gamedevlists-windows

[GD-Windows] Visual C++ .Net

[Brian H. <bri...@py...>]

This was asked on the algorithms list, but since I haven't seen it posted 
here I thought I'd go ahead and bring it up:

Is anyone here seriously thinking about upgrading their MSVC to MSVC 
.Net?  Information on the latter is very sparse at Microsoft's Web site, 
and -- here's the killer -- even if it offers incremental improvements, 
apparently it won't support generating code for Win95(!).  That is just a 
flat out, no-brainer, deal killer for me because of our particular products.

Anyway, I'm curious as to other people's perception of it.

-Brian

RE: [GD-Windows] Visual C++ .Net

[Stefan B. <sbo...@te...>]

> Is anyone here seriously thinking about upgrading their MSVC to MSVC
> .Net?  Information on the latter is very sparse at Microsoft's Web
site,
> and -- here's the killer -- even if it offers incremental
improvements,
> apparently it won't support generating code for Win95(!).  That is
just a
> flat out, no-brainer, deal killer for me because of our particular
> products.

   The Win95 comment only applies to .NET and DX8.1. If you don't use
these API's you're fine regardless of which compiler you use.

   Apart from that... Well, Visual Studio.NET is a *massive* product,
which contains a lot of cool things. Just a few things off the top of my
head:

---- VC++ v7:

     - The v7 compiler is much more standards-compliant, but still lacks
a few things such as Koenig lookup (I *think*), partial template
specialisation and some other rarely-used gunk. This is apparently
scheduled for the next VC release.

     - The STL implementation is new and improved (although I still use
STLPort4.5 for better cross-platform compatibility). 

     - MFC7 is also a nice step forward (mostly due to better
design-time tools). Good for those quick and dirty tools (although using
C# / the CLR would be preferred where possible).

     - Better code generation.

     - Whole-program optimisation. When using this .obj files actually
contain intermediate representation of the code, and the machine code
gets generated at link-time.

     - Loads of fancy/cool stuff for those poor people who have to write
ActiveX/COM components. Not so interesting to most us perhaps.

---- IDE/debugger:

  - Opinions differ, but I find the new IDE vastly better, and extending
it is almost trivial if you use the .NET framework. You can do some very
nice stuff here. The editors supports collapsing, and IntelliSense is
marginally more clever than VC6. VisualAssist.NET is still worth getting
though.

  - The runtime and debugger support mini-dumps, which is sort of like a
core-dump. When an application crashes and there is no debugger
installed, you get the option to save a crash dump, which can be loaded
back into the VS.NET debugger for further analysis. Very cool. [Although
you don't actually need VS.NET for this... the functionality has been
there for a while in the Platform SDK debugger and DBGHELP.DLL]

---- .NET

  Think what you will about Microsoft, but.NET is amazing. For tools
there is just no other option for me. If only Mono/dotGNU were more
advanced I might even attempt switching over to writing game code using
the CLR. My only problem with it is that there is no support for
STL-like containers.

  For me, it's a very worthwhile upgrade (and a no-brainer since we're
all on MSDN anyway here and thus get it "for free"). It might break some
of your code, but then it's probably because the code was broken in the
first place!

Cheers,
Stef! :)
--
Stefan Boberg - R&D Manager, Team17 Software Ltd.
bo...@te...

RE: [GD-Windows] Visual C++ .Net

[Tom H. <to...@3d...>]

At 01:42 PM 1/22/2002, Stefan Boberg wrote:
>      - The v7 compiler is much more standards-compliant, but still lacks
>a few things such as Koenig lookup (I *think*), partial template
>specialisation and some other rarely-used gunk. This is apparently
>scheduled for the next VC release.

OK. Is the scope of a variable declared in a for loop correct now?

That is

for(int i = 0; i < 5; i++)
{
}
i = 1; // i is out of scope on all compilers except VC 6

Also .. do the Perforce and/or SourseSafe plug-ins still work correctly?

Tom

RE: [GD-Windows] Visual C++ .Net

[Jon W. <hp...@mi...>]

If you turn off compiler extensions, "i" will be out of scope in
VC6, too. The problem with that is that Windows headers certainly
don't compile with language extensions turned off :-(

Cheers,

			/ h+


> -----Original Message-----
> From: gam...@li...
> [mailto:gam...@li...]On Behalf Of
> Tom Hubina
> Sent: Tuesday, January 22, 2002 2:53 PM
> To: gam...@li...
> Subject: RE: [GD-Windows] Visual C++ .Net
>
>
> At 01:42 PM 1/22/2002, Stefan Boberg wrote:
> >      - The v7 compiler is much more standards-compliant, but still lacks
> >a few things such as Koenig lookup (I *think*), partial template
> >specialisation and some other rarely-used gunk. This is apparently
> >scheduled for the next VC release.
>
> OK. Is the scope of a variable declared in a for loop correct now?
>
> That is
>
> for(int i = 0; i < 5; i++)
> {
> }
> i = 1; // i is out of scope on all compilers except VC 6
>
> Also .. do the Perforce and/or SourseSafe plug-ins still work correctly?
>
> Tom
>
>
> _______________________________________________
> Gamedevlists-windows mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-windows
>

RE: [GD-Windows] Visual C++ .Net

[Dirk R. <ri...@ph...>]

P4 did work in beta2 but everytime a project was loaded (and we have quite a
lot in our workspaces) there was an annoying message box telling me that all
the source control info was lost from the project files and that the system
hopes that the scc system can recover from this state. The reason for this
was probably the fact that p4 does not store scc info in the project files.
I filed it as a bug but i have not tried the later vc7 versions.

Dirk

-----Original Message-----
From: gam...@li...
[mailto:gam...@li...]On Behalf Of
Tom Hubina
Sent: Tuesday, January 22, 2002 11:53 PM
To: gam...@li...
Subject: RE: [GD-Windows] Visual C++ .Net


At 01:42 PM 1/22/2002, Stefan Boberg wrote:
>      - The v7 compiler is much more standards-compliant, but still lacks
>a few things such as Koenig lookup (I *think*), partial template
>specialisation and some other rarely-used gunk. This is apparently
>scheduled for the next VC release.

OK. Is the scope of a variable declared in a for loop correct now?

That is

for(int i = 0; i < 5; i++)
{
}
i = 1; // i is out of scope on all compilers except VC 6

Also .. do the Perforce and/or SourseSafe plug-ins still work correctly?

Tom


_______________________________________________
Gamedevlists-windows mailing list
Gam...@li...
https://lists.sourceforge.net/lists/listinfo/gamedevlists-windows

RE: [GD-Windows] Visual C++ .Net

[Stefan B. <sbo...@te...>]

> >      - The v7 compiler is much more standards-compliant, but still 
> >lacks a few things such as Koenig lookup (I *think*), partial 
> >template specialisation and some other rarely-used gunk. This is 
> >apparently scheduled for the next VC release.
> 
> OK. Is the scope of a variable declared in a for loop correct now?

   You can control this by passing the /Zc:forScope option. The default
is the old VC6 behaviour, simply because doing anything else would break
lots of old VC6 code.

> Also .. do the Perforce and/or SourseSafe plug-ins still work 
> correctly?

  SS definitely works, since you get SS6.0c with the Enterprise
Architect edition of VS.NET. I've not tried Perforce with it however, so
I would not know how well that's integrated.

Cheers,
Stef! :)
--
Stefan Boberg - R&D Manager, Team17 Software Ltd. 
bo...@te...