32-bit numbers. Separate signed and unsigned accessors have been
added and used where appropriate.
Definitely not for 0.99.5.
svn path=/trunk/; revision=20472
* <epan/crypt/crypt-md5.h> must come after <glib.h> because of
'guint8' etc.
* Include <wiretap/file_util.h> because of eth_fopen().
svn path=/trunk/; revision=20456
a patch to avoid the warning "implicit declaration of function
'strptime'" in editcap.c
glib.h is included just after the define __USE_XOPEN and include <time.h>
svn path=/trunk/; revision=20455
remove ENABLE_WSUG setting from config.nmake and replace it with GLIB_MAJOR_VERSION >= 2 (as Glib 1.x cannot open a browser to show a HTML file - it's just currently? not implemented)
svn path=/trunk/; revision=20446
- Win32 only: try to load help page from local user-guide.chm (if HHC_DIR is set in config.nmake) and if that fails:
- show corresponding help page directly from www.wireshark.org in the configured web browser.
The "internet way" will be the only way for all UNIX based versions, as they cannot read .chm files.
svn path=/trunk/; revision=20445
The U3 Deployment Guide recommends a sub-directory of U3_DEVICE_DOCUMENT_PATH, so the u3util creates a "My Captures" sub-directory to use as the personal data file directory.
svn path=/trunk/; revision=20441
- initially show the "My Documents" folder
- don't use the "open folder" preference setting on save and alike dialogs (we don't do it in the GTK versions either)
svn path=/trunk/; revision=20433
add a function get_persdatafile_dir() that will return the users personal default data dir Win32:"My Documents" UNIX:"" (for the current dir)
svn path=/trunk/; revision=20431
The problem was that when dissecting the set, if a sub-dissector didn't consume any bytes it was assumed that the correct field hadn't been matched.
This fix matches the field if the sub-dissector consumes no bytes and we know that the length of the field is zero. This is only allowed on the first pass when we are not matching ANYs.
I think this is a fairly safe fix - I've tried it with some other ASN.1
I've also changed dissect_ber_octet_string() to show the zero length fields in the dissection. This shows the fields as "<MISSING>" which is not quite the right explanation as the field is definitely present. Something like "<EMPTY>" or "<ZERO LENGTH>" may be better - but I'm not sure of the reasoning behind "<MISSING>".
svn path=/trunk/; revision=20429
As a nice side effect, the dir/file part of the dialogs is bigger now, which makes handling of the dialogs much nicer/faster. Yes, it still fits into 800*600 screen resolution.
Needs testing, I might have placed a g_free() call at the wrong place ...
svn path=/trunk/; revision=20423