> From: Rocky Bernstein <[email protected]> > Date: Thu, 8 Jan 2026 15:59:18 -0500 > Cc: [email protected], [email protected], [email protected] > > > I was discouraged. I learned that MI, with its binary-encoded interface, > is also far from optimal. > Possibly, > > that's why MI hasn't become more widely adopted. The current thought and > hope (well, actually, from > a > > decade ago), would be to support more modern remote debugging protocols, > such as Microsoft's > debug > > protocol, the one that JVM/Eclipse uses, and the one that V8 uses. (There > are also the DBGp and > Xdebug > > debug protocols) > > > > The great thing about "standards" is that there are so many to choose > from! But one thing is clear: > MI isn't a > > popular choice. > > GDB supports DAP since v14. > > Ok, but I don't understand what that has to do with the topic of this thread: > realgud and gud, neither of which > support DAP yet.
It has to do with your comments about GDB/MI being problematic, and that it would be good for GDB to support more modern remote debugging protocols. DAP is such a protocol. > And to be precise, gud.el doesn't directly support MI either. The additional > support was done by a Google > Summer of Code student and put in an add-on package called gdb-mi.el. That's an outdated view. gdb-mi.el is nowadays the main GDB interface in Emacs, and the command "M-x gdb" invokes it, not gud.el. "M-x gdb" runs GDB with the "-i mi" option by default. This has been so since Emacs 24.
