add an expert item. For unusable lengths throw ReportedBoundsError but
try to continue on otherwise.
Based on a patch from Mike Morrin in bug 3884.
svn path=/trunk/; revision=44439
Also (for a few files):
- create/use some extended value strings;
- remove unneeded #include files;
- remove unneeded variable initialization;
- re-order fcns slightly so prefs_reg_handoff...() at end, etc
svn path=/trunk/; revision=44438
- untermintated range-string array can/will cause crash
(e.g. 'tshark -G values');
- remove all 'if (tree)':
calling col_...() and sub-dissectors under 'if (tree)' not OK;
- simplify code for heuristic test;
- find_dissector() calls need to be done only once
(not on each prefs callback);
- do some whitespace/indentation changes.
svn path=/trunk/; revision=44436
lets us pass a NULL pinfo to expert_add_info_format() and
expert_add_undecoded_item(), which makes it possible to use those
routines deep in the bowels of many dissectors. As a proof of concept
remove the recent pinfo additions to packet-afp.c. This should also make
it easier to fix bug 3884.
svn path=/trunk/; revision=44435
- Add svn:keywords and svn:eol-style properties;
- col_...() and expert...() shouldn't be done under 'if (tree)';
- #if 0 an unused value-string;
- Use val_to_str_const as appropriate;
- Do some reformatting (whitespace changes).
svn path=/trunk/; revision=44426
RLC: changed so reassembly fail flag is set per channel instead of globally.
RLC: added reset channel function
Added support for reconfigurin HSDSCH flows, and E-DCH flows (just needs to figure out when it should become active for 'overloaded' ports), also changed slightly how ehs headers are configured in nbap.
Removed extra (debugging) port in nbap
This is the last patch set for a while.
svn path=/trunk/; revision=44423
- remove 'if (tree)'
* col_...() shouldn't be called under 'if (tree)';
* new-style dissector should alwyas return same 'bytes dissected'
(independent of 'if (tree)');
- create/use extended value string;
- Use consistent indentation;
svn path=/trunk/; revision=44414
- put back return NULL in line 947 that got removed in r44384: it is mandatory to avoid accessing non initialized variables. Change rlc_channel_assign a bit to fix what was the root cause of this removal (I guess)
- put back add_channel_info in dissect_rlc_am (removed by error?)
- fix some typo errors
- fix indentation a bit
svn path=/trunk/; revision=44393
- remove 'if (tree)' around calls to col_...() fcns and to sub-dissectors;
- call proto_get_id_by_filter_name() in proto_reg-handoff...() (not init routine);
- do some whitespace and formatting changes.
svn path=/trunk/; revision=44385
Add dissection of sub-options in DNS OPT RRs (RFC 2671)
From me
* use/prefer proto_tree_add_item
* Remove dns_optcode_description function (Use value_string...)
svn path=/trunk/; revision=44383
packet-classicstun.c
packet-reload-framing.c (probably)
packet-reload.c
packet-sccp.c
packet-sua.c
packet-tcp.c
packet-xmcp.c
\epan\gcp.c
The following files unnecessarily recreate keys because of the previously destructive nature of emem_tree_*32_array functions:
packet-btl2cap.c
packet-nfs.c
packet-rpc.c
packet-scsi-osd.c
packet-stun.c (per Bug 7569)
These could be cleaned up, but it's not like the key recreation is burning CPU cycles.
svn path=/trunk/; revision=44380
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=7563 :
Previous patch solved the reported problem. This patch should cover input
validation for the VC Id and Link Rate fields properly.
Note link_rate==0 (unknown) is supported, as some hardware/firmware may not set
this field, but vc_size==0 is defined invalid.
From me: also initialize m_sdh_line_rate when the input is garbage.
svn path=/trunk/; revision=44377
FP: fixed exotic bug where fakes counter was sometimes not reset, RRC: fixed bug where num_chans_per_flow was not properly cleared
svn path=/trunk/; revision=44375