The linked blog entry for /Qspectre suggests intention to include
/Qspectre support for VS 2015 Update 3, but this has not happened yet
(at least, not with VS Community 2015 14.0.23107.178).
Change-Id: I001c8fa512457f1edc753b460634f13d2ff0ed7a
Fixes: v2.5.1rc0-611-g0ebcd27377 ("Trust CMake's Visual C++ version detection.")
Reviewed-on: https://code.wireshark.org/review/27225
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot
Reviewed-by: Gerald Combs <gerald@wireshark.org>
erf.c:2562:9: error: this statement may fall through [-Werror=implicit-fallthrough=]
Change-Id: Ib516a689e078a9e1eea96d692ffbbaab398f2bcb
Reviewed-on: https://code.wireshark.org/review/27271
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot
Reviewed-by: Guy Harris <guy@alum.mit.edu>
conversation_hash_tables_dialog.cpp:52:67: error: type qualifiers ignored on cast result type [-Werror=ignored-qualifiers]
Change-Id: I07c815a7233af41c840087a35eaf9172c06161f3
Reviewed-on: https://code.wireshark.org/review/27268
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot
Reviewed-by: Guy Harris <guy@alum.mit.edu>
extcap-base.c:293:43: error: cast between incompatible function types from ‘void (*)(void *)’ to ‘void (*)(void *, void *)’ [-Werror=cast-function-type]
extcap-base.c:302:45: error: cast between incompatible function types from ‘void (*)(void *)’ to ‘void (*)(void *, void *)’ [-Werror=cast-function-type]
extcap-base.c:321:42: error: cast between incompatible function types from ‘void (*)(void *)’ to ‘void (*)(void *, void *)’ [-Werror=cast-function-type]
Change-Id: Ifd9151d04412c5e29636dc14c57a327cc12d33f3
Reviewed-on: https://code.wireshark.org/review/27265
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Reviewed-by: Dario Lombardo <lomato@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Guy Harris <guy@alum.mit.edu>
The spec states that connSupervisionTimeout = Timeout * 10 ms
Change-Id: I89494c74d80b63c85f001540ea79850736457b21
Reviewed-on: https://code.wireshark.org/review/27255
Reviewed-by: Stig Bjørlykke <stig@bjorlykke.org>
The installer is not presenting applicable text in the description
box when highlighting the new file type and codec plugin options.
Change-Id: I25fbc7800a594796ac8f4f15b39431a89b404d0e
Signed-off-by: Jaap Keuter <jaap.keuter@xs4all.nl>
Reviewed-on: https://code.wireshark.org/review/27248
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Some preference modules did not exist, so importing would not work.
While vuze-dht used to have a configurable protocol name, it no longer
supports that, so remove it from the list.
Add some sanity checks to prevent bug 14316 from going undetected again,
and to ensure that dead code is removed when prefs are removed.
Change-Id: I5df809af66a6c19f9eb9a6b75d5e60c96008cde4
Fixes: v2.3.0rc0-971-g268841f3e0 ("Combine Decode As and port preferences for tcp.port dissector table.")
Reviewed-on: https://code.wireshark.org/review/27227
Petri-Dish: Anders Broman <a.broman58@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
They've been replaced by the Python scripts.
Change-Id: I8add9c9ea0a6bdd68b2fa3841977863c0ea9a761
Reviewed-on: https://code.wireshark.org/review/27243
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Prior to the switch from AsciiDoc to Asciidoctor we converted
release_notes.html to NEWS using elinks or lynx, which in turn generated
ASCII output. It was sufficient to read NEWS in PowerShell using
Get-Content, which defaults to ASCII.
We now use tools/html2text.py, which generates UTF-8. Switch Get-Content's
encoding to match. Note that Notepad detects file encodings heuristically,
and that we might want to use a BOM.
Bug: 14636
Change-Id: Ibd92ef7ad642631a938bb4d75a2d83f479099032
Reviewed-on: https://code.wireshark.org/review/27240
Reviewed-by: Gerald Combs <gerald@wireshark.org>
Petri-Dish: Gerald Combs <gerald@wireshark.org>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
If it's run as "tshark", it should run TShark, not Wireshark.
Bug: 14643
Change-Id: I0d4e6fa64e42b7a2e2d4b89b53db62748b4f288d
Reviewed-on: https://code.wireshark.org/review/27245
Reviewed-by: Guy Harris <guy@alum.mit.edu>
Convert capinfos output to UTF-8 in getCaptureInfo.
Normalize our command paths, otherwise "./run/RelWithDebInfo/..." might
be interpreted as the command "." with flags "/run", "/RelWithDebInfo",
etc. on Windows.
Change-Id: Ib7336a016db3ee0805739fc44913cb9c6895aaad
Reviewed-on: https://code.wireshark.org/review/27239
Petri-Dish: Gerald Combs <gerald@wireshark.org>
Tested-by: Petri Dish Buildbot
Reviewed-by: Gerald Combs <gerald@wireshark.org>
They were flipped. See https://wiki.wireshark.org/SampleCaptures,
Bluetooth_HCI_and_OBEX_Transaction_over_USB.ntar.gz, packets 153136/153140
for an example.
Change-Id: Iaac853fad16e97ff88ba38a7b4c5cbbdd13052b3
Reviewed-on: https://code.wireshark.org/review/27206
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Loading an old Wireshark profile with certain deprecated preferences
could result in a crash due to type confusion. If the new preference was
a range type, then four bytes of the pointer (address) to the range was
overwritten with the numeric value of the deprecated preference.
Minimal reproducer:
tshark -opgm.udp.encap_ucast_port:0 -r ../test/captures/empty.pcap
Bug: 14316
Change-Id: Ia8dc24f81f6b2e6494448dadffe810606765cb9e
Fixes: v2.3.0rc0-971-g268841f3e0 ("Combine Decode As and port preferences for tcp.port dissector table.")
Reviewed-on: https://code.wireshark.org/review/27226
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Let the Save dialog fixup the extension on accepting the dialog.
Otherwise it is possible that files are silently overwritten without
prompting. Additionally, if a user decides to save a pcapng file as
"foo.pcap", do not try to rename it to "foo.pcap.pcapng".
This change is limited to macOS and Linux because Windows uses a
different file dialog. Tested with both macOS and Linux.
Bug: 14600
Change-Id: Ie0bc1f579766a04f0aad96dcd5daba3fffef9764
Reviewed-on: https://code.wireshark.org/review/27188
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
The default Qt behavior for extension adjustment is quite bad. When the
file type filter is changed, the extension always becomes "gz" because
"pcap.gz" happens to be the first extension in the list. It also did not
check that the last suffix is actually a valid extension (e.g.
"capture.2018.01" became "capture.2018.gz").
Improvements:
- Respect the "compression" checkbox when adjusting the filename.
- Replace the extension only if it is a known one, append otherwise.
- Use a better default extension (from "wtap_default_file_extension").
Affects only macOS and Linux since Windows has its own native dialog.
See also https://bugreports.qt.io/browse/QTBUG-67993
Bug: 14600
Change-Id: I8cd0788f2abac0c6d7e29490b1ebb381f5a926d0
Reviewed-on: https://code.wireshark.org/review/27186
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Apple bsdpd uses the same routine to parse BSDP suboptions as it uses to parse
the DHCP options, which means that the "pad" (0) and "end" (255) options (as
described in RFC 2132) are also accepted as BSDP suboptions. Just like when
used as DHCP options, they do not follow the usual TLV template: They do not
have a length field and do not have any value, so they always consume exactly
one byte.
This change enhances the BSDP suboption dissector to accept the "pad" (0) and
"end" (255) suboptions, without any stored length or value.
Apple firmware/software does not issue BSDP "pad" or "end" suboptions, but will
tolerate them in received packets. At least one 3rd-party BSDP implementation
(the Dell KACE K2000 appliance) includes a BSDP "end" suboption in packets it
sends. Prior to this fix, function dissect_vendor_bsdp_suboption was expecting
a length for these suboptions, leading to dissection failing with error
"Suboption 255: no room left in option for suboption length".
For further discussion -- in which the exact same issue is found to affect
VMware virtual machine firmware -- refer to the VMware Communities forum thread
at https://communities.vmware.com/message/2459144#2459144 .
Interestingly, when Apple's bsdpd finds an "end" BSDP suboption, it simply
records that an "end" was encountered, and continues parsing until the whole of
the vendor options blob is consumed. The BSDP suboption dissector required no
modification to match that behavior.
Testing Done: Built Wireshark on Linux amd64. Loaded a BSDP ACK[LIST] from a
Dell KACE K2000 appliance; Previously it would issue an error about there
being insufficient room for the length of the "end" suboption, and now it
parses correctly. Modified the packet to include a string of "0" and "255"
suboptions, and observed that they were parsed as expected: One byte each,
no subtree, no length, and parsing continues afterwards. 200,000 iterations
with tools/fuzz-test.sh using the original BSDP packet, 4,000 of which were
under Valgrind.
Change-Id: I1786414b2ef0b8726d989a566d0e8a3525d516b8
Reviewed-on: https://code.wireshark.org/review/27210
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Move */ to a separate line below the SPDX identifier.
Change-Id: Id1032215449cfccae0933147b45e04b65e0b727f
Reviewed-on: https://code.wireshark.org/review/27211
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Move UAT file creation to config.py.
Run the text2pcap and some of the clopts tests under our default
environment.
Use "in" instead of "has_key".
Change-Id: Ie5c70fb33c29676672bed7bf8205cff0bba77f8a
Reviewed-on: https://code.wireshark.org/review/27234
Petri-Dish: Gerald Combs <gerald@wireshark.org>
Reviewed-by: Gerald Combs <gerald@wireshark.org>
dissect_ldss_transfer had a trivial read overrun: "line" was not
NUL-terminated, and strtol/g_ascii_strtoull will keep reading and discarding
any leading whitespace, so a malformed LDSS packet (with only whitespace
characters following the tag on a "Size:"/"Start:"/"Compression:" line) could
trigger a read overrun.
Let's replace the tvb_memdup with tvb_get_string_enc, which does some checking
of the input characters (which, it seems, must always be ASCII), and produces a
neat NUL-terminated string.
Testing Done: On Linux x64, ran "valgrind tshark -r fuzz-2018-04-23-14422.pcap"
without the fix (to reproduce the failure), and then with the fix, and
observed that no errors were reported anymore after the fix. 60,000 iters of
fuzz-test with ldss_filtered.pcap as input, plus 1,000 iters under valgrind.
Launched wireshark and opened ldss_filtered.pcap, and examined the dissection
of the "ldss and tcp" packets; All looks good.
Bug: 14615
Change-Id: I3fccc4ffbe315a3cff6ea03cc7db37f884b0582c
Reviewed-on: https://code.wireshark.org/review/27204
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
ENABLE_CHM_GUIDES is no longer in CMakeOptions.txt
Change-Id: I217ac89f12c95e66591465e3230c19968dcc0bde
Reviewed-on: https://code.wireshark.org/review/27209
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
Change-Id: I541bd728c159e95c2d5daa8ce0bfea3961ff1db9
Reviewed-on: https://code.wireshark.org/review/27203
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot
Reviewed-by: João Valverde <j@v6e.pt>
Gerald Combs