This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH for expand_end_bindings
At 20:15 22.05.99 , mark@codesourcery.com wrote:
> >>>>> "Franz" == Franz Sirl <Franz.Sirl-kernel@lauterbach.com> writes:
>
> Franz> May I suggest to handle --enable-checking now the same way
> Franz> as WARN_CFLAGS? Enabled for development, disabled for
> Franz> releases?
>
>Hmm. I think the answer to this depends on just how much
>--enable-checking slows things down. Why don't you try running the
>entire top-level `make check' both ways, and post the result? I'm
>afraid this may be a significant blow-up, which could really hose
>people with many multilibs to test, etc. However, if it doesn't take
>much longer, I'm all for turning it on. Definitely a useful
>sanity-check.
Ok, on a 300 MHz PPC604e running linux-2.2.8/glibc-2.1.1 the bootstrap
takes about 18% longer, the testsuite is running 1-2% slower, so for me a
complete bootstrap/check run is about 12% slower.
For development I find this acceptable, but maybe it's not acceptable for
others. I'll switch it on for my testcompiler by default. Just one thing,
is it possible to modify the testsuite (or dejagnu?) in a way that doesn't
hide the remaining problems in XFAIL or PASS?
Franz.