Pull the code to register plugin taps, and the loop to register built-in
taps, into a single register_all_tap_listeners() routine.
This leaves it up to libwireshark, not to the programs using it, to know
how to register them.
If, for example, you run out of file descriptors, dup() can fail, and
ws_dup() is a wrapper around it on UN*X. Don't just pass the result of
ws_dup() to ws_fdopen(); instead, save its result, check against -1 and,
if it's -1, give up, otherwise pass it to ws_fdopen().
This addresses Coverity CID 1471708.
Also, if ws_fdopen() fails, close the descriptor we got from ws_dup();
this closes a possible FD leak.
Fix various warnings with the following changes:
Pass a list of include directories to lupdate. Fixes:
ui/qt/proto_tree.cpp:57: Qualifying with unknown namespace/class ::ProtoTree
and similar warnings.
Use QT_TRANSLATE_NOOP instead of QT_TR_NOOP. Fixes:
ui/qt/lte_rlc_statistics_dialog.cpp:993: tr() cannot be called without context
ui/qt/lte_mac_statistics_dialog.cpp:911: tr() cannot be called without context
ui/qt/font_color_preferences_frame.cpp:28: tr() cannot be called without context
ui/qt/font_color_preferences_frame.cpp:29: tr() cannot be called without context
ui/qt/font_color_preferences_frame.cpp:30: Discarding unconsumed meta data
Add Q_OBJECT to the class definition. Fixes:
ui/qt/models/filter_list_model.cpp:120: Class 'FilterListModel' lacks Q_OBJECT macro
The following warnings were not fixed. This might require moving IOGraph
to its own file:
ui/qt/io_graph_dialog.cpp:320: Qualifying with unknown namespace/class ::IOGraphDialog
ui/qt/io_graph_dialog.cpp:555: Qualifying with unknown namespace/class ::IOGraphDialog
ui/qt/io_graph_dialog.cpp:1059: Qualifying with unknown namespace/class ::IOGraphDialog
ui/qt/io_graph_dialog.cpp:1485: Qualifying with unknown namespace/class ::IOGraphDialog
Displaying statistics with tshark results in a memory leak.
tshark -r <any pcap file> -z dhcp,stat -q
==26971==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 24 byte(s) in 1 object(s) allocated from:
#0 0x7f89a4bae518 in calloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9518)
#1 0x7f8989af2918 in g_malloc0 (/usr/lib/x86_64-linux-gnu/libglib-2.0.so.0+0x53918)
init_stat_table allocates a table_stat_t. This is used as private data while
the tap listener is running but it's not freed afterwards.
This patch adds a finish callback for the tap listener where the
table_stat_t is freed.
qt-5.12.2/include/QtCore/qstring.h:291:31: note: candidate 1: QString
QString::arg(double, int, char, int, QChar) const
Q_REQUIRED_RESULT QString arg(double a, int fieldWidth = 0, char
fmt = 'g', int prec = -1,
qt-5.12.2/include/QtCore/qstring.h:975:16:
note: candidate 2: QString QString::arg(int, int, int, QChar) const
inline QString QString::arg(int a, int fieldWidth, int base, QChar
fillChar) const
wireshark/ui/qt/rtp_analysis_dialog.cpp:926:77:
error: ISO C++ says that these are ambiguous, even though the worst
conversion for the first is better than the worst conversion for the
second: [-Werror]
.arg(abs(r_calc.start_time_ms - f_calc.start_time_ms), 0,
'f', 6)
Change-Id: I6a27adff3b03bcfeac8fb56ceb0833d2707000b5
As running pd_ui_->advancedView->expandAll() takes a noticeable amount
of time and so would introduce significant lag while typing a string
into the Search box, we instead debounce the call to
updateSearchLineEdit(), so that it doesn't run until a set amount of
time has elapsed with no updates to the Search field.
If the user types something before the timer elapses, the timer restarts
the countdown.
Fixes#17107
Set the parent of the QMessageBox in the constructor instead of calling
QMessageBox::setParent(). The latter inherits from QDialog, and it
"clear[s] the window flags specifying the window-system properties for
the widget (in particular it will reset the Qt::Dialog flag)."
(See https://doc.qt.io/qt-5/qdialog.html#details )
This makes the dialog properly appear instead of attempting to save a file
with comments to a file type that does not support comments silently failing.
Fixes#17146.
The patch reintroduces WiresharkDialog::captureFileClosed() method and
calls captureFileClosing() and captureFileClosed() in right order.
Both methods call updateWidgets() at its end.
All dialogs were reviewed and captureFileClosing/Closed methods updated
when appropriate.
captureEvent() method in multiple dialogs changed to captureFileClosing/Closed
as it does same actions - looks like old style of detecting of capture
file closing.
Set the Packet List scrollbar page step to be equal to the height
of the scrollbar.
This makes page stepping using the packet list slider more natural.
In OverlayScrollBar return the real ScrollBar sliderPosition to ensure
the correct value is used when handling the actionTriggered signal in
vScrollBarActionTriggered().
This improves turning on and off auto scroll during capture when page
stepping using the packet list slider, because the value is propagated
after this signal.
The patch changes:
- Removed first_packet_mac_addr staff. It was commented out many years ago...
- Removed delta_timestamp item. Not used.
- Sequence verification takes into account timestamp therefore it is able to detect delayed packets more clearly. As consequence of it, #16330 is solved.
- If packet is delayed, it is not used in calculation of diff/jitter/skew. It just mess output. As consequence of it, #16330 is solved.
I checked output with many RTP streams and looks well. But I have no
sample with wrapped timestamp and I have just a few samples with
missing/reordered packets. Nevertheless all are calculated same way as
before and #16330 is solved too.
When waveform start much later than at time 0s, centering of waveform
was incorrect. The reason was that range was taken before rescalingAxes.
Patch solves the issue.
Reason for #16452 is that RTP Player dialog is destroyed before tap
finishes when Escape key pressed. The only solution I found is to check
whether dialog does exists when tap finishes. If so, data are processed
and shown. If dialog was destroyed in meanwhile, no data processing is done.
Start of call column is shown as seconds from start of capture nowadays.
This patch add option to show it as Time of Day. It adds symetry to
other voice processing dialogs
VoIP Calls dialog and RTP Streams dialog has now option to apply display
filter dialog during processing packets.
Filter checkbox is activated during dialog open when display filter is active.
New field apply_display_filter had to be added to voip_calls_tapinfo_t and
_rtpstream_tapinfo/rtpstream_tapinfo_t structures.
Preferences are extended with advanced settings:
gui.decimal_places1 - 2
gui.decimal_places2 - 4
gui.decimal_places3 - 6
rtp_analysis_dialog, rtp_player_dialog and rtp_stream_dialog uses new settings
for formating numbers. Same information in all dialogs uses same settings.
It solves request #15481.
Note: Other UI dialogs can be adapted later.
!1257 solved issue with duplication of information, but removed all
calls from VoIP Calls dialog. This patch solves the issue.
It was tested with many samples and provides same output as 3.4 branch.
We should not prevent a user from using a valid though perhaps enigmatic display
filter just because dfilter_deprecated_tokens() returns true.
Commit 3a53b8643b added ColoringRulesDialog::isValidFilter()
which would return false when dfilter_deprecated_tokens() returned true. Remove
the dfilter_deprecated_tokens() test.
Also fix memory leak for valid display filters.
Closes#17092
The fix solves issue #16952. It reverts commit 88813716 which introduced memory leak which causes the issue. The original issue with duplicating entries is solved too.
Because commit was cherry picked to 3.4.0 (might be in more branches), this patch should be cherry picked too.
The fix solves issue #16952. It reverts commit 88813716 which introduced memory leak which causes the issue. The original issue with duplicating entries is solved too.
Because commit was cherry picked to 3.4.0 (might be in more branches), this patch should be cherry picked too.
When capture contains just RTP streams without signalizations, there
is no way how to play many streams in one player (Analyze button plays
just two streams). This patch adds Play Streams button to dialog and
allows a user to select as many streams they wish.
Adds a pre-commit hook for detecting and replacing
occurrences of `g_malloc()` and `wmem_alloc()` with
`g_new()` and `wmem_new()`, to improve the
readability of Wireshark's code, and
occurrences of
`g_malloc(sizeof(struct myobj) * foo)`
with
`g_new(struct myobj, foo)`
to prevent integer overflows
Also fixes all existing occurrences across
the codebase.
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.
We can't determine the version number, as there's nothing in the header
to indicate the version with which we were compiled, nor is there an API
to determine the version with which we're running.
Make sure NSView.wantsLayer is true by setting QT_MAC_WANTS_LAYER=1 at
startup if we're running on Big Sur and we were built with a version of
Qt susceptible to QTBUG-87014. Fixes#17075?
Every press of Play Stream or Prepare Filter caused incorrect increasing
of Packets count and added Comments.
The reason was that callinfo statistics were not clear before recap
therefore all new values were added to exiting ones.
Patch solves it.
Current CSV/YAML export from RTP stream exports just pure/unformated items. Therefore e.g. SSRC looks different way than shown in dialog. This patch adds additional columns with formated values.
In addition, export uses same method how to get values from a record as dialog uses.
Selecting Edit / Preferences on my asan+ubsan build brings up the
following warning from ubsan:
/media/sf_wireshark.git/ui/qt/models/column_list_model.cpp:273:9:
runtime error: load of value 25, which is not a valid value for type 'bool'
The problem is in ColumnListModel::populate(), where a ListElement is
populated from fmt_data. The ListElement's "changed" component is not
initialized, though.
It looks like "changed" is not used anywhere. This patch removes it from
the ListElement struct.
For Qt 5.11 and newer use horizontalAdvance() instead of boundingRect().width()
to calculate the width of a QString to position the hover rectangle position,
and to select which byte(s) to highlight.
Closes#17033.
Notes:
1. Wireshark internal max limit for ring buffer files is 100,000
2. Wireshark internal limit before warning the user is 65,535
3. GTK: The old GUI did not limit the value for this parameter
4. Qt: This value was limited to 1,000
Change:
Set the GUI limit to match the warning limit (65,535)
Closes#15968#17059
- set width of pull down list so items are visible
- allow user to resize columns in UAT editor
- resize columns on open and changing data
- if a UAT file does not exist, display the UAT name in lower right
where pathname is displayed for files
- pad pathname on right to account for right tilt of italic fonts
Note: a future change may be to not resize columns where user has
changed width.
The Export Objects Content-Type filter combobox should not sort the
header of "All Content-Types" in with the list of content types, but
should ensure that it is always the first item, especially as the first
item position is used to show all content. This is particularly an issue
in some localizations; e.g., すべてのコンテントタイプ alphabetizes
after actual content types. Fixes bug #17048
QFont::ForceIntegerMetrics is deprecated, and the floating-point
variant of QFontMetrics is no longer used anyway since 96eec0beb9.
Change the font_width_ member to int, as should have been done in
that commit.
CaptureOptionsDialog allocates a SparkLineDelegate in its constructor.
It should set itself as parent of the SparkLineDelegate.
Without a parent, the SparkLineDelegate is never freed and each
invocation of Capture / Options leaks memory.
Don't call the columnString() and discard the result in order to force
colorization; instead, add a separate method to force colorization and
use that.
This avoids the need to choose a column; we were using 1 as the column
number, but column numbers are zero-origin, so that's column 2, which
isn't guaranteed to exist (a crash ensued if it didn't).
Improves the resolution of interval steps that can be selected in IO Graph.
Selectable interval steps follow a scheme of 1 -> 2 -> 5 -> 10.
Having a broad choice of different intervals is important for visualizing.
Pull the value-formatting code in proto_custom_set into
proto_item_fill_display_label. Use that in FieldInformation::toString
instead of fvalue_to_string_repr. Fixes#16911.
Add ByteViewText::updateLayoutMetrics, which fetches the character width
and line height. Call it whenever our font changes and when we're about
to paint. Blind attempt at fixing #15819.
Systemd journal entries aren't file-type-specific; they're found in both
systemd journal entry blocks in pcapng files and in systemd journal
export files. Give it a record type, for use with both file types.
This fixes#16955.
It also means that you can open a systemd journal export file and save
it as a pcapng file.
Start the limit at 2^32-1, as we use a guint32 to store the frame
number.
With Qt prior to Qt 6, lower the limit to 53 million packets; this
should fix issue #16908.
This reverts commit 5df2925434.
The problem only showed up in tfshark.c, and was caused by tfshark.c
using stuff from ui/urls.h but not *including* ui/urls.h.
If you use it, GCC 9.3.0 seems to think there's a missing parenthesis
somewhere, just as the version of clang++ in my version of Xcode does,
even though other versions of GCC don't. I'm clearly missing something
obscure about C here; I give up.