Re: gdb emacs reporting 'Function "x_error_quitter" not defined'

[Nick Roberts]

 > I think that is not a good idea.  Compiling with -O2 and with -O0
 > produces two very different programs, so doing that by default with
 > CVS sources means that we lose a lot of testing potential, as users of
 > the CVS version will be running a program that is utterly different
 > from what users of the released Emacs will see.  I expect quite a few
 > bugs to remain unearthed until the pretest stage.

Well I'm not sure that they're utterly different but it does seem to be
standard practice to compile with the same level of optimisation for debugging
as release. It make debugging a little more awkward but, as you say, it
ensures that the program you are testing is the same one that the users will
be using. If developers wish to use -O0, then thats their perogative, but I
don't think it should be the default. One could change from -O0 to -O2
shortly before a pretest, like ..err, now?.

Nick