osmocom/wireshark
14
0
Fork 1
Code Issues Pull requests Projects Releases Wiki Activity
31276 commits 26 branches 0 tags 795 MiB
Commit graph

383 commits

Author SHA1 Message Date
Bill Meier
6812b68eb1 From Yaniv Kaul: constify parameters 
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=4422

 From me: Fix a number of instances where the function prototype or
  the function definition wasn't changed so there was a mismatch 
  thus causing Windows (but not gcc) compilation errors.

svn path=/trunk/; revision=32365
 2010-04-03 18:18:50 +00:00
Bill Meier
d32b4c0758 Revert SVN #32360 until Windows compilation errors corrected. 
svn path=/trunk/; revision=32361
 2010-04-02 15:18:03 +00:00
Bill Meier
049f9eac85 From Yaniv Kaul: constify parameters 
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=4422

svn path=/trunk/; revision=32360
 2010-04-02 14:37:49 +00:00
Ronnie Sahlberg
4f3db1a1a8 Change printing of all BASE_OCT values to use the alternate form 
%#o instead of %o

This means that the value will be printed with a leading 0, which is the 
standard way to denominate that a value is in base octal.



svn path=/trunk/; revision=32241
 2010-03-19 02:05:18 +00:00
Jaap Keuter
7be36221d8 Fix for bug 4563: 
Always have a tvb when adding a tree item.

svn path=/trunk/; revision=32146
 2010-03-08 21:04:47 +00:00
Guy Harris
500eb99cd9 Add a third date format, ABSOLUTE_TIME_DOY_UTC, to show UTC with the 
date as YYYY/DDD, where DDD is a 1-origin day of year.  Move the formats
to a "time_fmt.h" file, included by the headers that use it.  Have
abs_time_to_str() and abs_time_secs_to_str() take the date format value,
rather than a Boolean "show this as UTC" flag, as an argument.  Document
the ABSOLUTE_TIME_ formats a bit better.  Use that format in the CCSDS
and VCDU dissectors, rather than having those dissectors do the
formatting themselves.

svn path=/trunk/; revision=32034
 2010-02-27 19:01:27 +00:00
Gerasimos Dimitriadis
6fd1cec787 From Jakub Zawadzki: 
Remove prepare_bits_string from proto.c; Use
equivalent function decode_bits_in_field in to_str.c

svn path=/trunk/; revision=31665
 2010-01-25 18:26:47 +00:00
Gerasimos Dimitriadis
acd35de1f4 Minor simplification in length calculation of proto_tree_add_bits 
and _format_value

svn path=/trunk/; revision=31640
 2010-01-24 12:08:14 +00:00
Gerasimos Dimitriadis
383f6a65dc Add _format_value variants of the proto_tree_add_bits function 
for different header field types; now possible to add bits item
for FT_FLOAT and FT_INT types.

svn path=/trunk/; revision=31633
 2010-01-23 16:17:17 +00:00
Gerasimos Dimitriadis
88835212b1 Use true/false strings instead of actual numeric values when adding a boolean 
field as a bits item and no custom strings are defined.

svn path=/trunk/; revision=31375
 2009-12-28 21:20:27 +00:00
Guy Harris
8e236e55de For fields of type FT_ABSOLUTE_TIME, have the "display" value be one of 
ABSOLUTE_TIME_LOCAL or ABSOLUTE_TIME_UTC, indicating whether to display
the date/time in local time or UTC. (int)ABSOLUTE_TIME_LOCAL ==
(int)BASE_NONE, so there's no source or binary compatiblity issue,
although we might want to eliminate BASE_NONE at some point and have the
BASE_ values used with integral types start at 0, so that you can't
specify BASE_NONE for an integral field.

svn path=/trunk/; revision=31319
 2009-12-19 03:17:44 +00:00
Guy Harris
a4b7135ce3 Rename BASE_STRUCTURE_RESET to BASE_DISPLAY_E_MASK, to clarify that it's 
a mask to select the base_display_e value from a display field in a
header_field_info structure.

Never select that value by masking out the BASE_RANGE_STRING flag bit,
as that won't continue to work if more flag bits, or other bitfields,
are added.  Instead, mask with BASE_DISPLAY_E_MASK.

Note that the base_display_e value and BASE_RANGE_STRING flag are only
for integral field types, and clarify what BASE_DISPLAY_E_MASK is.

Give at least one of the reasons why hiding protocol fields is not
considered a good idea.

svn path=/trunk/; revision=31249
 2009-12-12 03:15:28 +00:00
Guy Harris
4fde145bd2 Add an argument to abs_time_to_str() and abs_time_secs_to_str() 
indicating whether the time should be shown as local time or UTC.  For
now, always pass FALSE, meaning "show as local time".

Clean up some stuff in the SNMP dissector, use abs_time_secs_to_str()
for times with one-second resolution, and update a comment in various
macros in the WSP dissector, while we're at it.

svn path=/trunk/; revision=31227
 2009-12-10 22:19:29 +00:00
Stig Bjørlykke
185488655e No need to do string compare for FT_PROTOCOL. 
svn path=/trunk/; revision=31018
 2009-11-19 16:12:08 +00:00
Stig Bjørlykke
a1c8cdcc7c Added support for some missing custom column types: 
FT_NONE: Print nothing
 FT_PROTOCOL: Print "Yes" if protocol exists in packet
 FT_IPv6: Print address

The changes in r29551 made wireshark crash for this columns.

svn path=/trunk/; revision=31016
 2009-11-19 13:29:57 +00:00
Stig Bjørlykke
b510b97ff1 Return FALSE from proto_tracking_interesting_fields() when no tree. 
svn path=/trunk/; revision=30533
 2009-10-12 08:19:30 +00:00
Stig Bjørlykke
ec95db36a7 Check for valid tree in proto_get_finfo_ptr_array() and 
proto_tracking_interesting_fields().

This fixes a crash when using Lua script without packet list coloring.

svn path=/trunk/; revision=30532
 2009-10-12 08:16:50 +00:00
Kovarththanan Rajaratnam
996e27dbc7 Set INITIAL_NUM_PROTOCOL_HFINFO to a more sensible value, e.g. 1500. Currently it is set to 200 even though we have way more protocols in Wireshark: 
$ ./tshark -G protocols | wc -l
1066

svn path=/trunk/; revision=30501
 2009-10-11 09:35:34 +00:00
Kovarththanan Rajaratnam
ee9ea561d2 As pointed out by Bill [1] we leak some header_field_info structs on shutdown. This change plugs that mem leak. 
[1] http://www.wireshark.org/lists/wireshark-dev/200910/msg00074.html

svn path=/trunk/; revision=30500
 2009-10-11 09:18:51 +00:00
Kovarththanan Rajaratnam
34e05c76f7 Replace g_malloc + memset with g_new0 
svn path=/trunk/; revision=30499
 2009-10-11 09:02:04 +00:00
Kovarththanan Rajaratnam
ca53b00b7e Set tree_is_expanded to NULL after freeing it 
svn path=/trunk/; revision=30498
 2009-10-11 08:57:08 +00:00
Kovarththanan Rajaratnam
b998cb04de Use the slice allocator if GLIB_CHECK_VERSION > (2,10,0) 
svn path=/trunk/; revision=30497
 2009-10-11 08:52:34 +00:00
Jaap Keuter
ee2043792f Tighten header field checks. 
svn path=/trunk/; revision=30410
 2009-10-08 21:26:26 +00:00
Kovarththanan Rajaratnam
e981de7a99 Don't attempt to fill custom columns if we don't have a valid column expression. 
svn path=/trunk/; revision=30407
 2009-10-08 16:35:58 +00:00
Stig Bjørlykke
5bc9611d56 Removed a redundant prototype. 
Made wrs_count_bitshift static.

svn path=/trunk/; revision=30387
 2009-10-07 08:37:18 +00:00
Bill Meier
bf85ecb308 Rename hfinfo.ref_count to hfinfo.ref_type since that's now how it's used. 
svn path=/trunk/; revision=30216
 2009-09-30 14:02:49 +00:00
Stig Bjørlykke
ad14865ce0 Removed unused variable 'hfinfo' in proto_tree_set_time(). 
svn path=/trunk/; revision=30197
 2009-09-29 19:16:52 +00:00
Kovarththanan Rajaratnam
778fe81f72 (Trivial) White space changes 
svn path=/trunk/; revision=30095
 2009-09-23 16:27:37 +00:00
Bill Meier
5bbec764db (Trivial) Rework a comment slightly. 
svn path=/trunk/; revision=30080
 2009-09-22 20:17:31 +00:00
Kovarththanan Rajaratnam
56b8d36e03 Introduce a few FT_XXX lengths defines and use those instead of hard coding constants 
svn path=/trunk/; revision=29846
 2009-09-10 18:22:33 +00:00
Gerald Combs
c238175acf Move the infinite loop check from proto_tree_add_node to 
TRY_TO_FAKE_THIS_ITEM. Partial fix for bug 3986.

svn path=/trunk/; revision=29811
 2009-09-08 22:56:55 +00:00
Kovarththanan Rajaratnam
d2db67480c Use g_strlcpy when possible 
svn path=/trunk/; revision=29808
 2009-09-08 20:06:52 +00:00
Anders Broman
c6d4c80371 From : Didier Gautheron 
speed up a little proto_custom_set.
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=3972

svn path=/trunk/; revision=29789
 2009-09-08 10:33:15 +00:00
Kovarththanan Rajaratnam
571e69fdfa We don't currently fake fields if they are referenced indirectly through some other fields (currently only FT_PROTOCOL's can be indirectly referenced) . However, there doesn't seem to be any reason for this restriction. We already have proto_tree_set_fake_protocols() to disable 'faking' if any clients requires this. So, this changelist removes this restriction on indirectly referenced fields. 
svn path=/trunk/; revision=29742
 2009-09-06 11:30:41 +00:00
Kovarththanan Rajaratnam
820ce62af5 This fix was supposed to be in r29731 which some how only contained whitespace changes. 
svn path=/trunk/; revision=29732
 2009-09-06 07:33:07 +00:00
Kovarththanan Rajaratnam
9c512f7cb4 Introduce TRY_TO_FAKE_THIS_REPR which is an optimization similar to the existing TRY_TO_FAKE_THIS_ITEM. The purpose of TRY_TO_FAKE_THIS_REPR is to avoid generating the string representation for the protocol item. There is no point in doing this if the protocol tree isn't visible. In this case the result will be discarded any way. 
svn path=/trunk/; revision=29731
 2009-09-06 07:24:54 +00:00
Kovarththanan Rajaratnam
4df25480b7 The frame dissector disables the TRY_TO_FAKE_THIS_ITEM() optimization by marking the protocol tree as permamently visible. It only needs to disable the optimization temporarily while it creates the protocol item it intends to use proto_item_append_string() on 
svn path=/trunk/; revision=29730
 2009-09-06 06:33:13 +00:00
Kovarththanan Rajaratnam
17f010119a From Jakub Zawadzki via. Bug 3330: 
* Fix memleak (df->deprecated in dfilter_free())
* Free protocol hash tables on cleanup.
* Free protocols list on cleanup.
* Free memory allocated by fgetline() in parse_services_file()

From me:

* proto.c: set gmc_hfinfo to NULL after free
* proto.c: switch order of g_free() and g_list_remove() in proto_cleanup()

svn path=/trunk/; revision=29656
 2009-09-01 18:16:55 +00:00
Kovarththanan Rajaratnam
8cba530c81 From Richard Brodie via. Bug 3913: 
Free ptr array allocated through proto_find_finfo()

svn path=/trunk/; revision=29654
 2009-09-01 17:09:36 +00:00
Kovarththanan Rajaratnam
cd49063473 Revert r29614 until I figure out why buildbot doesn't like it 
svn path=/trunk/; revision=29615
 2009-08-29 07:53:48 +00:00
Kovarththanan Rajaratnam
cc948586bf Use G_STRINGIFY/G_PASTE 
svn path=/trunk/; revision=29614
 2009-08-29 06:42:23 +00:00
Jörg Mayer
4d8506486f After asking on the list and receiving no feedback, I'll 
just commit this:

Change the checks for type FT_IPv6 to no longer require
a lenght of exaclty 16 bytes, but something between
0 and 16 bytes. That way, we can filter on prefixes
that do not provide the whole length of 16 bytes.

svn path=/trunk/; revision=29594
 2009-08-28 09:38:34 +00:00
Anders Broman
32e7cddf70 From Didier Gautheron: 
move custom col set from  proto_tree_set_xxx() to
epan_dissect_fill_in_columns().
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=3930

I'v added the new functions to the .h files.

svn path=/trunk/; revision=29551
 2009-08-25 21:10:35 +00:00
Kovarththanan Rajaratnam
da1e555499 Don't dereference pointer before checking for NULL 
svn path=/trunk/; revision=29503
 2009-08-21 23:02:59 +00:00
Kovarththanan Rajaratnam
8a3f4f53eb Delay TRY_TO_FAKE_THIS_ITEM() in proto_tree_add_bits_ret_val() until we know it is safe. 
Thanks to Martin for pointing this out in r29493.

svn path=/trunk/; revision=29494
 2009-08-21 17:11:02 +00:00
Martin Mathieson
044aabb9ea Don't try TRY_TO_FAKE_THIS_ITEM in proto_tree_add_bits_ret_val(), as otherwise 
the side effect 'return_value' getting set doesn't happen.

svn path=/trunk/; revision=29493
 2009-08-21 15:16:27 +00:00
Kovarththanan Rajaratnam
27f7d88c84 Also fake empty field_info's by gracefully handling NULL field_info pointer elsewhere. 
svn path=/trunk/; revision=29490
 2009-08-21 11:03:30 +00:00
Kovarththanan Rajaratnam
0f6b75b7ea Move TRY_TO_FAKE_THIS_ITEM to all public proto_xxx methods. Currently we rely on our internal (static) helper functions to do the faking. However, when they fake a protocol tree, they usually return a valid tree (although a dummy tree it is still a valid one). This means that we end up calling proto_tree_set_representation + friends on it for no reason. By moving TRY_TO_FAKE_THIS_ITEM to the public proto_xxx methods we're able to bail immediately (when we're allowed to do so). 
svn path=/trunk/; revision=29480
 2009-08-19 19:08:34 +00:00
Kovarththanan Rajaratnam
460eacaae1 Move the NULL pointer 'tree' check under TRY_TO_FAKE_THIS_ITEMs umbrella 
svn path=/trunk/; revision=29453
 2009-08-17 11:21:33 +00:00
Kovarththanan Rajaratnam
69babcb4a6 Don't do PROTO_REGISTRAR_GET_NTH twice. 
svn path=/trunk/; revision=29452
 2009-08-17 11:13:36 +00:00