I've implemented a very simple callback mechanism which provides exactly this. I've tried GHook from GLib before, but this doesn't seem to be the right thing, as it's too inflexible for the purpose here.
So I've implemented a callback function in main.c which receives all "events" and spreads them to menu, statusbar and itself.
I would see this implementation as a prototype which may need improvements. Please comment the changes.
svn path=/trunk/; revision=13330
"main_set_for_capture_file()"; it should only deal with menus, not
anything else - and it gets called while the menus are being set up,
which is before the main window has been completely created, so
"main_widgets_show_or_hide()", which is called by
"main_set_for_capture_file()", gets errors trying to show or hide
widgets the pointers to which are null.
svn path=/trunk/; revision=13328
move another two capture related fields (iface and cfilter) from cfile to capture_opts
also move the handling of capture related command line options from main.c to capture.c, that way a future privilege seperated capture program can use the same code to parse it's command line than Ethereal.
It might be even possible to share this parser code even with Tethereal, didn't took a closer look at this.
svn path=/trunk/; revision=13320
cf_ functions, have separate ones for different classes of routines, and
use gboolean when the return value is just "success" or "failure" - that
way you don't get compiler warnings if a case statement isn't handling a
particular status value if the routine in question won't return it.
svn path=/trunk/; revision=13293
This includes: all functions in file.h now have a cf_ prefix, will have doxygen tags, will have the capture_file *cf as the first parameter and I tried to generalize the return values for non trivial functions.
Hopefully, I didn't introduced any new bugs, as I had to change a lot of files...
svn path=/trunk/; revision=13289
don't use global cfile at all but only an untpyed handle to call the cf_... functions in file.c
move the save_file member from capture_file to capture_opts, as it's only used while capturing and while preparing it
svn path=/trunk/; revision=13276
one.
"get_basename()" doesn't modify its argument, and its callers don't
modify the substring pointed to by the result, so make it take a "const
char *" as an argument and return a "const char *".
"find_last_pathname_separator()" doesn't modify its argument, so make it
a "const char *" - but some of its callers pass a non-"const" "char *"
and modify the result, so don't make its return value a "const char *".
And, as none of its callers are outside "filesystem.c", make it static.
In "about_folders_page_new()", have separate variables for pathnames
returned as "const char *" (which are cached by the routine that returns
them, so you can't modify them - and can't free them, so get rid of the
commented-out "g_free()" calls for them) and pathnames returned as "char
*" (which are allocated anew for each call, and can be modified, but
have to be freed).
Clean up white space.
svn path=/trunk/; revision=12881
that return the next packet from a set of {chronologically sorted,
sequential-by-file} packets; it doesn't need to have a loop over all
those packets, or any code to write packets.
Supply those abstractions, change the code that merges packets to do its
own writing, and have the Ethereal version manage a progress bar and
have the mergecap version print packet numbers in verbose mode, as the
common merge code used to do.
svn path=/trunk/; revision=12427
indication - success, read failure, write failure - and have their
callers handle read failures by looking for the file that got the read
failure and reporting the failure in question.
Free up the err_info string returned by "wtap_read()" after using it.
svn path=/trunk/; revision=12423
mergecap.c (get rid of the verbose printing of information for each
packet).
Have "merge_append_files()" return FALSE only on a write error, as
"merge_files()" does.
Sort the routines in "merge.c" in the order from "merge.h".
svn path=/trunk/; revision=12422
and use that information to provide better error messages.
Have "merge_open_outfile()" do all the work of filling in the
merge_out_file_t structure, with the values to use passed as arguments.
Get rid of some structure members that used to be used solely to pass
information to "merge_open_outfile()".
Add a "cf_merge_files()" routine to do the merging and reporting of errors.
svn path=/trunk/; revision=12420
really more of an Ethereal/Tethereal component than a libethereal
component (nothing else in libethereal knows about capture files); move
it back out of libethereal. (The range stuff doesn't; we leave it in
libethereal.)
svn path=/trunk/; revision=11898
they should ultimately be split into files with routines that handle
ranges, which are just subsets of [0,2^32), and packet ranges, which are
subsets of the packet list, possibly specified by a range.
Move them into epan, so they can be used by, for example, utilities that
handle ranges, such editcap.
svn path=/trunk/; revision=11890
already been marked and have unmark_frame() do likewise. Don't mess
with the marked frame count in mark_all_frames().
Be a little more paranoid about the marked frame count in other places.
svn path=/trunk/; revision=11775
"print_ok_cb()", and have "print_packets()" just work on a
"print_stream_t" handed to it, so that different platforms can open the
printer/print file in different ways (opening the file is probably not
going to be platform-dependent, but opening the printer will be).
svn path=/trunk/; revision=11544
a structure containing a pointer to print operations for that object and
a pointer to the private subclass-dependent data for that object, with
subclasses for text and PostScript, and use those rather than the old
scheme where a print format was passed as an argument - or where (as in
the case of printing summary information in Tethereal) we just printed
as text even if "-T ps" was selected.
Check whether those routines succeed or get an I/O error writing output.
Clean up indentation.
svn path=/trunk/; revision=11514
Also move ncp222.py, x11-fields, process-x11-fields.pl,
make-reg-dotc, and make-reg-dotc.py.
Adjust #include lines in files that include packet-*.h
files.
svn path=/trunk/; revision=11410
they have LF at the end of the line on UN*X and CR/LF on Windows;
hopefully this means that if a CR/LF version is checked in on Windows,
the CRs will be stripped so that they show up only when checked out on
Windows, not on UN*X.
svn path=/trunk/; revision=11400
the "text vs. PostScript" choice. The "text vs. PostScript" choice
should probably ultimately be done with a generic set of print methods,
to handle various platform-native print mechanisms more cleanly (and
perhaps the dialog box code for "export as {PDML,PSML}" should be
separate from the "export as text"/"print" dialog).
svn path=/trunk/; revision=11342
is per-packet and the packet has an encapsulation type we don't know
about, so handle it on reads as well as errors - show an error message
noting that we had a packet with a network type we don't know about, and
show the extra info returned for that error giving details.
It shouldn't return WTAP_ERR_UNSUPPORTED, however, so just give the
"wtap_strerror()" error for that case.
svn path=/trunk/; revision=11340
add a config.nmake option to control whether to build
libethereal.dll or not;
remove "./wiretap" from PATH to prevent problems due to
wrongly-loaded files;
build dissector.lib with MSVC;
move "print.c" and "ps.c" to the dissector helpers, as "print.c"
imports variables from packet-frame.c and packet-data.c, which
are in libethereal;
move "g711.c" out of the dissector helpers, as they're used only
by Ethereal in a tap, not in Tethereal or in any dissector;
add a .def file for libethereal;
arrange to declare global variables exported from libethereal
with "__declspec(dllimport)" when building programs that import
those variables;
update the NSIS installer.
Make the "configure" script define ETH_VAR_IMPORT as "extern".
svn path=/trunk/; revision=10834
when using GTK2 code for rendering the error. In order to correctly
render the error message, it must be XML escaped.
TODO: track down the remaining places where this XML escaping is
required, and fix it there too (not sure if they exist though).
svn path=/trunk/; revision=10764
-ps: added formatting hints for ghostscript, so pdf conversion will be much better
-ps: print a thin line at the top and bottom of each page
-ps/text: add an option to start a new page for every packet (formfeed)
svn path=/trunk/; revision=10660
print the header before each summary line, and print a blank line
separating the summary line and the remaining information.
svn path=/trunk/; revision=10614
with "string" rather than "ascii", to make it clearer what they're
involved with.
Use "gtk_toggle_button_set_active()", not
"gtk_toggle_button_set_state()" (the latter is a deprecated alias for
the former, probably dating back to GTK+ 1.0[.x] - 1.2[.x] and later
have "gtk_toggle_button_set_active()").
Do *NOT* change the radio buttons for the type of string search to do
based on whether we're doing a string search or not - doing so means we
don't correctly remember the type of string search.
Get rid of code to fetch some values that we don't subsequently use.
svn path=/trunk/; revision=10609
put the definition of it back under HAVE_LIBPCAP and don't add an extra
declaration in "file.h", as there's no longer code that needs to refer
to it if HAVE_LIBPCAP isn't defined.
svn path=/trunk/; revision=10350
I don't know whether this is the optimal patch, but it does the job.
file.h: extern declaration of auto_scroll_live
file.c: always declare auto_scroll_live
svn path=/trunk/; revision=10347
"main_filter_packets()", to force the filtering to be done even if the
filter is the same as the current one; this is necessary in order to
make sure "Follow TCP Stream" gets the packets processed even if you're
filtering the stream that's currently filtered in.
svn path=/trunk/; revision=10209
read errors; there are separate implementations for Ethereal (pops up an
alert box) and Tethereal (prints an error message).
Use those routines in the ASN.1 dissector.
svn path=/trunk/; revision=10152
"cf_XXX_failure_alert_box()" routines that put the alert box up, and
directly call the "alert_box.h" routines for OS errors.
svn path=/trunk/; revision=10028
attempt to write to a file (or close a file opened for writing).
Get rid of no-longer-needed #includes of <epan/filesystem.h>.
svn path=/trunk/; revision=10027
filter expressions; use that in a number of places, so we use the same
alert box. (More work is needed to figure out the right way to handle
some other "dfilter_compile()" failures.)
Use the error message from the display filter as the primary error, as
that's the message that tells you what the underlying problem is. (The
GNOME HIG says "In most situations the user should only need the primary
text to make a quick decision", so the primary text should tell you
what's wrong with the filter, not just that it's invalid. If there are
messages from the display filter code that don't give enough
information, or are a bit cryptic, such as "Unexpected end of filter
string," those should be fixed in the display filter code.)
Improve the error used if an empty filter is used for "find frame".
svn path=/trunk/; revision=10025
"simple_dialog()"; NULL might be #defined to be a pointer expression on
some platforms, causing compiler warnings (and, on platforms where a
null pointer doesn't have all its bits 0, possibly causing misbehavior,
although I don't think there are any such platforms on which Ethereal
runs).
Don't allow 0 as button mask argument to "simple_dialog()".
Squelch a compiler warning.
Report fatal problems as errors, not warnings.
Report file I/O errors with "file_open_error_message()".
Report file write errors (including those reported by "close()", e.g.
some errors writing to an NFS server) when saving raw packet data to a
file.
svn path=/trunk/; revision=9915
addition to an error code, an error info string, for
WTAP_ERR_UNSUPPORTED, WTAP_ERR_UNSUPPORTED_ENCAP, and
WTAP_ERR_BAD_RECORD errors. Replace the error messages logged with
"g_message()" for those errors with g_strdup()ed or g_strdup_printf()ed
strings returned as the error info string, and change the callers of
those routines to, for those errors, put the info string into the
printed message or alert box for the error.
Add messages for cases where those errors were returned without printing
an additional message.
Nobody uses the error code from "cf_read()" - "cf_read()" puts up the
alert box itself for failures; get rid of the error code, so it just
returns a success/failure indication.
Rename "file_read_error_message()" to "cf_read_error_message()", as it
handles read errors from Wiretap, and have it take an error info string
as an argument. (That handles a lot of the work of putting the info
string into the error message.)
Make some variables in "ascend-grammar.y" static.
Check the return value of "erf_read_header()" in "erf_seek_read()".
Get rid of an unused #define in "i4btrace.c".
svn path=/trunk/; revision=9852
"capture_file" structure. Keep it locally, instead.
Check for errors when printing packets.
Report failure to open a print destination and failure to write to a
print destination differently.
Don't have the "print preamble" and "print final" routines return
success/failure indications - revert to the old scheme where they
didn't, and have the callers use "ferror()" to check for errors.
Report write errors when printing dissections in Tethereal.
Report print errors as errors, not warnings.
svn path=/trunk/; revision=9828
"file_close_error_message()" - but just use "file_write_error_message()"
for UNIX-style errors, under the assumption that a close will only fail
because a buffer-flushing write fails or because "close()" itself fails
when, for example, pushing unsynced NFS client-side writes out over the
wire.
Make several routines in "print.c" return success/failure indications.
Check for write errors when printing "Follow TCP Stream" stuff or saving
it to a file.
svn path=/trunk/; revision=9825
"cf_write_error_message()"/"file_write_error_message()".
Use "file_open_error_message()" instead of "cf_open_error_message()" in
some places we missed in the previous checkin.
Catch ENOSPC and EDQUOT in "file_open_error_message()".
Use "file_open_error_message()" rather than "file_write_error_message()"
to report errors when creating the file to which we're saving the
"Follow TCP Stream" data.
svn path=/trunk/; revision=9823
translate UNIX errno values to a somewhat friendly message format
string.
Rename "file_open_error_message()" in "file.c" to
"cf_open_error_message()", make "cf_open_error_message()" use the new
"file_open_error_message()" for UNIX errno values, have "do_capture()"
in "capture.c" use "file_open_error_message()" to report errors from
"open()", and make "cf_open_error_message()" static as nothing outside
"file.c" uses it.
Do similar stuff in "tethereal.c".
svn path=/trunk/; revision=9821
As this will always be a Cancel of a running operation, this parameter was removed.
This makes us also able to use a stock button for this.
svn path=/trunk/; revision=9774
b.) added new feature "Edit->Go To First Packet" "Edit->Go To Last Packet" with corresponding menu and toolbar items
c.) added new feature "View->Zoom In" / "View->Zoom Out" / View->Normal Size" with corresponding menu and toolbar items
This feature will act as a "size offset" to the current fontsize, so that the packet list/tree view/... will have a larger/smaller font size.
The value is stored inside the recent file.
d.) Win32 only: Try to get the win32 system font and fontsize at program startup and show the menus/dialogs and such with the same font and fontsize like other win32 windows.
This makes the program make a *lot* more feel like a normal win32 program.
svn path=/trunk/; revision=9753
them and running all taps on them, but not reconstructing the packet
list. Use that in the IO-stat tap rather than "redissect_packet()"; the
latter does more work and redraws the display, neither of which are
necessary.
Call the filter callback when the Calc field is changed, to redraw the
graphs; that change also fixes things so that it's called when the
Filter field is changed.
Rename the "filter_button" member of an io_stat_graph_t to
"filter_field", as it's not the "Filter:" button, it's the text field
containing the filter expression.
svn path=/trunk/; revision=9659
packets in a loop; it handles reading the packet from the capture file,
creating and updating the progress bar, handling the stop button,
and handling range specifications.
Use that for printing and saving packets. (There might be other loops
that should use it as well.)
svn path=/trunk/; revision=9620
filter's colors" bug, by storing the pointer to the matching color_filter_t
structure in the frame_data structure.
Replace "frame" and "frames" by "packet" and "packets" in many places.
svn path=/trunk/; revision=9607
use the common display filter dialog infrastructure in the MGCP
service response time tap;
add common infrastructure for updating the titles of tap dialogs
when the capture file name changes.
svn path=/trunk/; revision=9366
"opened capture files" and "display filter used",
the settings will be saved in the file "recent" in the users config path
svn path=/trunk/; revision=9275
captures, as it has to compute the width of an auto-resizing column in
every row. Just pick fixed widths for the columns (and tune the width
of the "Protocol" column so that it's not narrower than the column
title).
svn path=/trunk/; revision=9219
pointers to the first *and* last child, in the "proto_node" structure
itself. That saves us one level of indirection and memory allocation,
and lets us append to a tree by appending to the last child directly,
rather than having to scan through the list of siblings of the first
child to find the end of that list.
svn path=/trunk/; revision=9171
last columns, if any, with that format, and use that to speed up
processing of columns with a particular format and checking whether
we're displaying a column with a particular format.
svn path=/trunk/; revision=9147
This makes the CulmulativeBytes field make more sense since if we want
something to be a TimeReference frame it is likely that we also want to
measure BOTH time and number of bytes (==culmulative bytes) until the event we are looking at.
svn path=/trunk/; revision=8956
return "<no file>" if there is no file loaded yet instead of crashing in io-stat
io-stat only print the label for the top tick on the y axis to make it look less cluttered
svn path=/trunk/; revision=8781
"set_menus_for_selected_packet()" and
"set_menus_for_selected_tree_row()", and have them decide whether to
enable or disable menu items based on whether that structure indicates
that a packet or field is selected and, if one is, on its properties.
Pass to the "selected packet enabled" routine for a menu item the
"frame_data" and "edt" members of the "capture_file" structure, and pass
to the "selected tree row enabled" routine the "field_info" member of
that structure.
Clear "cf->current_frame" if no packet is selected.
svn path=/trunk/; revision=8525
don't have to worry about the file name being null when
"cf_get_display_name()" is called. (It should never be called before we
have a capture file open; if it is, that's a bug.)
svn path=/trunk/; revision=8482
capture temporary files, it's "<capture", and for saved capture files,
it's the last component of the pathname of the file. Use that in
various places when displaying the file name.
svn path=/trunk/; revision=8474
One can now select a packet and mark it as a TimeReference packet using the menu.
A TimeReference packet will be indicated by having all timestamp related column entries replaced by the string *REF*
A TimeReference packet will always be displayed in the packet pane, and overrides any display filters.
When a frame is a TimeReference frame, all later frames will calculate the TimeRelativeToFirstPacket relative to the timestamp of the TimeReference frame instead of the first frame of the capture.
You can have any number of TimeReference frames you like.
svn path=/trunk/; revision=8459
the option to print only marked packets similarly, rather than as
"Suppress unmarked packets" (for consistency, and because the latter
isn't unlike a double negative).
svn path=/trunk/; revision=8451
Only calculate culmulative bytes on those packets that are actually displayed in the packet pane.
When a display filter is applied, culmulative bytes should only be calculated on thoise packets that are actually displayed.
svn path=/trunk/; revision=8418
Ethereal presents a column to display culmulative bytes into the capture.
A new column type is added : Culmulative Bytes.
While PacketLength column type specifies the number of bytes in the current packet,
Culmulative Bytes specifies the culmulative number of bytes from the start of the capture.
svn path=/trunk/; revision=8359
so we shouldn't torment the users by offering it.
Check the string type and convert it to an internal representation in
the GUI code; have the search code deal only with the internal
representation.
Save the case-sensitivity flag, and the indication of where string
searches look, along with other search parameters.
Upper-casify the string, for case-insensitive searches, in the GUI code;
don't save the upper-casified string, so it doesn't SHOUT at you when
you next pop up a "find" dialog.
Convert the hex value string to raw binary data in the GUI code, rather
than doing so in the search code. Check that it's a valid string.
Connect the signals to the radio buttons after the pointers have been
attached to various GUI items - the signal handlers expect some of those
pointers to be attached, and aren't happy if they're not.
Have "find_packet()" contain a framework for searching, but not contain
the matching code; instead, pass it a pointer to a matching routine and
an opaque pointer to be passed to the matching routine. Have all the
routines that do different types of searching have their own matching
routines, and use the common "find_packet()" code, rather than
duplicating that code.
Search for the Info column by column type, not by name (the user can
change the name).
When matching on the protocol tree, don't format the entire protocol
tree into a big buffer - just have a routine that matches the text
representation of a protocol tree item against a string, and, if it
finds a match, sets a "we found a match flag" and returns; have that
routine not bother doing any more work if that flag is set.
(Unfortunately, you can't abort "g_node_children_foreach()" in the
middle of a traversal.)
Free the generated display filter code after a find-by-display-filter
finishes.
svn path=/trunk/; revision=8306
isn't sufficient if the line to be printed is more than twice the
current size of the line buffer. Make it twice the size of the line to
be printed, instead.
svn path=/trunk/; revision=7921
specified by the user, so if the frame with that frame number isn't
being displayed, just refer to it as "That frame", not "The frame with
that frame number". (That's even OK if they *did* specify it.)
(If there's no such frame, that "shouldn't happen" in cases where it was
implicitly specified, e.g. by getting it from a list of fragments, or a
"{request,reply} is in frame XXX" field, or clicking on a point in a TCP
graph.)
svn path=/trunk/; revision=7258
frame and the previous frame in the capture - a frame that might not be
displayed, so you don't know what it was - rather than the previous
frame in the display, as is intended. Fix that.
svn path=/trunk/; revision=6708
in continue_tail_cap_file, the following test :
else if (*err != 0)
sometimes reads uninitialized memory (so the test might be incorrect).
We now set *err to 0 at the beginning of the function.
svn path=/trunk/; revision=6418
- moved color_filter_t in color.h
- change color_filter_t to use color_t instead of GdkColor
This changed allowed to remove the last gtk includes in file.c. It is
now completely free of any gtk related code.
svn path=/trunk/; revision=6324
- created a few packet_list_xxx functions (ui_util.h gtk/packet_list.c
gtk2/packet_list.c) ;
- removed almost all "gtk/xxx" and "gtk2/xxx" includes in file.c
The only remaining includes are related to color filters. We have to
make color_filter_t GUI independent by replacing GdkColor with color_t.
I'll work on this later.
svn path=/trunk/; revision=6311
All the deprecated widgets have not been replaced yet :
GtkList and GtkCList ==> GtkTreeView conversion :
- color_dlg.c
- column_prefs.c
- decode_as_dlg.c : done
- dfilter_expr_dialog
- filter_prefs.c
- main.c
- plugins_dlg.c : done
GtkCTree ==> GtkTreeView conversion : done
GtkText ==> GtkTextView conversion : done
Remaining problems :
- gtk_font_selection_dialog_set_filter doesn't exist anymore (but hasn't
been removed from the documentation). I don't know how to filter the
font selection dialog to get only fixed width fonts ;
- we have to remove GUI prefs which are not usefule anymore : tree line
style and tree expander style.
svn path=/trunk/; revision=6153
information.
Fix the types of some variables (make the file position in
"read_cap_file()" a "long", as Wiretap supports "long" offsets, and make
processed-packet counts in packet-processing loops "int"s, as the total
packet count in a "capture_file" structure is an "int").
svn path=/trunk/; revision=6112
equivalents for the toplevel directory. The removal of winsock2.h will
hopefully not cause any problems under MSVC++, as those files using
struct timeval still include wtap.h, which still includes winsock2.h.
svn path=/trunk/; revision=5932
Ethereal sometimes creates a progress dialog bar and then, if
the processing is fast, quickly destroys it. The resulting
"flash" can be disconcerting. This set of patches ensures a
progress bar is either not created or is displayed for a minimum
time.
svn path=/trunk/; revision=5916
Allow "-" as the output file name in Wiretap, referring to the
standard error.
Optimize the capture loop.
Fix some of the error-message printing code in Ethereal and Tethereal.
Have Wiretap check whether it can seek on a file descriptor, and pass
the results of that test to the file-type-specific "open for output"
routine. Have the "open for output" routines for files where we need to
seek when writing the file return an error if seeks don't work.
svn path=/trunk/; revision=5884
padding when printing text and use "putline" when printing PostScript.
Eliminate "hexdump", as it's just a special case of "putline".
Have "proto_tree_print_node()" just call "print_line()".
Get rid of "print_ps_hex()" - the font used for the hex dump and the
protocol tree are the same, so there's no need to switch fonts.
svn path=/trunk/; revision=5798
routines; instead, rename "print_hex_data_common()" to
"print_hex_data_buffer()" as it prints a buffer in hex and ASCII (or
EBCDIC), and have it take the print format as an argument.
Have it use "print_line()" to print lines, and fix "print_line()" so
that it prints a buffer *without* a newline at the end, printing the
newline itself.
svn path=/trunk/; revision=5729
Have "wtap_open_offline()", if asked to open a FIFO, return that error
if it was asked to open the file for random access.
svn path=/trunk/; revision=5643
in the "packet_info" structure instead, as we don't need a pointer for
every single frame in the capture file, just for each frame for which we
currently have an open "epan_dissect_t".
svn path=/trunk/; revision=5614