Mail Archives: djgpp/1997/10/29/09:13:21
On Mon, 27 Oct 1997, Paul Moore wrote:
> Is GCC (and hence DJGPP) likely to come back up to speed on the
> standard C++ implementation anytime soon?
What ``standard'' are we talking about here? AFAIK, no ANSI C++
standards were issued lately, they just keep introducing new features
and significant changes into the Draft for what seems like forever.
> It's quite a long way behind
> the "current state of the art" if I recall correctly
IMHO, the ``current state of art'' is a hodgepodge of incompatible
features and extensions produced because compiler vendors are trying
to shoot a moving target--the evolving standard--and the target moves
too fast. I have yet to see a C++ program using the ``state of the
art'' that can be compiled without infinite #ifdef'ery with two
different compilers, even if they both support these hot features.
> Of course, I could be wrong - but 2.7.2.1 has been around for quite a
> time now, and the standard has changed a lot recently...
If you need to live on the edge, you could always download the latest
snapshots of the next GCC release in the works (they used to call it
``the compiler of the day'', and not for nothing).
Personally, I refuse to work with a language (or features thereof) for
which a reasonably-accepted standard doesn't exist. I find it a waste
of my efforts to debug what in the end boils down to immature
technology badly implemented by over-zealous vendors which try to
throw in all of the hottest features mentioned in the last-night
conference of the X3J16 comittee, so it looks better in the ads. I
prefer to excercise will-power and do my job using more stable tools.
Note that I have no relation whatsoever to the GCC development team,
so please don't take the above as expressing anything but my own
private views.
- Raw text -