Draw in black, erase by just redrawing the current pixmap.
Will apply to tcp_graph.c if/when this gets through the buildbots.
svn path=/trunk/; revision=44308
Updating 'qtshark_de.ts'...
Found 72 source text(s) (16 new and 56 already existing)
Kept 13 obsolete entries
Updating 'qtshark_fr.ts'...
Found 72 source text(s) (16 new and 56 already existing)
Kept 10 obsolete entries
Similar-text heuristic provided 3 translation(s)
Also remove $Id$, it is xml file (auto)generated by Qt (lupdate)
svn path=/trunk/; revision=44305
( It is link Tshark Help not Wireshark (already in bottom !))
Also remove $Id$, it is Xml file (auto)generated by Qt Design
svn path=/trunk/; revision=44303
/home/jmayer/work/wireshark/svn/trunk/ui/gtk/capture_dlg.c:209:19: error: ‘compile_bpf_w’ defined but not used [-Werror=unused-variable]
svn path=/trunk/; revision=44299
Fix cmake (and possibly autotools) build for:
lib/libqtui.a(interface_tree.cpp.o): In function `SparkLineDelegate':
/home/jmayer/work/wireshark/svn/trunk/ui/qt/sparkline_delegate.h:10: undefined reference to `vtable for SparkLineDelegate'
This would not show up using qtcreator due to the magic of qmake.
I must admit that I don't fully understand what I am doing here, but
I read the following thread to the bottom and one of the last posts
contained the solution:
http://stackoverflow.com/questions/1095298/gcc-c-linker-errors-undefined-reference-to-vtable-for-xxx-undefined-refere
svn path=/trunk/; revision=44280
Note: It still fails with
Linking CXX executable qtshark
lib/libqtui.a(interface_tree.cpp.o): In function `SparkLineDelegate':
/home/jmayer/work/wireshark/svn/trunk/ui/qt/sparkline_delegate.h:10: undefined reference to `vtable for SparkLineDelegate'
collect2: error: ld returned 1 exit status
svn path=/trunk/; revision=44257
Expand show version preference.
implementation of Steven's suggestion of a radio button (actually
dropdown) of the 4 possibilities.
I changed the default to have the version on both in
the welcome screen and on the window as that's how it was and people tend not to change the preferences I think.
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=6437
svn path=/trunk/; revision=44153
cairo do it. The clipping that was done worked properly for horizontal/verical
lines only, but would not for diagonal ones.
svn path=/trunk/; revision=44125
tcp_graph.c needlessly does this itself, but it only has to deal with horizonal
and vertical lines....
When have returned to original, unzoomed state, set zoom to be 'in' for
the middle mouse button, as thats all that makes sense.
svn path=/trunk/; revision=44124
isn't an SN=1023 for the channel.
Now I understand the time origin toggle - make it show the capture time
by default now.
svn path=/trunk/; revision=44099
- only setting the colour when it has changed
- only calling cairo_stroke() when we need to rather than after every
element
These improvements were previously made to rlc_lte_graph.c. I did look
at trying to extract what these 2 graph modules had in common, but gave up pretty quickly
when I saw how much they had diverged...
svn path=/trunk/; revision=44087
Valgrind log:
= 47,616 bytes in 1 blocks are possibly lost in loss record 41,578 of 41,608
= at 0x4C26ABB: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
= by 0xAA2BA53: gdk_pixbuf_new (in /usr/lib64/libgdk_pixbuf-2.0.so.0.2200.1)
= by 0x10379F2F: ??? (in /usr/lib64/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-xpm.so)
= by 0x1037A4FD: ??? (in /usr/lib64/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-xpm.so)
= by 0xAA3132E: gdk_pixbuf_new_from_xpm_data (in /usr/lib64/libgdk_pixbuf-2.0.so.0.2200.1)
= by 0x4589B5: xpm_to_widget (gui_utils.c:572)
= by 0x4E3FC4: about_wireshark (about_dlg.c:86)
svn path=/trunk/; revision=44066
long list of elements. Now the graph is like greased lightening!
I will look later to see if I can apply the same change to the TCP
graphs.
svn path=/trunk/; revision=44001
do the same type thing for the ui_desciptions.
eg put a "place holder" in the UI descriptions and
make a call from main_menubar.c i.e add_filter_menu()
and do the UI stuff there.
svn path=/trunk/; revision=43999
If column is based on frame data packet_list_get_widest_column_string()
was constructing his own frame_data with biggest *value* for given column
(which might be not the longest string) and later fill column for it.
Which generally should work (time precision is fixed),
but in r41185 to lot col_*() function was added check if fd->flags.has_ts is set.
packet_list_get_widest_column_string() was passing frame_data with flags.has_ts unset.
To make maintaince easier, instead of fixing the problem with flags.has_ts,
generate column buffer for each frame, and find the one which has longest string.
svn path=/trunk/; revision=43991
Updating 'qtshark_de.ts'...
Found 69 source text(s) (23 new and 46 already existing)
Updating 'qtshark_fr.ts'...
Found 69 source text(s) (25 new and 44 already existing) Kept 1 obsolete entries
There is new string to translate....
svn path=/trunk/; revision=43989
graph until know that we were able to choose a TCP session.
This could fail if e.g. a MAC LTE frame had > 1 bearers carrying TCP
segments.
svn path=/trunk/; revision=43977
user_data is NULL in the callbacks from the menu_bar, and when preparing filters we need to find the filter widget with g_object_get_data(G_OBJECT(data), E_DFILTER_TE_KEY).
The object is saved in the global variable popup_menu_object.
Use that in the filter callbacks from the menu bar.
- remove unused widget from a couple of callbacks.
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=7506
svn path=/trunk/; revision=43961
Put generated, with *proper* types version of marshaler (VOID:OBJECT,OBJECT) inside bytes_view.c
It was also only blocker of #undef GTK_DISABLE_DEPRECATED, so remove it.
XXX generated version of marshaler looks portable, but it'd be best to generate it using
glib-genmarshal during compilation.
svn path=/trunk/; revision=43947
this should fix crashes in the interface list handling when the
interface list contains pipes initliazed by scan_local_interfaces() or
add_pipe_cb()
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=7511
svn path=/trunk/; revision=43922
gtk_combo_box_text_get_active_text() and gtk_combo_box_get_active_text() [from old-gtk-compat.h]
returns *newly allocated string (...) must be freed with g_free().*
svn path=/trunk/; revision=43918
shown in two places and there's no reason to show it in a third.
Use CommDlg_OpenSave_GetFilePath instead of CommDlg_OpenSave_GetSpec so
that previews work properly for folders.
svn path=/trunk/; revision=43877
Cheat a little and always make BytesView focused to make render blue-highlight instead of grey /inactive/ one.
XXX, we should have our own BytesView class and allow user define his own style in css.
svn path=/trunk/; revision=43859
so that we can properly associate a widget with create, update, and
destroy events. Only used by Qt so far but it should be easy enough to
add to GTK+.
Rename ui/qt/progress_dialog.{h,cpp} to progress_bar.{h,cpp}. Show a
progress bar in the status bar of the main window instead of creating
a separate dialog. Note that we still need to add a "cancel" mechanism
and display the task and item titles somewhere.
Thus began the War Against Gratuitous Dialogs.
svn path=/trunk/; revision=43833
Bissecting shows, that the problem was introduced by commit
r43009 | etxrab | 2012-06-03 01:53:04 +0200 (So, 03 Jun 2012) | 1 line
From Evan Huus Convert a bunch more GTK-2 calls to their cross-version ws_ equivalent
To get the problems back, compile with -DBREAK_GTK3
svn path=/trunk/; revision=43787
filter make sure we check its syntax. Return a boolean. --This line,
and those below, will be ignored--
M ui/gtk/capture_file_dlg.c
svn path=/trunk/; revision=43757
and win32_open_file behave similarly. _snwprintf is "banned". Use
StringCchPrintf instead. Tested on Windows. I'll test on OS X shortly.
svn path=/trunk/; revision=43756
redrawing the graph.
Don't mess around with a static pointer when allocating segments. This
seems unecessary and can result in the last one being leaked.
These changes were already made to rlc_lte_graph.c, which is based upon
the tcp graph code.
svn path=/trunk/; revision=43740
3; if so, that might indicate that GTK+ 3 doesn't have gtkmarshal.h.
Try not attempting to include it with GTK+ 3.x - if that causes other
failures, well, I guess that means that, when building the BytesView
widget under GTK+ 3, we can't use the gtk_marshal_ functions....
svn path=/trunk/; revision=43737
gtk_widget_get_allocation() (gtk_widget_set_window() doesn't have them),
but *do* put parentheses around its "x" argument.
Fix gtk_widget_set_allocation() to set the allocation.
svn path=/trunk/; revision=43735
packet_panes.c(304) : error C2220: warning treated as error - no 'object' file generated
packet_panes.c(304) : warning C4244: 'function' : conversion from 'gdouble' to 'int', possible loss of data
packet_panes.c(304) : warning C4244: 'function' : conversion from 'gdouble' to 'int', possible loss of data
Add casts, old code did the same casts so this should be OK.
svn path=/trunk/; revision=43730
object for each line - rather create one and use it for the entire list,
then destroy it.
Simplify the allocation of segments, by no longer messing around with a
static that may or may not be used. Also, there was a good chance the
last one would be leaked (if it was not part of the stream we are
interested in).
Note that both of these improvements would also apply to tcp_graph.c -
will wait a couple of days to be sure the changes are good.
svn path=/trunk/; revision=43654
When showing the TCP (and UDP and other protocols) port numbers in the
framelist columns, they are left aligned. This is sub-optimal if you are
scanning the list for a certain port.
This patch, very small, will properly right justify these numbers:
UNRESOLVED DST PORT
UNRESOLVED SRC PORT
DST PORT
SRC PORT
svn path=/trunk/; revision=43642
Fix an uninitialized-variable error that was breaking compile with gcc.
Also remove some cruft that looks like it was copy-pasted from
ui/gtk/tcp_graph.c
Fixes bug 7460.
svn path=/trunk/; revision=43635
disable hide checkboxes with error messages
Updated preference dialog to include text if hide checkbox is disabled.
Used error message for "Local Interfaces" dialog.
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=2598
svn path=/trunk/; revision=43622
IO Graph Fixes
- Allow counting *any* type of field, FT_NONE, FT_PROTOCOL, whatever, by moving
FT_NONE case to default case per Jakub's suggestion.
- Fix get_it_value() for FT_PROTOCOL fields by breaking CALC_TYPE_COUNT_FRAMES
and CALC_TYPE_COUNT_FIELDS out of the adv_type switch statement just like
COUNT_TYPE_FRAMES.
- Avoid an overflow in pixmap_clicked_event() that was causing crashes when
clicking outside the data portion of the graph (when the capture hasn't enough
files to fill the entire graph window),
svn path=/trunk/; revision=43621
Fix leaks
- don't g_strdup a string just to use it in a g_strdup_printf
- clean up properly in error cases in lua bindings
- misc. other missing g_free() calls
- one missing fclose() in the new 80211_utils
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=7454
svn path=/trunk/; revision=43617
Add a new name resolution option: whether or not use the configured (in the OS)
name resolver (e.g., DNS) to resolve network names. When this option is disabled
but network name resolution is enabled then Wireshark will resolve only those
names that it can from local sources. This includes (at least, AFAIK):
- name resolutions that Wireshark picks up on from DNS packets it decodes
- the "user hosts file" (~/.wireshark/hosts on *NIX)
- what Wireshark reads out of capture file (the PCAPNG name resolution block)
This new preference defaults to "use external resolvers" for backward
compatibility (so people turning on network name resolution will get the old
behavior).
This option can be set via Edit->Preferences and on the command line; there
remain several UIs (e.g., the "open capture file" dialog, the
View->Name Resolution menu, etc.) that don't have the new option yet.
Also expand on the "description" for the name resolution preferences: these
are used not only in the tooltips but are also written to the preferences
file. The previous text didn't include enough context when written do the
preferences file.
svn path=/trunk/; revision=43605
implicitly by the #define name and string they were defined to; not all
UATs neatly fit into any of the categories, so some of them were put
into categories that weren't obviously correct for them, and one - the
display filter macro UAT - wasn't put into any category at all (which
caused crashes when editing them, as the GUI code that handled UAT
changes from a dialog assumed the category field was non-null).
The category was, in practice, used only to decide, in the
aforementioned GUI code, whether the packet summary pane needed to be
updated or not. It also offered no option of "don't update the packet
summary pane *and* don't redissect anything", which is what would be
appropriate for the display filter macro UAT.
Replace the category with a set of fields indicating what the UAT
affects; we currently offer "dissection", which applies to most UATs
(any UAT in libwireshark presumably affects dissection at a minimum) and
"the set of named fields that exist". Changing any UAT that affects
dissection requires a redissection; changing any UAT that affects the
set of named fields that exist requires a redissection *and* rebuilding
the packet summary pane.
Perhaps we also need "filtering", so that if you change a display filter
macro, we re-filter, in case the display is currently filtered with a
display filter that uses a macro that changed.
svn path=/trunk/; revision=43603
There is no guarantee that uat->category is non-null, so check whether
it's null before checking whether it's UAT_CAT_FIELDS or not.
Fixes bug 7444.
#BACKPORT
svn path=/trunk/; revision=43602
"Find Next Mark" was duplicated in the Edit menu and "Find Previous Mark" was
missing (the 2nd "Find Next Mark" would, if selected, find the previous mark).
svn path=/trunk/; revision=43600
some warnings from GTK+.
For interior nodes in the preference module tree, we create a page even
if the module itself has no preferences, so there's *something* we can
show if the user clicks on it. (Showing the top-level protocols page is
a bit weird, and requires us to keep track of it.)
svn path=/trunk/; revision=43596
Yes, the following files can now be deleted:
\wireshark\ui\gtk\prefs_nameres.c
\wireshark\ui\gtk\prefs_nameres.h
\wireshark\ui\gtk\prefs_print.c
\wireshark\ui\gtk\prefs_print.h
\wireshark\ui\gtk\prefs_protocols.c
\wireshark\ui\gtk\prefs_protocols.h
\wireshark\ui\gtk\prefs_taps.c
\wireshark\ui\gtk\prefs_taps.h
I thought removing them was part of the patch for bug 7402, but
they still show up in the SVN (perhaps I didn't create the patch
correctly).
Make it so.
svn path=/trunk/; revision=43591
Remove refresh button from wireless toolbar
With a "refresh interfaces" menu option and device hotplug
notification in place there is no longer need for a refresh
button in the toolbar.
svn path=/trunk/; revision=43570
Refresh wireless toolbar too
The recently improved refresh interfaces code forgot the wireless
toolbar. Make sure to refresh it too.
svn path=/trunk/; revision=43568
proto_tree_draw() from r43189 is lazy, and it don't fills tree with all item labels.
To fix it, move 'gbl_resolv_flags' hack to proto_tree_model_get_value().
XXX could we pass resolving flags to proto_item_fill_label()?
svn path=/trunk/; revision=43550
platform-dependent work is now done in get_interface_type().
That means we also don't need to protect the handling of IF_DIALUP or
even IF_AIRPCAP in #ifdefs, just in case we add support for those types
in other OSes.
Nothing's uisng the capture_ethernet_16.xpm image here, so don't bother
including it. (Does *anybody* use it? I couldn't find any uses, but
maybe we should use that.)
svn path=/trunk/; revision=43531