ANNOUNCEMENT: Live Wireshark University & Allegro Packets online APAC Wireshark Training Session
April 17th, 2024 | 14:30-16:00 SGT (UTC+8) | Online

Wireshark-dev: Re: [Wireshark-dev] [Wireshark-commits] rev 39422: /trunk/gtk/ /trunk/gtk/: main

From: Tony Trinh <tony19@xxxxxxxxx>
Date: Fri, 14 Oct 2011 18:23:19 -0400
On Fri, Oct 14, 2011 at 3:50 PM, Guy Harris <guy@xxxxxxxxxxxx> wrote:

On Oct 14, 2011, at 8:37 AM, Tony Trinh wrote:

> Actually, the valid #define for Lua code throughout Wireshark is HAVE_LUA_5_1 (not HAVE_LUA). HAVE_LUA works in Windows but not other OS's (such as OSX).

...which probably means "such as UN*Xes"; that sounds like a difference between using autoconf and config.nmake.

If we don't have a HAVE_LUA definition independent of the Lua version, that sounds like a bug in the configure script.  Does CMake define HAVE_LUA, HAVE_LUA_5_1, both, or neither?

CMake defines HAVE_LUA_5_1 but not HAVE_LUA. On the other hand, config.nmake defines both, and it's actually the only file to define HAVE_LUA. For reference, there are 3 files that define HAVE_LUA_5_1 and 7 source files that test for it (trivial fix).


> It's really named "make_menu_actions()". I named the function based on what it does, not based on who calls it. That function (and make_menu_xml()) can easily be used outside the context of Lua menus, and there's nothing about them that fundamentally binds them to Lua. That said, it doesn't matter enough to me if their names include the word "lua" since they might change at a later point when someone sees wider use for them.

"Someone" meaning "the Python interface", for example?

Yes. Good example.