AIX I/O failures

Mark Mitchell mark@codesourcery.com
Thu Jan 25 14:25:00 GMT 2001


>>>>> "Gabriel" == Gabriel Dos Reis <gdr@codesourcery.com> writes:

    Gabriel> That doesn't match the discussion found in the de-facto
    Gabriel> pre-standard reference: ARM.  I've just checked :-)

That's an authoritative reference.  

I can certainly tell you what people *did* with placement new,
independent of what Bjarne and his co-authors wished they would
do. :-) We built error-checking tools that did run-time type checking
of their programs.  We had to learn a lot about what the industry
actually did with C++, and what people expected.  There were a lot of
people who cared about the *when* along with the *where*, and a lot of
code that looked suspciously like the patch I posted...

Anyhow, it's all moot: this is a working solution.  

Until another one is found, it's what we've got.

Thus far, I think there is probably another solution -- your reference
trick.  It's clever, and if you say it's conforming, I'm not inclined
to doubt you!  But, I do think it penalizes every use of I/O, and I
don't think there's any portability win, so I don't find it
compelling.

But, hey, not my call -- see earlier post:

  If my "opaque" hacks cause people to be sufficiently appalled that
  they do something cleaner, and thereby satisfy the goal, I will be not
  all embarssed -- even if not one line of the code I wrote remains! :-)

My job is to get a release of the compiler and the library out the
door.  For now, I'll check in the patch, and let y'all sort it out
afterwards. :-)

In the meantime, people can make progress testing and fixing other
problems on AIX, which is the important thing.

Thanks,

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com


More information about the Libstdc++ mailing list