Clarification of the scope of this group

[Tim Clacy]

I was advised in a recent thread that, because one of the compilers that I
use does not support 'namespace', I should not use this group but ask
elsewhere; perhaps comp.lang.c++.but_without_namespace'? What's the general
consensus?

Tim

[Unforgiven]

Tim Clacy wrote:

I was advised in a recent thread that, because one of the compilers
that I use does not support 'namespace', I should not use this group
but ask elsewhere; perhaps comp.lang.c++.but_without_namespace'?
What's the general consensus?

We talk about everything that is mentioned in the C++ standard, nothing
more, nothing less. If your compiler does not support namespaces advice
about how to circumvent that falls out of the domain of Standard C++ and
into the domain of compiler-specific issues, which are not topical here.
Instead you should ask in a group where your compiler is topical.

--
Unforgiven

"Most people make generalisations"
Freek de Jonge

[Tim Clacy]

Unforgiven wrote:

Tim Clacy wrote:

I was advised in a recent thread that, because one of the compilers
that I use does not support 'namespace', I should not use this group
but ask elsewhere; perhaps comp.lang.c++.but_without_namespace'?
What's the general consensus?
We talk about everything that is mentioned in the C++ standard,

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is purely for
abstract, theoretical language issues and not practical issues with actual
language implementations?
nothing more, nothing less. If your compiler does not support
namespaces advice about how to circumvent that falls out of the
domain of Standard C++ and into the domain of compiler-specific
issues, which are not topical here. Instead you should ask in a group
where your compiler is topical.

[Tim Clacy]

Unforgivable wrote:

If your compiler does not support namespaces advice about how
to circumvent that falls out of the domain of Standard C++ and
into the domain of compiler-specific issues

I've lost count of the number of times I've read advice to avoid
'namespace', and yet here you are suggesting that to ask a question here
related to a workaround for a compiler that doesn't support namespace is
inappropriate; effectively, you are saying 'don't use namespace', but if you
need to use 'namespace' and you're compiler doesn't support it, then ask
elsewhere :-)

[Unforgiven]

Tim Clacy wrote:

Unforgivable wrote:

If your compiler does not support namespaces advice about how
to circumvent that falls out of the domain of Standard C++ and
into the domain of compiler-specific issues

I've lost count of the number of times I've read advice to avoid
'namespace', and yet here you are suggesting that to ask a question
here related to a workaround for a compiler that doesn't support
namespace is inappropriate; effectively, you are saying 'don't use
namespace', but if you need to use 'namespace' and you're compiler
doesn't support it, then ask elsewhere :-)

I have never said 'don't use namespace'. I'd also say that if you *need* to
use namespace and your compiler doesn't support it, you need a different
compiler. ^_^

--
Unforgiven

"Most people make generalisations"
Freek de Jonge

[Unforgiven]

Tim Clacy wrote:

Unforgiven wrote:

Tim Clacy wrote:

I was advised in a recent thread that, because one of the compilers
that I use does not support 'namespace', I should not use this group
but ask elsewhere; perhaps comp.lang.c++.but_without_namespace'?
What's the general consensus?

We talk about everything that is mentioned in the C++ standard,

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is
purely for abstract, theoretical language issues and not practical
issues with actual language implementations?

VC7.1 comes pretty close in terms of standards-compliance (although it has
quite a few non-standard extensions), so does Comeau. But you are basically
right with your statements. We do not discuss any specific implementation,
but C++ 'as it should be'. So if you have issues that are specific to a
certain compiler, and not to C++ as a language, this isn't the place for
them. The only exception of course is when you have a question of the sort
'This doesn't compile in my compiler, should it according to the standard or
am I doing something wrong?'

--
Unforgiven

"Most people make generalisations"
Freek de Jonge

[osmium]

Tim Clacy writes:

We talk about everything that is mentioned in the C++ standard,
THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is purely

for abstract, theoretical language issues and not practical issues with actual
language implementations?

As I see it, discussion is limited to the latest best effort of the major
compilers, rather than to the standard. The major out standing issue is
templated code and how object files magically appear. But there may be some
wiggle room specifically included in the standard on that issue. Whether
there is official wiggle room or not, there is a Gentlemen's Agreement that
some things are OK and some are not. After immersion in the culture for a
while, you will kind of absorb, by osmosis or something, what the rules
really are.

I looked at your other post you refer to. The battle you are fighting is
that all old compilers were supposed to have been put in a bonfire and
burned when the standard was approved. You appear to have one that escaped
destruction. Think smallpox.

[Tim Clacy]

Unforgiven wrote:

Tim Clacy wrote:

Unforgivable wrote:

If your compiler does not support namespaces advice about how
to circumvent that falls out of the domain of Standard C++ and
into the domain of compiler-specific issues

I've lost count of the number of times I've read advice to avoid
'namespace', and yet here you are suggesting that to ask a question
here related to a workaround for a compiler that doesn't support
namespace is inappropriate; effectively, you are saying 'don't use
namespace', but if you need to use 'namespace' and you're compiler
doesn't support it, then ask elsewhere :-)

I have never said 'don't use namespace'. I'd also say that if you
*need* to use namespace and your compiler doesn't support it, you
need a different compiler. ^_^

I don't know where you're from, but you should come visit Earth some time;
it's an awful place where, sometimes, you have no choice but to work with
tools that are flawed. I very much like the sound of your world though; how
do you get there? Do you do any commercial development there or is it purely
academic?

[Tim Clacy]

Unforgiven wrote:

Tim Clacy wrote:

Unforgiven wrote:

Tim Clacy wrote:
I was advised in a recent thread that, because one of the compilers
that I use does not support 'namespace', I should not use this
group but ask elsewhere; perhaps
comp.lang.c++.but_without_namespace'?
What's the general consensus?

We talk about everything that is mentioned in the C++ standard,

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is
purely for abstract, theoretical language issues and not practical
issues with actual language implementations?

VC7.1 comes pretty close in terms of standards-compliance (although
it has quite a few non-standard extensions), so does Comeau. But you
are basically right with your statements. We do not discuss any
specific implementation, but C++ 'as it should be'. So if you have
issues that are specific to a certain compiler, and not to C++ as a
language, this isn't the place for them. The only exception of course
is when you have a question of the sort 'This doesn't compile in my
compiler, should it according to the standard or am I doing something
wrong?'

VC7.1 is so new the ink is still wet on the boxes. If you want to limit
discussion here to those using compilers that are 100% bang-up to date with
the latest language hacks, I think you'll be talking to yourself; which,
perhaps, is what you would prefer?

[Tim Clacy]

osmium wrote:

Tim Clacy writes:

We talk about everything that is mentioned in the C++ standard,

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is
purely for abstract, theoretical language issues and not practical
issues with actual language implementations?

As I see it, discussion is limited to the latest best effort of the
major compilers, rather than to the standard. The major out standing
issue is templated code and how object files magically appear. But
there may be some wiggle room specifically included in the standard
on that issue. Whether there is official wiggle room or not, there
is a Gentlemen's Agreement that some things are OK and some are not.
After immersion in the culture for a while, you will kind of absorb,
by osmosis or something, what the rules really are.

I looked at your other post you refer to. The battle you are
fighting is that all old compilers were supposed to have been put in
a bonfire and burned when the standard was approved. You appear to
have one that escaped destruction. Think smallpox.

osmium,

I don't think I'm the only one in the battle; there are millions of us.
Forums like this are our refuge; it bites hard if we're not let in because
the generals don't supply us with regulation shoe-laces :-)

[Thomas Wintschel]

"Tim Clacy" <no*******@nospamphaseone.nospamdk> wrote in message
news:3f*********************@news.dk.uu.net...

osmium wrote:

Tim Clacy writes:

We talk about everything that is mentioned in the C++ standard,

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is
purely for abstract, theoretical language issues and not practical
issues with actual language implementations?

As I see it, discussion is limited to the latest best effort of the
major compilers, rather than to the standard. The major out standing
issue is templated code and how object files magically appear. But
there may be some wiggle room specifically included in the standard
on that issue. Whether there is official wiggle room or not, there
is a Gentlemen's Agreement that some things are OK and some are not.
After immersion in the culture for a while, you will kind of absorb,
by osmosis or something, what the rules really are.

I looked at your other post you refer to. The battle you are
fighting is that all old compilers were supposed to have been put in
a bonfire and burned when the standard was approved. You appear to
have one that escaped destruction. Think smallpox.

osmium,

I don't think I'm the only one in the battle; there are millions of us.
Forums like this are our refuge; it bites hard if we're not let in because
the generals don't supply us with regulation shoe-laces :-)

The thing is, if you are looking for ways to deal with the idiosyncracies
of a particular compiler, you are better off asking people who *are*
using that compiler (in a newsgroup devoted to it) than asking people
who *may be* using that compiler (in some other place).

Given the fact that a lot of people new to the language make use of
this group, it is important to constantly draw the line as to where the
language leaves off and the file systems/gui libraries/compiler directives
etc. etc. begin.

Tom

[Unforgiven]

Tim Clacy wrote:

I don't know where you're from, but you should come visit Earth some
time; it's an awful place where, sometimes, you have no choice but to
work with tools that are flawed. I very much like the sound of your
world though; how do you get there? Do you do any commercial
development there or is it purely academic?

Yes I do commercial development, and I am quite aware of limitations that
can be in place due to corporate policy. However, I think you missed my
joke. I said that if you *need* namespaces you'd need a new compiler. If you
can work around the issue with your current compiler, you don't really need
them, hence my statement.

--
Unforgiven

[jeffc]

"Tim Clacy" <no*******@nospamphaseone.nospamdk> wrote in message
news:3f*********************@news.dk.uu.net...

I was advised in a recent thread that, because one of the compilers that I
use does not support 'namespace', I should not use this group but ask
elsewhere; perhaps comp.lang.c++.but_without_namespace'? What's the general consensus?

Of course you should ask here. There are some "standards police" who don't
understand that C++ is a language implemented by compilers. Ther is no such
thing as a perfect compiler. Also, that we do not just discuss C++ as it
exists in the standard today, but C++ as it has existed over the years.

[jeffc]

"Tim Clacy" <no*******@nospamphaseone.nospamdk> wrote in message
news:3f*********************@news.dk.uu.net...

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is purely for abstract, theoretical language issues and not practical issues with actual
language implementations?

Yes, that's what he's suggesting. He's wrong though.

[jeffc]

"Thomas Wintschel" <th******@telus.net> wrote in message
news:V%1Db.7993$OJ.4485@edtnps84...

Given the fact that a lot of people new to the language make use of
this group, it is important to constantly draw the line as to where the
language leaves off and the file systems/gui libraries/compiler directives
etc. etc. begin.

As long as the "new people" understand it's not standard, and that the
solution is a C++ solution, there shouldn't be a problem.

[jeffc]

"Tim Clacy" <no*******@nospamphaseone.nospamdk> wrote in message
news:3f*********************@news.dk.uu.net...

I very much like the sound of your world though; how
do you get there?
You're here.
Do you do any commercial development there or is it purely academic?

Rhetorical question, right?

[tom_usenet]

On Sun, 14 Dec 2003 15:33:16 +0100, "Tim Clacy"
<no*******@nospamphaseone.nospamdk> wrote:

Unforgiven wrote:

Tim Clacy wrote:

I was advised in a recent thread that, because one of the compilers
that I use does not support 'namespace', I should not use this group
but ask elsewhere; perhaps comp.lang.c++.but_without_namespace'?
What's the general consensus?

We talk about everything that is mentioned in the C++ standard,

THE C++ standard? My understanding is that even VC 7.0 and gcc aren't
compliant in all respects. Are you suggesting that this group is purely for
abstract, theoretical language issues and not practical issues with actual
language implementations?

Maybe 90% of the code for most programs can be written to be entirely
standards compliant (and even portable) as long as you have a
reasonably recent compiler (even MSVC6!). For the other 10% of code,
including platform libraries and APIs, compiler workarounds, etc.,
etc., it makes sense to direct your question to an appropriate group
handling that compiler. Some compilers don't have active usenet
groups, but there are usually mailing lists (for, say, gcc) or tech
support people.

If you want a standard compliant workaround to avoid a particular
standard feature that your compiler doesn't support, I'd certainly
have no problem answering such a question here, so long as the
workarounds aren't too labourious. Others may choose to differ (and
even flame such questions), and that's their privilege.

Tom

C++ FAQ: http://www.parashift.com/c++-faq-lite/
C FAQ: http://www.eskimo.com/~scs/C-faq/top.html

[Jerry Coffin]

In article <3f*********************@news.dk.uu.net>,
no*******@nospamphaseone.nospamdk says...

[ ... ]

VC7.1 is so new the ink is still wet on the boxes.
VC++ 7.1 was officially released in February, and IIRC, MSDN subscribers
could download it around a month before that. IOW, it's a couple of
weeks short of a year old.
If you want to limit
discussion here to those using compilers that are 100% bang-up to date with
the latest language hacks, I think you'll be talking to yourself; which,
perhaps, is what you would prefer?

Comeau C++ has been even closer to conforming since quite a while before
VC++ 7.1 came out. Furthermore, it's dirt cheap and easily available
for nearly every platform around (and I believe he'll do custom ports to
other platforms as well).

To make a long story short, compilers that are at least awfully close to
conforming have been widely available for quite a while now. I'm not
saying every discussion has to use every obscure feature to be
interesting, but I also don't think compiler availability is a very good
reason to limit discussions either.

--
Later,
Jerry.

The universe is a figment of its own imagination.