This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [RFC] std::string(0);
Nathan Myers wrote:
On Fri, Jun 11, 2004 at 07:51:21PM +0200, Daniel Frey wrote:
Martin Sebor wrote:
A conforming implementation is not allowed to reject constructs
such as std::string (0) just because they might cause undefined
behavior at runtime.
OK, so this means we can't go the "easy" way. The remaining options are
(AFAICS):
a) Change the standard, make the above example illegal. Unlikely.
b) Ask the compiler folks to detected these cases and emit a warning.
Reporting this usage at compile time would be a valuable service.
The most frequent complaint I've heard is from users who expect
it to construct a "null" (i.e., empty) string. AFAIK, that's even
the behavior on platforms that let you read from address 0. I would
much prefer defining meaningful behavior for it than turning it
into an anal retentive error and breaking well-formed programs
or implementations. Constructing the empty string is IMO perfectly
reasonable, harmless, and has no adverse performance impact over
the current behavior (the one if that the requirement is presumably
trying to eliminate is there anyway in implementations that throw).
Martin