This solves a crash occurring when trying to free memory allocated by
GeoIP (cross-compiled with mingw(32|64)) with MSVC function
Bug: 13598
Change-Id: I757cff13660bd485d7ea91d10660e9bf86404728
Reviewed-on: https://code.wireshark.org/review/21090
Petri-Dish: Pascal Quantin <pascal.quantin@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Guy Harris <guy@alum.mit.edu>
The -Wimplicit error covers two warnings, trying to catch these cases:
Setting a global variable without declaring its type (-Wimplicit-int):
undeclared_type = 1;
More importantly, -Wimplicit-function-declaration catches the case where
a function is not declared (missing header, programming error, etc.).
Turn these warnings into errors, most likely it will be a programming
error that results in a link failure anyway. See also
https://fedoraproject.org/wiki/Changes/Fedora26CFlags
Also fix autotools checks not to trigger -Wimplicit-int and
-Wimplicit-function-declaration (in krb5 check due to missing include).
Tested on Ubuntu 16.04 (autotools & cmake) and Arch Linux (cmake),
configure/cmake output and config.h are identical.
Change-Id: I137284263f3b1223df6e6a893111c3640802631f
Reviewed-on: https://code.wireshark.org/review/19331
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
There is *NO* guarantee that sed will do anything useful with input that
doesn't have a newline and, in fact, the default sed in Solaris 11
doesn't produce *any* output if you hand it input with no newline.
Furthermore, doing xxx=`yyy` will, if the output of yyy has newlines at
the end, strip them; to quote section 2.6.3 "Command Substitution" of
the Single UNIX Specification:
The shell shall expand the command substitution by executing
command in a subshell environment (see Shell Execution
Environment) and replacing the command substitution (the text of
command plus the enclosing "$()" or backquotes) with the
standard output of the command, removing sequences of one or
more <newline> characters at the end of the substitution.
so there's no need to strip the newline.
Change-Id: Ia710f67a42739c1e218eb9fec53a54bde6e010da
Reviewed-on: https://code.wireshark.org/review/19016
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Guy Harris <guy@alum.mit.edu>
We do not use the STARTUP negotiation since a stream
can be captured in its middle but try to decompress
if the flag is present and fallback if it fails.
Change-Id: Iecbf49a45220b04be7808869c9884548eb1e7694
Signed-off-by: Benoît Canet <benoit@scylladb.com>
Reviewed-on: https://code.wireshark.org/review/17952
Reviewed-by: Pascal Quantin <pascal.quantin@gmail.com>
The qtchooser package in many (all?) distros installs executable symlinks to
qtchooser that can mislead AC_PROG_PATH if the corresponding Qt linguist
tools package is missing.
Bug: 12570
Change-Id: I98795a3a7b8a05214c56a0c7b50829ea3b2976fa
Reviewed-on: https://code.wireshark.org/review/16316
Petri-Dish: João Valverde <j@v6e.pt>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Adrian Simionov <daniel.simionov@gmail.com>
Reviewed-by: João Valverde <j@v6e.pt>
sshdump and ciscodump require libssh >= 0.6.0. Check for this version and
only build them if a sufficient version is present. (The other checks
stay; in particular, ssh_userauth_agent doesn't exist on Windows and
must be checked separately.)
Bug: 12507
Change-Id: Ibe42b0cbb597cd9c4aab51f44d48185940ad954b
Reviewed-on: https://code.wireshark.org/review/15882
Petri-Dish: Michael Mann <mmann78@netscape.net>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Dario Lombardo <lomato@gmail.com>
Tested-by: Dario Lombardo <lomato@gmail.com>
Reviewed-by: Michael Mann <mmann78@netscape.net>
Add SOCKET_LIBS and NSL_LIBS to global LIBS variables on platforms
where it is required.
Make configure checks for getaddrinfo/gethostname unconditional,
that is handled with #ifdefs if necessary.
Change-Id: Ia874038454fb9cf3bdbf8e6fd829f319e331837e
Reviewed-on: https://code.wireshark.org/review/14560
Reviewed-by: João Valverde <j@v6e.pt>
Let the compiler and linker handle the system default search path.
Generally give priority to /usr/local when doing "manual" searches.
Change-Id: I3bde7af1226305d94ddb4bc96cefe9ef91e26769
Reviewed-on: https://code.wireshark.org/review/14564
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: João Valverde <j@v6e.pt>
(When cross-compiling, this is *not* the compiler used to build wireshark itself).
Macro ax_prog_cc_for_build.m4 retrieved from http://git.savannah.gnu.org/gitweb/?p=autoconf-archive.git
Change-Id: I035059d9315e0081dcf1ffbafd9ce6d7ad46700c
Reviewed-on: https://code.wireshark.org/review/4503
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Eric Anderson <andersoe@cs.cmu.edu>
Reviewed-by: João Valverde <j@v6e.pt>
Use pkg-config if a zlib.pc file is available.
Remove the now redundant AC_TRY_LINK_FUNC test (there are no linker flags
for GTK+ here).
Change-Id: I7de744749eba7231ae0097b975144b76ffcf1bdb
Reviewed-on: https://code.wireshark.org/review/14263
Petri-Dish: João Valverde <j@v6e.pt>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: João Valverde <j@v6e.pt>
Use PKG_WIRESHARK_CHECK_SYSTEM_MODULES for consistency with other
system dependencies.
Give slightly more informative configure messages other than multiple
'checking for LUA'.
Check for headers using pkg-config to avoid undefined symbols in
config.h.
Change-Id: I8d3df9ccf6f718d7d055da4f713af60d46d89e33
Reviewed-on: https://code.wireshark.org/review/14173
Petri-Dish: João Valverde <j@v6e.pt>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: João Valverde <j@v6e.pt>
GNU coding standards recommend against it and automake is designed
around it.
This allows overriding the global build flags using AM_CFLAGS, etc.,
or per object flags, something that is difficult or impossible currently
because of automake precedence rules.
Change-Id: I3f1ea12e560af5a46b2f04b342b1882bbf123f12
Reviewed-on: https://code.wireshark.org/review/13455
Petri-Dish: João Valverde <j@v6e.pt>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: João Valverde <j@v6e.pt>
This reverts commit 67c5849744.
It's not *guaranteed* to be there in *every* UN*X, but it's in at least
the Single UNIX Specification V3, as well as in shells commonly used in
non-UNIX UN*Xes :-), so it'll be there in the UN*Xes we'll be run on.
Change-Id: I541f7607055a24d6933d10244f85eea60052a3d8
Reviewed-on: https://code.wireshark.org/review/13325
Reviewed-by: Guy Harris <guy@alum.mit.edu>
Fixes false "-Wvariadic-macros" because "-Wpedantic" came after
"-Wno-variadic-macros". While at it, avoid (unintentionally?) adding
-fPIE to all C++ programs (via CXXFLAGS).
Availability of flags is checked by appending a flag (in case the user
has something like CFLAGS=-Wno-error).
This removes a -fPIE check for CXXFLAGS and removes 5
-Woverlength-strings and 9 -Wvariadic-macros warnings from make with
Clang 3.7.0. configure and compile times were equal. (A diff between the
configure outputs showed no other changes.)
This reverts commit cf0d762d73 and applies
a different approach.
Ping-Bug: 10791
Change-Id: Ic7b4137e2d98b06bc7625091be9bc7dd69182586
Reviewed-on: https://code.wireshark.org/review/12175
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
If we didn't find a C++ compiler, we can't check whether the
non-existent C++ compiler supports a flag, so don't do so.
Change-Id: I3d3232acae2dfc40deb0b01f35656ef53c4f1640
Reviewed-on: https://code.wireshark.org/review/12132
Reviewed-by: Guy Harris <guy@alum.mit.edu>
It's an ancient obsolete option with a confusing name.
Change-Id: Ib10330cf859cdea18fed2077c6539e56350ef380
Reviewed-on: https://code.wireshark.org/review/11967
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Michael Mann <mmann78@netscape.net>
Quick fix for OSX build, AS_ECHO_N is not available for that autoconf
version.
Change-Id: I24a93622d15388ef84c0f099d1209052bb35856d
Reviewed-on: https://code.wireshark.org/review/12013
Reviewed-by: Peter Wu <peter@lekensteyn.nl>
Use -isystem instead of -I for external headers with GCC/clang to squash
all the noise.
cmake already uses -isystem by default for supported platforms/compilers.
Change-Id: Ia6c9d1eb9b894fda6f48c531094d792e16fd39fc
Reviewed-on: https://code.wireshark.org/review/11947
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
sshdump is an extcap module that allows dumping from a remote host using an ssh connection.
It goes with the existing extcap plugin interface.
Change-Id: I8987614fdd817b8173a50130812bc643a4833bca
Reviewed-on: https://code.wireshark.org/review/11402
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Anders Broman <a.broman58@gmail.com>
Note the "initial". This is woefully incomplete. See the "to do" lists
below and in the code.
This differs a bit from the GTK+ version in that you specify one or more
streams to be decoded.
Instead of showing waveforms in individual widgets, add them all to a
single QCustomPlot. This conserves screen real estate and lets us more
easily take advantage of the QCP API. It also looks better IMHO.
Change a bunch of checks for QtMultimediaWidgets to QtMultimedia. We
probably won't use the widgets until we make 5.0 our minimum Qt
version and plain old QtMultimedia lets us support Qt 4 more easily
(in theory at least).
Add resampling code from libspeex. I initially used this to resample
each packet to match the preferred rate of our output device, but this
resulted in poorer audio quality than expected. Leave it in and use to
create visual samples for QCP and to match rates any time the rate
changes. The latter is currently untested.
Add some debugging macros.
Note that both the RTP player and RTP analysis dialogs decode audio data
using different code.
Note that voip_calls_packet and voip_calls_init_tap appear to be dead
code.
To do:
- Add silence frames where needed.
- Implement the jitter buffer.
- Implement the playback timing controls.
- Tapping / scanning streams might be too slow.
Change-Id: I20dd3b66d3df53c9b1f3501262dc01458849f6b4
Bug: 9007
Reviewed-on: https://code.wireshark.org/review/10458
Petri-Dish: Gerald Combs <gerald@wireshark.org>
Reviewed-by: Gerald Combs <gerald@wireshark.org>
Bug: 10791
Change-Id: I58c35c757039e69111a39100f5ccb306e098d591
Reviewed-on: https://code.wireshark.org/review/10519
Petri-Dish: Michael Mann <mmann78@netscape.net>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Guy Harris <guy@alum.mit.edu>
Reviewed-by: Michael Mann <mmann78@netscape.net>
On many systems, Qt5 requires that code compiled against it must compile
with -fPIC. The preferred place to determine whether this is necessary
is at configure time. This change adds the auto-tool code to make that
happen and removes the hack put in place as an interim solution.
Bug: 11230
Change-Id: I6e583c67bb67d168a30c0af05a8cd0a070522ba0
Reviewed-on: https://code.wireshark.org/review/10281
Petri-Dish: Jeff Morriss <jeff.morriss.ws@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Jeff Morriss <jeff.morriss.ws@gmail.com>
That's one of the mechanisms the autotools documentation suggests using
to handle generated header files.
Using it means that "make dist" will no longer try to build those files
(they're not part of the distribution - and they can't be, as the files
would be different for Qt 4 and Qt 5, and need to be generated by the
uic from the same version of Qt as the one against which Wireshark is
being built).
This means we don't need to try to find uic, moc, or rcc if we're not
building with Qt, so don't do so.
Change-Id: Id2aadb8289598b82e14e4ed402ff8cdc15fdef74
Reviewed-on: https://code.wireshark.org/review/9583
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Guy Harris <guy@alum.mit.edu>
For example, Qt 4's uic produces .h files that don't compile with Qt 5,
as they use header #include paths that work with Qt 4's headers but not
Qt 5's headers.
Change-Id: I50c7bd15fca05475180a933a6c77955dc686c0c5
Reviewed-on: https://code.wireshark.org/review/9567
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Guy Harris <guy@alum.mit.edu>
If you don't specify --with-qt or --without-qt, we build with Qt iff we
find it; we don't fail if we don't find it.
If you specify --with-qt, we build with Qt if we find it and fail if we
don't find it.
If you specify --without-qt, we don't look for Qt and don't build with
it.
This is all independent of --with-gtk2 or --with-gtk3.
Change-Id: I508d3281192bda9168fc46aba6011687c83ef818
Reviewed-on: https://code.wireshark.org/review/8861
Reviewed-by: Guy Harris <guy@alum.mit.edu>
First, search for packages with the version number without the period (bug
11219).
Second, don't look for Lua 5.3 because we don't work with it. If what we find
(without pkg-config's help) is Lua 5.3, disable Lua support (bug 10881).
Cmake support by Peter Wu (originally Ie73e5b53640f10432881a9671c0a605f7f027ed8):
Note the check for "lua<=5.2.99" instead of "lua<5.3" since cmake does not
support the latter syntax. Tested with lua5.2, lua5.1 and lua (5.3) installed.
Bug: 11219
Ping-Bug: 10881
Change-Id: I382d07ca00eafc6111cd4e9faa2b66f6b8f95b6e
Reviewed-on: https://code.wireshark.org/review/8783
Petri-Dish: Jeff Morriss <jeff.morriss.ws@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
The TOOL variant will automatically search for host prefixed scripts
(e.g. armv7a-unknown-linux-gnueabi-pcap-config) before falling back
to the default (i.e. pcap-config). This makes cross-compiling a bit
simpler.
Change-Id: Ia450839693b5550f798634a7e8b82c2a661b088c
Reviewed-on: https://code.wireshark.org/review/8481
Petri-Dish: Alexis La Goutte <alexis.lagoutte@gmail.com>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Alexis La Goutte <alexis.lagoutte@gmail.com>
Without a major version number, it behaves as before, picking whatever
version it finds, and preferring Qt 5 to Qt 4. With a major version
number, it looks only for the version in question.
Bug: 10793
Change-Id: Idf6c2c61e84bb87f7b601d8f09c33f31b67bf46d
Reviewed-on: https://code.wireshark.org/review/8052
Petri-Dish: Guy Harris <guy@alum.mit.edu>
Tested-by: Petri Dish Buildbot <buildbot-no-reply@wireshark.org>
Reviewed-by: Guy Harris <guy@alum.mit.edu>
Otherwise, just because a flag is appropriate for $CC, that doesn't
necessarily mean it's appropriate for $CFLAGS_FOR_BUILD.
(We don't use CFLAGS_FOR_BUILD for many things, so it's probably not
worth making a lot of effort to throw all the warning flags into it if
we're cross-compiling; we *do* throw them in for native compilation,
which is what most developers use and test with, so that should be
sufficient to let the warnings catch problems with the build tools we're
building.)
Change-Id: Iad9d611b4687e9e154f9871f741f3c8f3b307c88
Reviewed-on: https://code.wireshark.org/review/8026
Reviewed-by: Guy Harris <guy@alum.mit.edu>
It can be set if either 1) this is Windows (where we're assumed to be
using WinPcap, which includes calls to set the buffer size) or 2) we
have pcap_create() (in which case we also have pcap_set_buffer_size(),
at least in a normal libpcap release).
Use that rather than testing "defined(_WIN32) ||
defined(HAVE_PCAP_CREATE)"; that makes it a bit more obvious what's
being tested.
Change-Id: Id9f8455019d19206b04dd6820a748cb97ae5ad12
Reviewed-on: https://code.wireshark.org/review/7816
Reviewed-by: Guy Harris <guy@alum.mit.edu>
Pascal Quantin