On Fri, Jan 25, 2008 at 10:23:03AM -0800, David Daney wrote: > Joe Buck wrote: > >On Fri, Jan 25, 2008 at 10:18:01AM -0700, Tom Tromey wrote: > >>>>>>>"Joe" == Joe Buck <[EMAIL PROTECTED]> writes: > >>Joe> On the other hand, old-timers are used to the commands being the way > >>Joe> they are. So I guess that a command that takes a proper argument > >>Joe> should have a different name. > >> > >>You can have it both ways, somewhat: > >> > >>define pt > >> if $argc == 0 > >> set debug_tree ($) > >> else > >> set debug_tree ($arg0) > >> end > >>end > > > >Good idea! If anyone wants to patch the commands, that's how it should > >be done, I think. > > > >>Personally I'd just change to purely using $arg0 and have people > >>adjust their habits. But then, I've used my own argument-taking pt > >>for years :) > > > >But since people can have it both ways, why not? > > Because it won't work at all with an old gdb?
When was $arg0 added to gdb? And why would a gcc developer need to use an old gdb?
