On Jul 21, 2000, Greg McGary <[EMAIL PROTECTED]> wrote:

> If I read you correctly, we can slip this in more easily if we keep
> the script growth to a minimum.

It would help, but I still don't think autoconf is the place to
support this feature.

> That makes some sense, but potentially opens a large can of worms.  In
> general, people won't want a BP library unless they're debugging, and
> in that case they'll want BPs for the code that uses the library.

I know of a Brazilian GNU/Linux distribution (www.Conectiva.com.br)
that ships their Server edition with some form of bounds checking
enabled by default.  I'm pretty sure their users would love to be able
to build applications with the libraries they ship without being
forced to use some particular compiler switch.  Of course, this would
reduce the amount of checking that can be done, but it wouldn't make
it totally useless.

> However, this situation might not work at all if B's interface
> includes data structures that contain pointers.

I see.  Converting the data structures would probably be too much work
to do at every function call, if doable at all.

If that's the case, maybe you should add the __BP_ prefix to *all*
functions and data symbols, not only to those that use pointers.
Then, users would *have* to use the compiler switch, and they'd get
the correct results for the tests, because the compiler would then
introduce the prefix regardless of the signature of the function.

> If I can come up with a modified AC_CHECK_FUNCS that doesn't bloat
> the configure script much

Bloating configure is just a minor issue.  What I'm really looking for
is some solution that doesn't require changes in autoconf.  If, after
some discussion, we agree that changing autoconf is the way to go,
I'll be all for it, but you haven't convinced me yet :-)

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Reply via email to