[freetds] replacement iconv

Craig A. Berry craigberry at mac.com
Mon Dec 15 22:43:11 EST 2003


At 11:32 AM -0500 12/15/03, Lowden, James K wrote:
>
>Well, Craig, you *do* sound a little grumpy, but probably with cause.  I was
>relying on my experience, which certainly doesn't encompass all linkers.
>Even though no one's complained to date about linker problems with our
>libiconv replacement, Frediano's choice (and your defense of it) will
>probably save us headaches in the 0.62 release. 

Sorry for the tone.  Unrelated stress plus having previously lost a
similar argument in relation to the other replacement functions have
made me a little impatient.

>P.S.  Did your research say anything about the priority of object files vs.
>libraries?  If you link in iconv.o, rather than iconv.a, will the linker pay
>any more attention?  AFAIK, there are no standards that govern any of this;
>it's just a matter of practice. 

I think your odds improve, but I don't remember explicit commitments.
I know of one case (VMS) where you're guaranteed to get a multiply
defined symbol warning regardless of whether the duplicates are in
object files, object libraries, or shareable images (dynamic
libraries).
-- 
________________________________________
Craig A. Berry
mailto:craigberry at mac.com

"... getting out of a sonnet is much more
 difficult than getting in."
                 Brad Leithauser


More information about the FreeTDS mailing list