this", and if its "stop this operation"
button gets clicked, don't have its handler delete the progress dialog;
instead, leave the deletion of the window up to the code using the
dialog, as it'll do so when it stops the operation.

Make the "read file" operation destroy the dialog; all the other
operations already did so (as they just broke out of the loop doing the
operation, rather than returning).

Don't catch the "destroy" operation on the dialog box - its handler
appeared never to get called; we can just free the "progdlg_t" for the
dialog in "destroy_progress_dlg()", right after destroying the dialog
box widget.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2122 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoUse "progdlg_t *" rather than "void *" as the handle for a progress
guy [Fri, 7 Jul 2000 09:30:56 +0000 (09:30 +0000)]
Use "progdlg_t *" rather than "void *" as the handle for a progress
dialog box; that lets us do some type-checking, but we can still typedef
it to an incompletely-defined structure to hide the implementation
details from the caller.

Make "create_progress_dlg()" take, as an argument, the title to put in
the "stop the operation" button, and use "Stop" rather than "Cancel" if
stopping the operation doesn't undo all the work it's done.

Thaw the clist if we break out of a "read the file" operation, as we
freeze it before the operation.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2121 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoUse "progdlg_t *" rather than "void *" as the handle for a progress
guy [Fri, 7 Jul 2000 07:01:58 +0000 (07:01 +0000)]
Use "progdlg_t *" rather than "void *" as the handle for a progress
dialog box; that lets us do some type-checking, but we can still typedef
it to an incompletely-defined structure to hide the implementation
details from the caller.

Make "create_progress_dlg()" take, as an argument, the title to put in
the "stop the operation" button, and use "Stop" rather than "Cancel" if
stopping the operation doesn't undo all the work it's done.

Thaw the clist if we break out of a "read the file" operation, as we
freeze it before the operation.

Have the handler for the "delete" event on the progress dialog box
return FALSE, to let GTK+ know that it should, in fact, delete the
window.  ("delete" event handlers should return TRUE if the window
shouldn't actually be deleted, FALSE if it should; they should not
return "void".)

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2120 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoSINIX-M systems don't know about strcasecmp(). I added the source file
girlich [Thu, 6 Jul 2000 10:03:44 +0000 (10:03 +0000)]
SINIX-M systems don't know about strcasecmp(). I added the source file
from the GNU C Library 2.0.7 like I did it with strncasecmp().

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2119 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd:
gram [Wed, 5 Jul 2000 17:24:29 +0000 (17:24 +0000)]
Add:
#define strcasecmp              stricmp

for prefs.c

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2118 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd support for a global "ethereal.conf" preferences file, stored in the
guy [Wed, 5 Jul 2000 09:41:07 +0000 (09:41 +0000)]
Add support for a global "ethereal.conf" preferences file, stored in the
same directory as the "manuf" file ("/etc" or "/usr/local/etc", most
likely).

Add a mechanism to allow modules (e.g., dissectors) to register
preference values, which:

can be put into the global or the user's preference file;

can be set from the command line, with arguments to the "-o"
flag;

can be set from tabs in the "Preferences" dialog box.

Use that mechanism to register the "Decode IPv4 TOS field as DiffServ
field" variable for IP as a preference.

Stuff that still needs to be done:

documenting the API for registering preferences;

documenting the "-o" values in the man page (probably needs a
flag similar to "-G", and a Perl script to turn the output into
documentation as is done with the list of field);

handling error checking for numeric values (range checking,
making sure that if the user changes the variable from the GUI
they change it to a valid numeric value);

using the callbacks to, for example, update the display when
preferences are changed (could be expensive);

panic if the user specifies a numeric value with a base other
than 10, 8, or 16.

We may also want to clean up the existing wired-in preferences not to
take effect the instant you tweak the widget, and to add an "Apply"
button to the "Preferences" dialog.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2117 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoGet rid of the stuff to let the callback that pops up the Preferences
guy [Wed, 5 Jul 2000 06:33:02 +0000 (06:33 +0000)]
Get rid of the stuff to let the callback that pops up the Preferences
dialog select a particular page - I think that was used only by the
filter code back when "Filter:" buttons popped up a Preferences dialog
with the Filter page (which is no longer a Preferences dialog page)
selected, but now there's a separate Filter dialog box.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2116 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPut the close button in an HButtonBox, so that it doesn't expand to the
guy [Wed, 5 Jul 2000 06:19:27 +0000 (06:19 +0000)]
Put the close button in an HButtonBox, so that it doesn't expand to the
width of the window.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2115 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPut the "Cancel" button in a progress dialog box in an HButtonBox,
guy [Wed, 5 Jul 2000 05:50:00 +0000 (05:50 +0000)]
Put the "Cancel" button in a progress dialog box in an HButtonBox,
rather than an Alignment; that appears to be the correct way to keep
buttons from expanding to the full width of the window.

Don't use "gtk_widget_show()" on every single widget in the progress
dialog box, use "gtk_widget_show_all()" on the top-level widget.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2114 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoTo make windows modal, use "gtk_window_set_modal()" rather than
guy [Wed, 5 Jul 2000 02:52:39 +0000 (02:52 +0000)]
To make windows modal, use "gtk_window_set_modal()" rather than
"gtk_grab_add()"; the former makes it a bit clearer what's being done,
and I think it may be considered the right way to do it (GTK+ remembers
the state of the window and appears to add and remove the grab as
appropriate).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2113 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd a "dlg_window_new()" routine, which creates a GTK_WINDOW_DIALOG
guy [Wed, 5 Jul 2000 02:45:42 +0000 (02:45 +0000)]
Add a "dlg_window_new()" routine, which creates a GTK_WINDOW_DIALOG
window and makes it transient for the top-level window; the
transient-for at least provides a hint to X window managers to

minimize the dialog if the main window is minimized;

keep the dialog on top of the main window in the Z order for
windows;

perhaps (if there are any window managers that actually *do*
this) even put it atop the main window in the X-Y plane (KWM
doesn't and I seem to remember that the Exceed X server for
Windows doesn't).

It's generally considered the Right Thing To Do for dialog boxes.

Use that routine to create dialog boxes, rather than doing it directly
in the code for that dialog box.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2112 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoSet the locale for Tethereal to the native environment; Ethereal already
guy [Wed, 5 Jul 2000 02:06:58 +0000 (02:06 +0000)]
Set the locale for Tethereal to the native environment; Ethereal already
does so, as a side-effect of calling "gtk_set_locale()".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2111 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoNote, in a comment, that "gtk_set_locale()" sets the C-language locale.
guy [Wed, 5 Jul 2000 02:04:16 +0000 (02:04 +0000)]
Note, in a comment, that "gtk_set_locale()" sets the C-language locale.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2110 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoChange version to 0.8.10
gram [Tue, 4 Jul 2000 12:10:06 +0000 (12:10 +0000)]
Change version to 0.8.10
In gtk/Makefile.nmake, add progress_dlg.obj

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2108 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoMore news for 0.8.10
gram [Tue, 4 Jul 2000 12:04:45 +0000 (12:04 +0000)]
More news for 0.8.10

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2107 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd a missing build dependency (debhelper) and change build dependency on
oabad [Mon, 3 Jul 2000 22:40:13 +0000 (22:40 +0000)]
Add a missing build dependency (debhelper) and change build dependency on
libsnmp4.0-dev to libsnmp4.1-dev (this should close bugs #66391 and #66018
from the Debian bug tracking system).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2106 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoGtkLabel is a subclass of GtkMisc, so it has alignment and padding
guy [Mon, 3 Jul 2000 19:42:36 +0000 (19:42 +0000)]
GtkLabel is a subclass of GtkMisc, so it has alignment and padding
parameters and doesn't have to be stuffed into a GtkAlignment to align
it on the left.  (Alas, GtkButton *isn't* a subclass of GtkMisc, so we
have to stuff the Cancel button into an alignment to keep it from
growing to the width of the window.)

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2105 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIf the user clicks the "Cancel" button in a progress-bar dialog box for
guy [Mon, 3 Jul 2000 09:34:27 +0000 (09:34 +0000)]
If the user clicks the "Cancel" button in a progress-bar dialog box for
reading a capture file, we should just clean up and return so that
Ethereal continues running with no capture file loaded, unless the read
is being done as a result of Ethereal being run with the "-r" flag, in
which case we still exit (although we may eventually choose to continue
running with no capture file loaded even in that case).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2104 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoRemove the progress bar from the status line, and, instead, for any
guy [Mon, 3 Jul 2000 08:36:52 +0000 (08:36 +0000)]
Remove the progress bar from the status line, and, instead, for any
potentially long-running operation that has a progress indicator, pop up
a modal dialog box with

an indication of what is being done;

a progress bar;

a "Cancel" button to stop the operation.

This:

leaves more room on the status line for a filter expression;

provides a mechanism to allow the user to cancel long-running
operations (although the way we do so may not back out of them
as nicely as the user might like, if it's not obvious what the
"right" way is or if the "right" way is difficult to implement
or involves doing as much work as letting the operation
continue);

means that, because the dialog box is modal, we don't have to
worry about the user performing arbitrary UI operations out from
under the operation and changing arbitrary bits of state being
used by that operation.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2103 f5534014-38df-0310-8fa8-9805f1628bb7

19 years ago"asn1_octet_string_decode()" sets the pointer to the raw octets to NULL
guy [Sun, 2 Jul 2000 07:10:16 +0000 (07:10 +0000)]
"asn1_octet_string_decode()" sets the pointer to the raw octets to NULL
if the octet string is zero-length; if the octets are to be interpreted
as a character string, we have to check for a null pointer and replace
it with a pointer to a null string.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2102 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoFix the comment at the beginning of the routine to give the right file
guy [Sun, 2 Jul 2000 03:25:48 +0000 (03:25 +0000)]
Fix the comment at the beginning of the routine to give the right file
name, and add the RFC number for ISAKMP to it.

Fix the structures for headers to treat all multi-byte quantities as
arrays of bytes; the certificate request header needs to be done that
way, so that it's *not* padded to a multiple of the size of the longest
integral element (the longest integral element is 2 bytes, but the size
of the header is 5 bytes, so certificate requests were being
misdissected), and other structures might require this as well - it also
catches code that doesn't use "pntohl()" or "pntohs()" to get the values
of integral fields bigger than 1 byte.

In fact, it *did* catch some; the configuration attribute header's
"identifier" field was being used without being put in host byte order.
In addition, that field required padding to be aligned on its natural
boundary, but the padding wasn't given as a structure member; given that
the length field was specified as one byte but extracted with
"pntohs()", and that the length field is 2 bytes in the other headers,
it probably should be 2 bytes in this header as well - make it so.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2101 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd support for OSI CLTP.
guy [Sat, 1 Jul 2000 08:55:28 +0000 (08:55 +0000)]
Add support for OSI CLTP.

Redo the way the variable part of COTP and CLTP packets are handled, by
having a single routine to handle it for all packets, rather than having
the dissector for each type of packet handle the parameters those
packets are supposed to have.

Fix "is_LI_NORMAL_AK()" to test (heuristically) whether an AK COTP
packet is normal rather than extended, rather than testing whether it's
normal, and invert the test where it's used (i.e. make what it does
correspond to what its name says, which means we invert the test it
does, and then change the code that uses it appropriately).

Add support for the "preferred maximum TPDU size" and "inactivity timer"
parameters, and fix various problems in the dissection of particular
parameters.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2100 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoMake "format_oid()" allocate a buffer that's big enough, format the OID
guy [Thu, 29 Jun 2000 09:37:02 +0000 (09:37 +0000)]
Make "format_oid()" allocate a buffer that's big enough, format the OID
string into it, and return a pointer to that buffer, rather than being
passed a buffer.  This gets rid of some additional fixed-length string
buffers (and makes it more closely resemble the "format_oid()" in the
libsmi-based "packet-snmp.c" under development).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2099 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoMake "format_oid()" allocate a buffer that's big enough, format the OID
guy [Wed, 28 Jun 2000 05:15:13 +0000 (05:15 +0000)]
Make "format_oid()" allocate a buffer that's big enough, format the OID
string into it, and return a pointer to that buffer, rather than being
passed a buffer.  This gets rid of some additional fixed-length string
buffers (and makes it more closely resemble the "format_oid()" in the
libsmi-based "packet-snmp.c" under development).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2098 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoDon't reposition the random stream in "ngsniffer_open()" if there isn't
guy [Wed, 28 Jun 2000 03:58:52 +0000 (03:58 +0000)]
Don't reposition the random stream in "ngsniffer_open()" if there isn't
a random stream open.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2097 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoDon't pass a pointer to the packet data buffer to "read_packet()"; let
guy [Tue, 27 Jun 2000 09:26:10 +0000 (09:26 +0000)]
Don't pass a pointer to the packet data buffer to "read_packet()"; let
it, rather than its callers, get that pointer from "cf->wth".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2096 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd routines to Wiretap to allow a client of Wiretap to get:
guy [Tue, 27 Jun 2000 07:13:42 +0000 (07:13 +0000)]
Add routines to Wiretap to allow a client of Wiretap to get:

a pointer to the "wtap_pkthdr" structure for an open capture
file;

a pointer to the "wtap_pseudo_header" union for an open capture
file;

a pointer to the packet buffer for an open capture file;

so that a program using "wtap_read()" in a loop can get at those items.

Keep, in a "capture_file" structure, an indicator of whether:

no file is open;

a file is open, and being read;

a file is open, and is being read, but the user tried to quit
out of reading the file (e.g., by doing "File/Quit");

a file is open, and has been completely read.

Abort if we try to close a capture that's being read if the user hasn't
tried to quit out of the read.

Have "File/Quit" check if a file is being read; if so, just set the
state indicator to "user tried to quit out of it", so that the code
reading the file can do what's appropriate to clean up, rather than
closing the file out from under that code and causing crashes.

Have "read_cap_file()" read the capture file with a loop using
"wtap_read()", rather than by using "wtap_loop()"; have it check after
reading each packet whether the user tried to abort the read and, if so,
close the capture and return an indication that the read was aborted by
the user.  Otherwise, return an indication of whether the read
completely succeeded or failed in the middle (and, if it failed, return
the error code through a pointer).

Have "continue_tail_cap_file()" read the capture file with a loop using
"wtap_read()", rather than by using "wtap_loop()"; have it check after
reading each packet whether the user tried to abort the read and, if so,
quit the loop, and after the loop finishes (even if it read no packets),
return an indication that the read was aborted by the user if that
happened.  Otherwise, return an indication of whether the read
completely succeeded or failed in the middle (and, if it failed, return
the error code through a pointer).

Have "finish_tail_cap_file()" read the capture file with a loop using
"wtap_read()", rather than by using "wtap_loop()"; have it check after
reading each packet whether the user tried to abort the read and, if so,
quit the loop, and after the loop finishes (even if it read no packets),
close the capture and return an indication that the read was aborted by
the user if that happened.  Otherwise, return an indication of whether
the read completely succeeded or failed in the middle (and, if it
failed, return the error code through a pointer).

Have their callers check whether the read was aborted or not and, if it
was, bail out in the appropriate fashion (exit if it's reading a file
specified by "-r" on the command line; exit the main loop if it's
reading a file specified with File->Open; kill the capture child if it's
"continue_tail_cap_file()"; exit the main loop if it's
"finish_tail_cap_file()".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2095 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPull the code to fetch display options from the display options dialog
guy [Tue, 27 Jun 2000 05:18:44 +0000 (05:18 +0000)]
Pull the code to fetch display options from the display options dialog
box into a common subroutine.

Don't call "change_time_formats()" to update the display unless the time
format actually changed.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2094 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoGet rid of a debugging fprintf.
guy [Tue, 27 Jun 2000 04:40:15 +0000 (04:40 +0000)]
Get rid of a debugging fprintf.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2093 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPatch from Ben Fowler to rename the global variable "cf" to "cfile", to
guy [Tue, 27 Jun 2000 04:36:03 +0000 (04:36 +0000)]
Patch from Ben Fowler to rename the global variable "cf" to "cfile", to
make it easier to use grep to find all references to it without getting
a lot of false hits and to check, after allocating the memory chunk for
"frame_data" structures, that the allocation succeeded.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2092 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoRename "format_value()" to "format_var()", and have it dynamically
guy [Mon, 26 Jun 2000 00:13:21 +0000 (00:13 +0000)]
Rename "format_value()" to "format_var()", and have it dynamically
allocate the buffer into which it formats the variable value, and return
that value.  This

1) makes it more closely resemble the formatting routine in the
   libsmi-based "packet-snmp.c" under development;

2) makes it less likely to overrun the buffer (we can't be
   certain how long the string "sprint_value()" generates will
   be, but we can make a reasonable guess as to the maximum size
   based on the type and size of the object we're formatting).

When *not* using "-lsnmp", dynamically allocate the buffers into which
we format octet strings and OID values, based on the size of the object
we're formatting, so that we don't overrun the buffer.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2091 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIn "asn1_oid_value_decode()", set "size" directly from "enc_len", rather
guy [Mon, 26 Jun 2000 00:08:48 +0000 (00:08 +0000)]
In "asn1_oid_value_decode()", set "size" directly from "enc_len", rather
than setting "eoc" from "asn1->pointer" and "enc_len" and then setting
"size" from "eoc" and "asn1->pointer", to make it clearer how it's being
set.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2090 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoMake the string buffers into which stuff is formatted larger, to handle,
guy [Sun, 25 Jun 2000 20:55:09 +0000 (20:55 +0000)]
Make the string buffers into which stuff is formatted larger, to handle,
for example, larger octet string values.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2089 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoReport changes for 0.8.10
gram [Sat, 24 Jun 2000 16:30:44 +0000 (16:30 +0000)]
Report changes for 0.8.10

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2088 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoJoerg Mayer's changes to add support for NetXRay file format version
guy [Sat, 24 Jun 2000 05:32:48 +0000 (05:32 +0000)]
Joerg Mayer's changes to add support for NetXRay file format version
2.002, as used by release 3.50 of the Network Associates Sniffer for
Windows; currently, we treat it just like the 2.001 version, so we
rename the version #define WTAP_FILE_NETXRAY_2_001 to
WTAP_FILE_NETXRAY_2_00x and use that for both 2.001 and 2.002.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2087 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoWrite and read the temporary file for "Follow TCP Stream" as a binary
guy [Sat, 24 Jun 2000 05:06:29 +0000 (05:06 +0000)]
Write and read the temporary file for "Follow TCP Stream" as a binary
file, not as a text file; that makes no difference on UNIX, but, as the
file *is* binary, it makes a difference on Win32 systems.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2086 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd link instructions for "tvbtest", so that we link with GLib.
guy [Thu, 22 Jun 2000 06:37:59 +0000 (06:37 +0000)]
Add link instructions for "tvbtest", so that we link with GLib.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2085 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd support for the reported length of tvbuffs and for the
guy [Thu, 22 Jun 2000 06:36:45 +0000 (06:36 +0000)]
Add support for the reported length of tvbuffs and for the
ReportedBoundsError exception.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2084 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoFix the display of volume names.
guy [Wed, 21 Jun 2000 04:41:30 +0000 (04:41 +0000)]
Fix the display of volume names.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2083 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agotvbuffify the IGMP dissector. There's still plenty more to do inside
gram [Tue, 20 Jun 2000 13:21:55 +0000 (13:21 +0000)]
tvbuffify the IGMP dissector. There's still plenty more to do inside
packet-ip.c.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2082 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agotvbuff-ify the TR MAC dissector.
gram [Tue, 20 Jun 2000 03:05:37 +0000 (03:05 +0000)]
tvbuff-ify the TR MAC dissector.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2081 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoMake the "isis_hello.source_id" an FT_BYTES field rather than an
guy [Mon, 19 Jun 2000 08:33:50 +0000 (08:33 +0000)]
Make the "isis_hello.source_id" an FT_BYTES field rather than an
FT_ETHER field; the ISIS spec doesn't say it's necessarily a 6-byte
Ethernet address (and, if it's FT_BYTES, you can test it in a filter
much the same way you test an Ethernet address).

Make "isis_hello.lan_id" an FT_BYTES field rather than an FT_STRING
field - it's an array of bytes, not a character string.

Don't require that "system ID" fields be 6 octets; use the size value
from the ISIS PDU header.  (This means that PDUs containing "system ID"
fields can't be described as C structures; dissect them by stepping the
offset instead.)

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2080 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoISIS is ISO standard 10589, not 10598.
guy [Mon, 19 Jun 2000 04:49:36 +0000 (04:49 +0000)]
ISIS is ISO standard 10589, not 10598.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2079 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoGet rid of string overrun.
gerald [Sun, 18 Jun 2000 22:12:14 +0000 (22:12 +0000)]
Get rid of string overrun.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2078 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoInstead of calling "snmp_set_full_objid()" to cause OIDs to be displayed
guy [Sat, 17 Jun 2000 05:56:22 +0000 (05:56 +0000)]
Instead of calling "snmp_set_full_objid()" to cause OIDs to be displayed
symbolically as full sequences, call "snmp_set_suffix_only(2)" to cause
them to be displayed symbolically as a module name and a name within
that module, as that might make it easier to find the RFC or whatever
that describes the object in question.

Don't just statically call it, though, on Linux, as that causes binaries
built on Red Hat releases prior to 6.2 to fail to run on 6.2, due to the
UCD SNMP 4.1.1 library used in RH 6.2 not being 100% binary-compatible
with the UCD SNMP libraries used in those prior releases.  Instead, on
Linux, try to "dlopen()" the "libsnmp.so" library and, if that succeeds,
try to find "snmp_set_suffix_only()" in that library - if that succeeds,
call it, otherwise try to find "ds_set_int()" in that library and, if
*that* succeeds, call it with the arguments that, in UCD SNMP 4.1.1, the
"snmp_set_suffix_only()" macro passes to it.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2077 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdd braces to squelch a warning from GCC 2.8.1 on Solaris 7/x86.
guy [Sat, 17 Jun 2000 03:05:02 +0000 (03:05 +0000)]
Add braces to squelch a warning from GCC 2.8.1 on Solaris 7/x86.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2076 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoSmall patch for the win32 makefile that improves the dependencies and
gram [Fri, 16 Jun 2000 15:05:26 +0000 (15:05 +0000)]
Small patch for the win32 makefile that improves the dependencies and
prints out exactly which executable is being linked.
From Graham Bloice <graham.bloice@trihedral.com>

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2075 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoChanges from Graham Bloice to support "Update list of packets in real
guy [Thu, 15 Jun 2000 08:02:43 +0000 (08:02 +0000)]
Changes from Graham Bloice to support "Update list of packets in real
time" during captures on Win32.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2074 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPatch from Joerg Mayer:
guy [Thu, 15 Jun 2000 07:49:25 +0000 (07:49 +0000)]
Patch from Joerg Mayer:

Make "tethereal -h" look as similar to "ethereal -h" as
possible.  The ethereal help was the "better" of the two.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2073 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPatch from Joerg Mayer to fix a problem reading Sniffer files:
guy [Thu, 15 Jun 2000 06:18:32 +0000 (06:18 +0000)]
Patch from Joerg Mayer to fix a problem reading Sniffer files:

Differentiate between LAPB and LAPD sync sniffer traces.
Personally I think there must be a better way to find out which
protocol is in the trace but I currently lack the time to look
at the remaining frame info.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2072 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoPatch from Joerg Mayer to fix a problem reading Sniffer files:
guy [Thu, 15 Jun 2000 06:13:08 +0000 (06:13 +0000)]
Patch from Joerg Mayer to fix a problem reading Sniffer files:

When trying to decode a sample trace from the NG offline sniffer
installation, one trace resulted in a "corrupted" error.  The
reason was, that the file was a version 2 file format.  That
format used type 8 for header purposes while version 4 uses it
for FRAME4.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2071 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoMark Clayton's patch to add support for capturing on ATM interfaces on
guy [Thu, 15 Jun 2000 04:23:06 +0000 (04:23 +0000)]
Mark Clayton's patch to add support for capturing on ATM interfaces on
Linux (call the until-now-unused "capture_clip()" routine for each
packet).

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2070 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoFix typos in some printf formats - "+%d", not "%+d"
guy [Thu, 15 Jun 2000 04:09:22 +0000 (04:09 +0000)]
Fix typos in some printf formats - "+%d", not "%+d"

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2069 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoConvert IPX-and-friend dissectors in packet-ipx.c to use
gram [Thu, 15 Jun 2000 03:49:00 +0000 (03:49 +0000)]
Convert IPX-and-friend dissectors in packet-ipx.c to use
tvbuffs.

In doing so, I realied that my recommendation for using
tvb_new_subset(pi.compat_top_tvb, -1, -1) was incorrect, because
some dissectors (ethernet!) change pi.len and pi.cap_len. So, I have
to take those two variables into account instead of using -1 and -1.

So, I provide a macro called tvb_create_from_top(offset), where
offset is the name of your offset variable. It is a wrapper around
tvb_new_subset().

I converted the lines that followed my suggestion to use
tvb_create_from_top().

In proto.c I added
proto_tree_add_debug_text(proto_tree*, const char*, ...)
It's much like proto_tree_add_text(), except that it takes no offset
or length; it's soley for temporarily putting debug text into the
proto_tree while debugging a dissector. In making sure that its
use is temporary, the funciton also prints the debug string to stdout
to remind the programmer that the debug code needs to be removed
before shipping the code.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2068 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoAdded a passel-o-codes from Gaetan Soltesz <gaetan@cisco.com>
gerald [Wed, 14 Jun 2000 19:53:37 +0000 (19:53 +0000)]
Added a passel-o-codes from Gaetan Soltesz <gaetan@cisco.com>

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2067 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoWhen building under Win32, "boolean" appears to be defined in a fashion
guy [Wed, 14 Jun 2000 07:16:07 +0000 (07:16 +0000)]
When building under Win32, "boolean" appears to be defined in a fashion
that causes "packet-x11.c" to fail to compile due to it defining its own
function named "boolean", so we rename the "packet-x11.c" "boolean()"
function to "add_boolean()".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2066 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agoIn "stringCopy()", make the character temporary unsigned, so that it
guy [Wed, 14 Jun 2000 00:24:39 +0000 (00:24 +0000)]
In "stringCopy()", make the character temporary unsigned, so that it
doesn't get sign-extended before being handed to "isgraph()".

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2065 f5534014-38df-0310-8fa8-9805f1628bb7

19 years agotos_str is no longer used. it was superseded by iptos_vals.
itojun [Tue, 13 Jun 2000 10:37:24 +0000 (10:37 +0000)]
tos_str is no longer used.  it was superseded by iptos_vals.

git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@2064 f5534014-38df-0310-8fa8-9805f1628bb7