Remove the major.minor version from the plugin path, i.e:
lib/plugins/X.Y/{epan,wiretap,codecs}
and use an unversioned path:
lib/plugins/{epan,wiretap,codecs}
Introduce a new naming policy for plugins that requires
name.so.ABI_VERSION.
This is a simplified filesystem layoutfor plugins some
important benefits such as:
* improves compatibility between Wireshark versions, because
a plugin that wasn't recompiled will be automatically picked
up, but only if it has a compatible ABI version in the file name.
* does not clash with Apple guidelines
* simpler for users to understand and apply
* just overall simpler and easier to maintain, removes a lot
of complexity from CMake code
It does impose more requirements on the plugin naming scheme
but this should be handled completely transparently
by the build system.
It would also be possible to add support for unversioned *.so file
extensions at the same time, although in ths case it is not possible
to support multiple Wireshark ABI versions with only *.so, of course.
This wasn't done here but it may or may not be a useful enhancement
in the future.
Follow-up to 90b16b4092.
System implementations of strcasestr are locale dependent; even
though we always force to a UTF-8 locale, the Turkic languages
produce a different result, as the upper case of i is U+0130,
Latin Capital Letter I With Dot Above, and the lower case of I
is is U+0131, Latin Small Letter Dotless i.
For the purposes of dissectors we want the locale-independent ASCII
search. (There may be exceptions in some user-facing GUI code.)
FreeBSD, NetBSD, and macOS (but not OpenBSD currently) provide
strcasestr_l, which could be used with the NULL (C) locale to
produce the desired result.
This replaces code in both absolute_time_string(), in capinfos, and
set_abs_ymd_time(), in libwireshark. Like absolute_time_string(), it
reports, in the formatted string, an error if the seconds value in the
nstime can't be converted to a struct tm (looking at *you*, Visual
Studio C library),
Have that routine handle all time precision values from 0 (seconds) to 9
(nanoseconds).
Add a ws_gmtime_r(), matching ws_localtime_r(), for use by that routine.
Replace our strptime code, which is from gnulib,
with the simpler and better NetBSD implementation.
This changes the ws_strptime() stub to unconditionally use
the internal implementation. Previously it would use the
system implementation of available. This is still possible
but is opt-in, i.e., code should add the necessary #ifdefs
and assume responsability for handling non-portable formats
or providing limited functionality on some platforms.
Text import allows the user to specify the strptime()
format freely, so in that case it makes sense to use the
system's implementation, and pass the responsability
for understanding the implementation and the supported
specifiers to the user.
Only fall back to our implementation if the system libc
lacks a strptime().
If AddressSanitizer is enabled, we build dumpcap with it, which
can be useful for catching issues - at least when run standalone.
Disable it when running in capture child mode, and ensure that the
sanitizers don't change dumpcap's exit code, since capture/capture_sync
uses the exitcode value and doesn't expect the sanitizer values.
Also disable LSan if prctl(PR_GET_DUMPABLE) is 2 on Linux, since
that means that the process will fail to run LSan and die with a
fatal error in the attempt due to not being able to ptrace. It
cannot be disabled through the normal method of using environment
variables since not being dumpable also means that /proc/self/environ
isn't readable. (LSan is able to deal with a dumpable value of 0,
temporarily setting it to 1, but not 2, because it's not allowed to
set it back to 2, due to CVE-2006-2451.)
Fix#18907
In certain situations using __MINGW64__ is not correct.
We want to have the condition apply using MinGW-w64 but also
using MSYS2, which the __MINGW64__ condition alone does not
capture.
Add a HAVE_MSYSTEM C define and use it where appropriate.
Forcing the use of a dedicated header to replace pcap.h is
unnecessary code and mental overhead in this case. We can
use config.h instead for the same purpose of defining a
macro symbol before including pcap.h.
_WIN32 is defined by the compiler, and is arguably a more reliable
test that WIN32. Switch to checking for _WIN32 in a couple of places in
the code.
Remove a WIN32 definition from config.h. It was added for the WinPcap
developer pack but we no longer use that.
Switch to the name "Logray" for the log analyzer. Rays are biological
cousins of sharks and more people like the name "Logray" in a completely
unscientific survey here. Apologies for any inconvenience this might
cause.
Rewrite ws_inet_pton{4,6} and ws_inet_ntop{4,6} without
GLib types.
Check for strerrorname_np() and use that is available,
to simplify error handling.
Add some minimal tests.
Currently used to define ssize_t on platforms that lack it.
Fix some Windows build errors caused by moving the definition into a
separate header.
Fix some narrowing warnings on Windows x64 from changing the definition
of ssize_t from long int to int64_t.
The casts in dumpcap are ugly but necessary. The whole code needs
to be rewritten for portability, or the warnings disabled.
Move epan_memmem() and epan_strcasestr() to wsutil/str_util.
Rename to ws_memmem() and ws_strcasestr(). Add compile time
check for a system implementation and use that if available.
We invoke those functions using a wrapper to avoid exposing
_GNU_SOURCE outside of the implementation.
In order to be able to defer solving all Qt6 API differences at once
I tried to reactivate the QT_MULTIMEDIA_LIB feature. I managed to fix
most problems but one problem remains in both Qt5 and Qt6 builds.
Without Qt[56]Multimedia, the following error exceeds my non-existing
C++ knowledge:
jmayer/work/wireshark/git/ui/qt/rtp_player_dialog.cpp:154:18: error: out-of-line definition of 'RtpPlayerDialog' does not match any declaration in 'RtpPlayerDialog'
RtpPlayerDialog::RtpPlayerDialog(QWidget &parent, CaptureFile &cf, bool capture_running) :
^~~~~~~~~~~~~~~
PCRE2 is the future of PCRE. The only advantage of GRegex is that
it comes bundled with GLib, which is not an advantage at all.
PCRE2 is widely available, the GRegex abstractions layer are not a
good fit and abstract things that don't need abstracting or that we
could handle better ourselves, there are open bugs (#12997) and
maintenance is spotty at best.
GRegex comes with many of the problems of bundled code, aggravated by
the fact that it completely falls outside of our control.
Besides the obvious limitation of being unavailable on Windows,
the standard is vague about getopt() and getopt_long() has many
non-portable pitfalls and buggy implementations, that increase
the maintainance cost a lot. Also the GNU libc code currently
in the tree is not suited for embedding and is unmaintainable.
Own maintainership for getopt_long() and use the musl implementation
everywhere. This way we don't need to worry if optreset is available,
or if the $OPERATING_SYSTEM version behaves in subtly different ways.
The API is under the Wireshark namespace to avoid conflicts with
system headers.
Side-note, the Mingw-w64 9.0 getopt_long() implementation is buggy
with opterr and known to crash. In my experience it's a headache to
use the embedded getopt implementation if the system provides one.
The following commits removed code that required the following defines,
so remove them:
c0711693ab HAVE_GETOPT_H (Partial; still required by CMake)
2925fb0850 HAVE_MKSTEMPS
0c889d6f5c HAVE_SYS_IOCTL_H
0c889d6f5c HAVE_SYS_SOCKIO_H
0c889d6f5c HAVE_STRUCT_SOCKADDR_SA_LEN
9c5049a80b HAVE_STRUCT_STAT_ST_FLAGS
fcntl.h appears to be available on all of our supported platforms,
including Windows. We've also been including it without HAVE_FCNTL_H
guards in a few places (e.g. sshdump.c) without any issues for some
time.
floorl is part of C99.
We use some private functions from MIT kerberos:
- krb5_free_enc_tkt_part()
- decode_krb5_enc_tkt_part()
- encode_krb5_enc_tkt_part()
but we already do that for krb5int_c_mandatory_cksumtype(),
which is newer than the above functions.
We use all of them only under HAVE_KRB5_PAC_VERIFY,
so we don't seem to need additional configure tests.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
ws_log_domains.h needs to be included before wslog.h to be used
to define WS_LOG_DOMAIN. Also the definition for enum ws_log_level
needs to be exported for other APIs so move that to ws_log_domains.h
and rename the file to ws_log_defs.h to reflect the new scope.
Experience has shown that:
1. The current logging methods are not very reliable or practical.
A logging bitmask makes little sense as the user-facing interface (who
would want debug but not crtical messages for example?); it's
computer-friendly and user-unfriendly. More importantly the console
log level preference is initialized too late in the startup process
to be used for the logging subsystem and that fact raises a number
of annoying and hard-to-fix usability issues.
2. Coding around G_MESSAGES_DEBUG to comply with our log level mask
and not clobber the user's settings or not create unexpected log misses
is unworkable and generally follows the principle of most surprise.
The fact that G_MESSAGES_DEBUG="all" can leak to other programs using
GLib is also annoying.
3. The non-structured GLib logging API is very opinionated and lacks
configurability beyond replacing the log handler.
4. Windows GUI has some special code to attach to a console,
but it would be nice to abstract away the rest under a single
interface.
5. Using this logger seems to be noticeably faster.
Deprecate the console log level preference and extend our API to
implement a log handler in wsutil/wslog.h to provide easy-to-use,
flexible and dependable logging during all execution phases.
Log levels have a hierarchy, from most verbose to least verbose
(debug to error). When a given level is set everything above that
is also enabled.
The log level can be set with an environment variable or a command
line option (parsed as soon as possible but still later than the
environment). The default log level is "message".
Dissector logging is not included because it is not clear what log
domain they should use. An explosion to thousands of domains is
not desirable and putting everything in a single domain is probably
too coarse and noisy. For now I think it makes sense to let them do
their own thing using g_log_default_handler() and continue using the
G_MESSAGES_DEBUG mechanism with specific domains for each individual
dissector.
In the future a mechanism may be added to selectively enable these
domains at runtime while trying to avoid the problems introduced
by G_MESSAGES_DEBUG.
The existing stuff doesn't appear to work (I tried it on 32-bit Ubuntu
18.04, and it did *not* add any flags to the compilation, as it appeared
not to conclude that they were necessary, even though they were).
Pull in the stuff from libpcap, which *does* appear to work. (it does
so in my 32-bit Ubuntu testing).
This should fix#17301.
While we're at it, fix cppcheck.sh so that it doesn't attempt to run
cppcheck on files that have been deleted.
If we're not on Windows, use clock_gettime(CLOCK_REALTIME) *if* we have
it; otherwise, fall back on gettimeofday().
(Note: neither Linux, nor macOS, nor Windows necessarily "have"
particular APIs; particular *versions* of Linux distributions
(kernel+libc) have them, particular *versions* of macOS have them, and
particular *versions* of Windows+MSVC have them.
And Linux, Windows and macOS aren't the only platforms on which we run.)
Fixes#17101.
It's possible to play opus payload with libopus (https://opus-codec.org/).
Closes#16882.
Helped-by: Pascal Quantin <pascal.quantin@gmail.com>
Signed-off-by: Lin Sun <lin.sun@zoom.us>
Signed-off-by: Yuanzhi Li <ryanlee@mail.ustc.edu.cn>
João Valverde
Harald Welte