This is the mail archive of the mailing list for the libstdc++ project.

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

Re: slow V3 configures (was Re: ICE during bootstrap.)

On Tue, Dec 05, 2000 at 11:03:24AM -0800, Joe Buck wrote:
> > we check for c++ declarations and c linkage. both are necessary. If you 
> > can figure out a faster way to do it, I'm all ears.
> Suppose you run one test that passes only if all the "common" functions
> are present and declared correctly, and if this test fails, then you run
> individual tests?  What remains is to decide what is "common".  One
> possible goal is to choose a set that will pass in one shot on most
> modern platforms -- but we still have a fallback on the other platforms.

I thought about this once, but I think the truly common set on "most
modern platforms" is too small to be of any use.  What led us to the
current situation were things like finding teeny tiny differences between
OS point versions in the linkage, and they kept defeating our assumptions.
So, no more assumptions.  :-)

If kind-hearted souls with fast out-of-box platforms wanted to post the
math entries of config.cache's for various platforms, then those users with
much slower machines could use them to prime the cache.  Freakishly unsafe,
but fast!  Sort of one-step back from a prebuilt binary package.


pedwards at disaster dot jaj dot com  |  pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools.  Fools are protected by more capable fools.

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