Re: PATCH to keep Jason from pointing out what a fool I am

[Mark Mitchell <mark at markmitchell dot com>]

>>>>> "Jeffrey" == Jeffrey A Law <law@hurl.cygnus.com> writes:

    Jeffrey>   In message
    Jeffrey> <199808252036.NAA26392@smtp.earthlink.net>you write:
    >>  This patch enables "-W -Wall" even for stage1 builds, if the
    >> native compiler is GCC.  It's handy if you're building a
    >> cross-compiler, or, if like me, you tend to just type `make'
    >> when making a small change to the C++ front-end.  Jeff?

    Jeffrey> It would be better if we only had the warning flags set
    Jeffrey> in one place in the Makefile.in or configure.in.  Having
    Jeffrey> them appear in multiple places in different files via
    Jeffrey> different mechanisms is a pain.

I don't quite understand this criticism.  I think that what you're
saying is:

  For stage 2/3 builds we enable warnings by setting WARN_CFLAGS in 
  Makefile.in.  For stage 1 builds, you (Mark) are proposing enabling
  warnings in configure.in.  That seems inconsistent.

Assuming I've paraphrased correctly, I'll try to answer the
criticism.  We *unconditionally* set warnings for stage2/3 builds
because we *know* the compiler used with be GCC, and so it will
understand -W -Wall.  Thus, there's no reason to use a configure
mechanism; we just set WARN_CFLAGS unconditionally in the Makefile.
For stage1, though, the compiler might not understand -W -Wall, so
this has to be done via a configuration check, and can't be set
unconditionally in Makefile.in.

    Jeffrey> jeff

-- 
Mark Mitchell 			mark@markmitchell.com
Mark Mitchell Consulting	http://www.markmitchell.com