Reward for implementing Export keyword. [ was: Re: C++ export keyword ]

Gabriel Dos Reis gdr@integrable-solutions.net
Mon Apr 14 02:04:00 GMT 2003


Martin Sebor <sebor@roguewave.com> writes:

[...]

| I missed the meeting, so I'm surprised to hear that there was such a
| wide spread lack of support for the paper given that no commercially
| available compiler (other than Comeau; even the EDG 3.0 based Intel
| C++ disables it) still implements the feature, and no vendor, AFAIK,
| has immediate or even near term plans to implement it. I was not in

Yeah, that argument was made many times.

For what it worths, I would like to make the point that I didn't hear
any vendor saying that they are *not* going to implement export.  Even
one of the proponent of the paper said (several times in public
forums) his company made a public commitment to implement the feature.

| favor of removing the feature when I was asked to review the paper,
| but given my experience with enabling export in our library I could
| certainly empathize with many of the points the paper makes about
| the difficulty and absence of benefits of using export.

That would have been interesting for the debate.  Dinkumware, the
other library vendor who sells an export-enabled version of C++
library, had an opposite point of view and supported "export".

When export was added to the language, the debate wasn't whether it
would be useful; the debate was whether it would be implementable.
Now, we have a proof that it is implementable. And, remember that
EDG's current implementation is the first release, which means that
now they have ideas to improve it.

-- Gaby



More information about the Gcc-bugs mailing list