information, as it'll fill in whatever stuff happens to be in the
capture info. We should somehow figure uot whether this is a capture we
just did (note: cf->is_tempfile does *NOT* necessarily tell us that, as
it's set for the result of a merge, and is cleared if we save the
capture) and fill in the interface information only if that's the case.
Put in a comment to note that we need to do that.
svn path=/trunk/; revision=42874
stuff - arrange that the interface summary list is set, if we've read a
capture file that has interface information, *regardless* of whether we
have libpcap/WinPcap or not.
That means that summary_fill_in() should fill in the interface
information for the summary if there's interface information from the
capture file, and summary_fill_in_capture() - which is called only if
HAVE_LIBPCAP is defined, and can exist only if HAVE_LIBPCAP is defined
(as it takes an argument of a type that's defined only if HAVE_LIBPCAP
is defined) - just fills in interface information from the capture
options and does so only if there's none from the file.
svn path=/trunk/; revision=42873
(otherwise, nobody sees them); do so.
Fix some cases where we weren't doing the proper post-processing after
doing the heuristics for format changes that didn't involve magic-number
changes (discovered because constructing the interface list is now being
done as part of that post-processing).
Fixes bug 7287.
svn path=/trunk/; revision=42872
Wiretap encapsulation values; rename the field in question encap_type to
emphasize that. (Code that looks at that field already assumes it's a
Wiretap encapsulation value.)
For live captures, map the LINKTYPE_ value to a Wiretap encapsulation
value.
wtap_encap_string() never returns NULL, so don't check for a null return
value.
svn path=/trunk/; revision=42871
cppcheck realized that if_info is known not to be null in that code
path, and therefore that checking whether it's null in that code path is
unnecessary. Remove it.
svn path=/trunk/; revision=42867
Fix an ancient copy-and-pasteo of mine ("me" here meaning Guy Harris,
not Evan Huus) - remove an unused data structure (used in the code I
copied and pasted to make this code, not used here).
svn path=/trunk/; revision=42866
(read a record header) from failure (got an EOF or an error). Make it
just return a Boolean.
If it fails in vwr_read(), don't overwrite *err_info (yes,
vwr_read_rec_header() might have set *err_info, so don't lose - and
leak! - the value it returned) - trust vwr_read_rec_header(), or the
routines it calls, to have set it. (If there's a code path where that
doesn't happen, that code path needs to be fixed; the setting of
*err_info in vwr_read() should *not* be restored.)
Thanks to Evan Huus for finding a useless variable with cppcheck, and
reporting it in bug 7295, provoking me to look at this.
svn path=/trunk/; revision=42865
- unpack an xz file with xzcat instead of gzcat
- remove some trailing whitespace
- the update of gnutls required two changes in the gnutls
build stuff: a) gnutls.pc is now generated from gnutls.pc.in
and b) we now need to build with --without-p11-kit
svn path=/trunk/; revision=42861
have the file open. Go back to doing it with a copy on Windows.
Explain what the problem is, and give a way in which we might be able to
make it work on Windows (without using any NT native API calls...).
svn path=/trunk/; revision=42859
saving files, and run it modal (which we're already doing with the
GtkFileChooserDialog); this means less callback-based state machine
stuff, simplifying the code paths a bit.
If we're saving a file before closing it, don't bother reloading it
after saving it.
svn path=/trunk/; revision=42855
file from which we got the error.
On a successful read, always clear out err - wtap_read() doesn't set *err
on success.
svn path=/trunk/; revision=42854
whole summary dialog into an editor-like dialog with an OK and Cancel buttons
(OK sets the new capture file comment, Cancel doesn't).
In order to keep the dialog the same regardless of the file type (and avoid
having a Cancel and OK button when there's no text field to edit), allow
users to create or edit capture-file comments even if the file type is not
PCAPNG (they can add a comment via the add/edit comment UI anyway).
Don't include a Clear button: the user can just Ctrl-A + backspace if they
want to do that.
Don't set the comment text to "[None]" if there's no comment, just leave it
blank.
Don't allow the user to create more than 1 Summary dialog at a time.
svn path=/trunk/; revision=42834
summary_update_comment() which is no longer necessary).
cf_update_capture_comment() has the advantage that it doesn't mark the file
as unsaved unless the comment actually changed.
svn path=/trunk/; revision=42832