http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1430
------- Comment #46 from jaap.keuter@xxxxxxxxx 2007-10-16 08:24 GMT -------
(In reply to comment #44)
> As /WX is already defined and this still not detects the unused variable I
> tried /analyze instead. This shows that the declaration of the static const
> true_false_string flags_set_truth in some of the files hides the declaration in
> epan\tfs.h. When removing my own declarations of flags_set_truth, it just wont
> compile any longer as flags_set_truth is not defined.
> When looking in epan\tfs.h there is a comment for the flags_set_truth that
> reads:
> * Old true_false_string from packet.c
> * Retained for backward compatibility until all dissectors are updated.
>
> What is this? What update should be required in *all* dissectors?
flags_set_truth should be considered an alias to tfs_set_unset. That is what
the comment is referring to.
The main problem is that a plugin in a DLL (note this is a Windows only
problem) can't use these predefined true_false_strings, since the data resides
in another DLL. It's just a limitation of the Windows dynamic linker, which
other platforms don't have.
So, the way to fix this is to include the declarations of the required TFS in
your plugin code wrapped in #ifdef _WIN32 ... #endif. This way it will be
compiled in locally in your Win32 plugin, while the other platforms just use
the provided strings from tfs.c. Sure, it's not so nice, but hey go complain to
Bill.
BTW: You'll have to do somthing similar if you want to use the xdlc helper
functions in a plugin.
--
Configure bugmail: http://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.