Re: M-x gdb troubles

[Stefan Monnier]

>> If I have a running gdb process (M-x gdb ... run), and accidentally do
>> M-x gdb again (instead of switching to the gdb buffer), things get
>> really messy...

>> Maybe the second M-x gdb should just switch to the current gdb buffer 
>> if I enter the same command line as a previous gdb run.

> How about this patch. The catch form would have to be added to the other
> debugger functions (dbx, pdb, perldb etc) to be consistent.

Yuck.  I'd much rather just throw an error and be done with it.
Let the user switch to the gdb buffer manually.


        Stefan


PS: while I'm here: it occurred to me that the gdb-ui stuff should be split
    into a "frame/window management" part (the gdb-many-windows stuff) and
    an "annotation=3" part.  The frame/window-management would be extended
    to include the "old GUD" style, would be made non-GDB specific, and
    moved to gud.el.  This way the new frame system (with a special
    gud-source-window and such) would also be available for other debuggers.

    Of course, this is only seen from a conceptual point of view.
    From a coding point of view, it might require a complete
    rewrite/redesign: I wouldn't know.  Tho I do understand that the
    frame/window-management is already partly separate so maybe it wouldn't
    be that hard to do.