This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: flag to dump core on error

On Wed, Feb 12, 2003 at 11:15:35AM +0100, Gabriel Dos Reis wrote:
> Geoffrey Keating <> writes:
> | Sometimes I hit a bug that depends on the exact memory layout,
> | including environment variables provided by make or the rest of the
> | build system, and my usual procedure is to write a program 'fake-gcc'
> | that prints out the environment and then another one that starts GCC
> | with exactly that environment.  That's a lot of work.  I'd much rather
> | have a flag that I could restart the build with and get a
> | coredump---the same compile wouldn't fail, but chances are something
> | will.
> | 
> | Thus, -dH (because a-h and A-G were taken).
> :-)  Too cute for me to understand.  I would have prefered averbose flags.

Yeah, I think -d is getting overloaded here.  There's a difference between
"dump" as in "dump rtl pass number 17" (most -d flags) and "dump" as in

It's a really good idea, though.  Thanks for adding this.


I would therefore like to posit that computing's central challenge, viz. "How
not to make a mess of it," has /not/ been met.
                                                 - Edsger Dijkstra, 1930-2002

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]