put this stuff into an external file but just doing a forward
declaration to squelch a compiler warning isn't the way to go - it
needs to be done right one day.
svn path=/trunk/; revision=19827
Here is a patch to fix a problem with notification parameters. Fix is:
Remove premature return statements from fNotificationParameters.
The real problem is that the original code is inconsistent about putting
a break in all cases of the switch blocks and some were falling through
to the default case where there was a return. I took the simple route
and removed the unnecessary returns - they were OK for default case but
not for the fall through.
Is there a standard about having breaks? If it is desired to have a
consistent coding style I will further update the code to conform.
Otherwise, could someone please commit this change as it does fix the bug?
Thanks,
Dave Richards
svn path=/trunk/; revision=19797
The new version of the dissector will compile without any patches to apply to
Wireshark/NDR.pm. Jelmer has improved the patch and pidl now support enums as
return type. Regarding the nspi.cnf file, it won't be useful anymore.
Regarding the dissector, I've noticed a bug in NspiGetMatches which isn't
dumped correctly. I'm currently investigating the problem and I'll notify the
list when a fixed version is available.
svn path=/trunk/; revision=19794
*) Remove maximum LDAP PDU size check - they can get large with either large attributes (e.g. CRLs, SPIFs) or with lots of results (see http://www.wireshark.org/lists/wireshark-users/200610/msg00197.html). The max size preference is also removed.
*) Support for dissecting LDAP controls including server side sorting and paged results. A new BER function is introduced to see if there is a dissector for a given OID.
*) Remove reference to removed BER preference in the LDAP reassembly preference.
*) Mark a LDAPURL as a URL
svn path=/trunk/; revision=19792
Attached is a fix for bug #1156: "If a dissector depends on a new
datasource created by a new tvb, that dissector's protocol can't be
filtered on."
Thanks,
Steve
I change the code slightly to avoid dereferenceing a NULL pointer.
While the original code was fine and safe on unix dereferencing a null pointer in this way would be unsafe and cause a segmentation violation on windows if tree is NULL.
svn path=/trunk/; revision=19791
here is a patch that transforms "valid lifetime" and "prefered lifetime" in a prefix option (in RA)
in human readable format. Following the RFC2461 Section 4.6.2, if the value is 0xffffffff, this means lifetime
is infinity.
svn path=/trunk/; revision=19779
pseudo-field by looking at the ID for hf_text_only. Apparently
some fields really don't have 'name' fields, but we still want
their info to be dumped out.
svn path=/trunk/; revision=19763
there are many reasons why some protocols actually need to be able to access the pinfo structure while determining the pdu size
svn path=/trunk/; revision=19751