2003/2005 conlfict

Using MSVC2003.net

std::vector<std::vector<int>> foo;

causes an error due to the >> that should be > >

but it compiles fine with MSVC2005.net. Is there a setting to
turn this off in 2005?

Carl
Thu Jan 10 09:14:39 CST 2008

Duane Hebert wrote:
> Using MSVC2003.net
>
> std::vector<std::vector<int>> foo;
>
> causes an error due to the >> that should be > >
>
> but it compiles fine with MSVC2005.net. Is there a setting to
> turn this off in 2005?

Not that I'm aware of.

-cd

Duane
Thu Jan 10 09:25:34 CST 2008


"Carl Daniel [VC++ MVP]" <cpdaniel_remove_this_and_nospam@mvps.org.nospam>
wrote in message news:uNc7Fu5UIHA.5164@TK2MSFTNGP03.phx.gbl...
> Duane Hebert wrote:
>> Using MSVC2003.net
>>
>> std::vector<std::vector<int>> foo;
>>
>> causes an error due to the >> that should be > >
>>
>> but it compiles fine with MSVC2005.net. Is there a setting to
>> turn this off in 2005?
>
> Not that I'm aware of.

Disabling extensions completely doesn't seem to make
this work properly either. I wonder if 2008 compiles this
as well.

Ben
Thu Jan 10 11:24:34 CST 2008


"Duane Hebert" <spoo@flarn.com> wrote in message
news:uzAGO05UIHA.4448@TK2MSFTNGP03.phx.gbl...
>
> "Carl Daniel [VC++ MVP]" <cpdaniel_remove_this_and_nospam@mvps.org.nospam>
> wrote in message news:uNc7Fu5UIHA.5164@TK2MSFTNGP03.phx.gbl...
>> Duane Hebert wrote:
>>> Using MSVC2003.net
>>>
>>> std::vector<std::vector<int>> foo;
>>>
>>> causes an error due to the >> that should be > >
>>>
>>> but it compiles fine with MSVC2005.net. Is there a setting to
>>> turn this off in 2005?
>>
>> Not that I'm aware of.
>
> Disabling extensions completely doesn't seem to make
> this work properly either. I wonder if 2008 compiles this
> as well.

I believe that the C++ standards committee changed the syntax very slightly
to make that work, and the newer versions implement the new parser.

Duane
Thu Jan 10 13:27:38 CST 2008


"Ben Voigt [C++ MVP]" <rbv@nospam.nospam> wrote in message
news:eM0h416UIHA.1164@TK2MSFTNGP02.phx.gbl...
>
> "Duane Hebert" <spoo@flarn.com> wrote in message
> news:uzAGO05UIHA.4448@TK2MSFTNGP03.phx.gbl...
>>
>> "Carl Daniel [VC++ MVP]"
>> <cpdaniel_remove_this_and_nospam@mvps.org.nospam> wrote in message
>> news:uNc7Fu5UIHA.5164@TK2MSFTNGP03.phx.gbl...
>>> Duane Hebert wrote:
>>>> Using MSVC2003.net
>>>>
>>>> std::vector<std::vector<int>> foo;
>>>>
>>>> causes an error due to the >> that should be > >
>>>>
>>>> but it compiles fine with MSVC2005.net. Is there a setting to
>>>> turn this off in 2005?
>>>
>>> Not that I'm aware of.
>>
>> Disabling extensions completely doesn't seem to make
>> this work properly either. I wonder if 2008 compiles this
>> as well.
>
> I believe that the C++ standards committee changed the syntax very
> slightly to make that work, and the newer versions implement the new
> parser.

I was under the impression that this would be true with the next
"version" of the c++ standard. Has that one been officially
released?

It would be good but I'm only even concerned due to some cross platform
requirements so until all of the compilers that we use are compliant...

Bo
Thu Jan 10 14:41:33 CST 2008

Duane Hebert wrote:
> "Ben Voigt [C++ MVP]" <rbv@nospam.nospam> wrote in message
> news:eM0h416UIHA.1164@TK2MSFTNGP02.phx.gbl...
>>
>> "Duane Hebert" <spoo@flarn.com> wrote in message
>> news:uzAGO05UIHA.4448@TK2MSFTNGP03.phx.gbl...
>>>
>>> "Carl Daniel [VC++ MVP]"
>>> <cpdaniel_remove_this_and_nospam@mvps.org.nospam> wrote in message
>>> news:uNc7Fu5UIHA.5164@TK2MSFTNGP03.phx.gbl...
>>>> Duane Hebert wrote:
>>>>> Using MSVC2003.net
>>>>>
>>>>> std::vector<std::vector<int>> foo;
>>>>>
>>>>> causes an error due to the >> that should be > >
>>>>>
>>>>> but it compiles fine with MSVC2005.net. Is there a setting to
>>>>> turn this off in 2005?
>>>>
>>>> Not that I'm aware of.
>>>
>>> Disabling extensions completely doesn't seem to make
>>> this work properly either. I wonder if 2008 compiles this
>>> as well.
>>
>> I believe that the C++ standards committee changed the syntax very
>> slightly to make that work, and the newer versions implement the
>> new parser.
>
> I was under the impression that this would be true with the next
> "version" of the c++ standard. Has that one been officially
> released?

No, it has not. Hopefully next year.

The usual complaint is that compilers are late in implementing new
features. Here they are too early? :-)

To be more serious, the compiler accepts the formally correct syntax,
which is important. Having extensions as well is allowed.

>
> It would be good but I'm only even concerned due to some cross
> platform requirements so until all of the compilers that we use are
> compliant...

It depends on what platforms you intend to cross between. There are
other compilers accepting this as well.


Bo Persson

Ben
Thu Jan 10 15:51:59 CST 2008


"Duane Hebert" <spoo@flarn.com> wrote in message
news:ev8id77UIHA.4440@TK2MSFTNGP06.phx.gbl...
>
> "Ben Voigt [C++ MVP]" <rbv@nospam.nospam> wrote in message
> news:eM0h416UIHA.1164@TK2MSFTNGP02.phx.gbl...
>>
>> "Duane Hebert" <spoo@flarn.com> wrote in message
>> news:uzAGO05UIHA.4448@TK2MSFTNGP03.phx.gbl...
>>>
>>> "Carl Daniel [VC++ MVP]"
>>> <cpdaniel_remove_this_and_nospam@mvps.org.nospam> wrote in message
>>> news:uNc7Fu5UIHA.5164@TK2MSFTNGP03.phx.gbl...
>>>> Duane Hebert wrote:
>>>>> Using MSVC2003.net
>>>>>
>>>>> std::vector<std::vector<int>> foo;
>>>>>
>>>>> causes an error due to the >> that should be > >
>>>>>
>>>>> but it compiles fine with MSVC2005.net. Is there a setting to
>>>>> turn this off in 2005?
>>>>
>>>> Not that I'm aware of.
>>>
>>> Disabling extensions completely doesn't seem to make
>>> this work properly either. I wonder if 2008 compiles this
>>> as well.
>>
>> I believe that the C++ standards committee changed the syntax very
>> slightly to make that work, and the newer versions implement the new
>> parser.
>
> I was under the impression that this would be true with the next
> "version" of the c++ standard. Has that one been officially
> released?

That's why I carefully avoided saying the syntax had been changed in the
standard.

As I understand it, the committee has officially approved the change. The
set of changes comprising the next "version" aren't complete yet.

>
> It would be good but I'm only even concerned due to some cross platform
> requirements so until all of the compilers that we use are compliant...
>

Duane
Thu Jan 10 17:45:03 CST 2008

>> I was under the impression that this would be true with the next
>> "version" of the c++ standard. Has that one been officially
>> released?
>
> No, it has not. Hopefully next year.
>
> The usual complaint is that compilers are late in implementing new
> features. Here they are too early? :-)

Yep.

> To be more serious, the compiler accepts the formally correct syntax,
> which is important. Having extensions as well is allowed.

Yep. I haven't had much of a problem so far with allowing extensions.

>>
>> It would be good but I'm only even concerned due to some cross
>> platform requirements so until all of the compilers that we use are
>> compliant...
>
> It depends on what platforms you intend to cross between. There are other
> compilers accepting this as well.

As far as I can see currently just Commeau and MS > 7.1 seem to accept
this. But that puts MS in good company IMO.

Oddly enough, we found it because one of our developers modified some
common source using VS8 and when another developer checked in the
SVN it failed with their VS7.1 build. So it was basically an MS to MS
cross.

Duane
Thu Jan 10 17:45:52 CST 2008

>> I was under the impression that this would be true with the next
>> "version" of the c++ standard. Has that one been officially
>> released?
>
> That's why I carefully avoided saying the syntax had been changed in the
> standard.
>
> As I understand it, the committee has officially approved the change. The
> set of changes comprising the next "version" aren't complete yet.

Ok. Thanks for the input.