gamedevlists-windows — List for discussing game development on the Windows platform

RE: [GD-Windows] WA_CLICKACTIVE vs. WM_MOUSEACTIVATE

[Nathan R. <Nat...@te...>]

Are you even recieving WM_NCHITTEST messages? If anything, that might prove
useful in tracking down which window is the culprit..

-----Original Message-----
From: Brian Hook [mailto:bri...@py...]
Sent: Tuesday, January 29, 2002 5:40 PM
To: gam...@li...
Subject: RE: [GD-Windows] WA_CLICKACTIVE vs. WM_MOUSEACTIVATE


> Oddly enough, I'm also not receiving WA_CLICKACTIVE in 
> WM_ACTIVATE...curiouser and curiouser.

Okay, it's gone from curious to annoying.  If I click on my app's title
bar, I get WM_MOUSEACTIVATE and WM_ACTIVATE/WA_CLICKACTIVE.  If I click
in the client area, however, I only get WM_ACTIVATE/WA_ACTIVE.  It's a
pretty vanilla window.  The only thing I can think of is that some of
the documentation throws around "Child window" as a term a lot, and
since the only child windows I have are created by Windows (title bar,
system menu), that my problem lies with that.

Brian


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

RE: [GD-Windows] WA_CLICKACTIVE vs. WM_MOUSEACTIVATE

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

> Oddly enough, I'm also not receiving WA_CLICKACTIVE in 
> WM_ACTIVATE...curiouser and curiouser.

Okay, it's gone from curious to annoying.  If I click on my app's title
bar, I get WM_MOUSEACTIVATE and WM_ACTIVATE/WA_CLICKACTIVE.  If I click
in the client area, however, I only get WM_ACTIVATE/WA_ACTIVE.  It's a
pretty vanilla window.  The only thing I can think of is that some of
the documentation throws around "Child window" as a term a lot, and
since the only child windows I have are created by Windows (title bar,
system menu), that my problem lies with that.

Brian

RE: [GD-Windows] WA_CLICKACTIVE vs. WM_MOUSEACTIVATE

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

Oddly enough, I'm also not receiving WA_CLICKACTIVE in
WM_ACTIVATE...curiouser and curiouser.


> -----Original Message-----
> From: gam...@li... 
> [mailto:gam...@li...] On 
> Behalf Of Brian Hook
> Sent: Tuesday, January 29, 2002 3:01 PM
> To: gam...@li...
> Subject: [GD-Windows] WA_CLICKACTIVE vs. WM_MOUSEACTIVATE
> 
> 
> Here's a weird one.  I'm trying to change my app so that when 
> a mouse click activates a window I automatically discard the 
> mouse click (so that unpausing doesn't inadvertently allow 
> the user to exit the game, restart, etc.).
> 
> I was originally going to do this programmatically by letting 
> my main app look at the timestamps of the last activation 
> message vs. incoming mouse down events, but the delta can 
> vary by a significant amount of time (at least 2ms).  Trying 
> to make this work by altering the tolerance seems iffy at best.
> 
> The next thing I decided to look at was WA_CLICKACTIVE, but 
> while looking at that I found WM_MOUSEACTIVATE and returning 
> MA_ACTIVATEANDEAT.  Unfortunately, my app NEVER sees a 
> WM_MOUSEACTIVATE
> -- I've confirmed this with Spy++.  This is on Win2K.  
> However, someone else I know sees this event just fine -- any 
> ideas what's going on here?
> 
> Failing WM_MOUSEACTIVATE, I'll look into WA_CLICKACTIVE and 
> set a flag to ignore the next mouse message, but now I'm 
> curious what's up with WM_MOUSEACTIVATE.
> 
> Brian
> 
> 
> _______________________________________________
> Gamedevlists-windows mailing list 
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-windows
> 

[GD-Windows] WA_CLICKACTIVE vs. WM_MOUSEACTIVATE

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

Here's a weird one.  I'm trying to change my app so that when a mouse
click activates a window I automatically discard the mouse click (so
that unpausing doesn't inadvertently allow the user to exit the game,
restart, etc.).

I was originally going to do this programmatically by letting my main
app look at the timestamps of the last activation message vs. incoming
mouse down events, but the delta can vary by a significant amount of
time (at least 2ms).  Trying to make this work by altering the tolerance
seems iffy at best.

The next thing I decided to look at was WA_CLICKACTIVE, but while
looking at that I found WM_MOUSEACTIVATE and returning
MA_ACTIVATEANDEAT.  Unfortunately, my app NEVER sees a WM_MOUSEACTIVATE
-- I've confirmed this with Spy++.  This is on Win2K.  However, someone
else I know sees this event just fine -- any ideas what's going on here?

Failing WM_MOUSEACTIVATE, I'll look into WA_CLICKACTIVE and set a flag
to ignore the next mouse message, but now I'm curious what's up with
WM_MOUSEACTIVATE.

Brian

Re: [GD-Windows] Installers?

[Rich <leg...@xm...>]

In article <200...@ys...>,
    "Steve Maier" <ma...@ys...>  writes:

> I use Nullsoft's NSIS. [...]
> It is quite powerful, but the only drawback is that you have to
> write the script file yourself.

Note: if you want Windows 2000 logo certification, you -must- use
Windows Installer-based installations.  The requirements for using
Windows Installer will only increase over time as MS has been forced
to 'own' the installation process due to the many user complaints
about poor install programs that corrupt the integrity of their
system.

Windows Installer is the future of installation programs on Windows.
Investing brain space in anything else is a dead end, IMO.
--
Ask me about my upcoming book on Direct3D from Addison-Wesley!
  Direct3D Book <http://www.xmission.com/~legalize/book/>
    Don't Support Spammers!  Boycott Fractal Painter 7!
     <http://www.xmission.com/~legalize/spammers.html>

Re: [GD-Windows] Installers?

[Steve M. <ma...@ys...>]

I use Nullsoft's NSIS.  Its free, you even get full source for it.  You can do the registry changes, register COM objects, you can even copy and rename files if you want.  It is quite powerful, but the only drawback is that you have to write the script file yourself.  Its an easy format and they do provide a few examples.

http://www.nullsoft.com/free/nsis/

Steve Maier, MCSD

---------- Original Message ----------------------------------
From: Kent Quirk <ken...@co...>
Date:  Mon, 28 Jan 2002 14:40:59 -0500

>We've used installshield for a long time, and frankly, I hate it. I
>haven't even been the one to deal with it, and I hate it. I especially
>hate that it doesn't seem to be very automatable -- I'd like our
>installer to get built from a makefile or the equivalent.
>
>Our install requirements aren't that tough -- make a couple of simple
>decisions, stick a couple of values in a registry, and copy some files.
>It's more than a zipfile extractor can do, but not too hard. 
>
>Anybody got some alternatives for me?
>
>It'd be especially nice if it was cheap, but that's not an absolute
>requirement.
>
>	Kent
>
>-- 
>-----------------------------------------------------------------------
>Kent Quirk                   | MindRover: "Astonishingly creative."
>Game Architect               | Check it out!
>ken...@co...      | http://www.mindrover.com/
>_____________________________|_________________________________________
>
>_______________________________________________
>Gamedevlists-windows mailing list
>Gam...@li...
>https://lists.sourceforge.net/lists/listinfo/gamedevlists-windows
>

Re: [GD-Windows] Installers?

[Rich <leg...@xm...>]

In article <3C5...@co...>,
    Kent Quirk <ken...@co...>  writes:

> Our install requirements aren't that tough -- make a couple of simple
> decisions, stick a couple of values in a registry, and copy some files.
> It's more than a zipfile extractor can do, but not too hard. 
> 
> Anybody got some alternatives for me?

See <http://izfree.sourceforge.net/> for some Windows Installer based
tools that would do the job.  Specifically, Visual Studio Installer
can easily do what you require.  That's an MS-supplied free download
for Visual Studio users.
--
Ask me about my upcoming book on Direct3D from Addison-Wesley!
  Direct3D Book <http://www.xmission.com/~legalize/book/>
    Don't Support Spammers!  Boycott Fractal Painter 7!
     <http://www.xmission.com/~legalize/spammers.html>

[GD-Windows] Installers?

[Kent Q. <ken...@co...>]

We've used installshield for a long time, and frankly, I hate it. I
haven't even been the one to deal with it, and I hate it. I especially
hate that it doesn't seem to be very automatable -- I'd like our
installer to get built from a makefile or the equivalent.

Our install requirements aren't that tough -- make a couple of simple
decisions, stick a couple of values in a registry, and copy some files.
It's more than a zipfile extractor can do, but not too hard. 

Anybody got some alternatives for me?

It'd be especially nice if it was cheap, but that's not an absolute
requirement.

	Kent

-- 
-----------------------------------------------------------------------
Kent Quirk                   | MindRover: "Astonishingly creative."
Game Architect               | Check it out!
ken...@co...      | http://www.mindrover.com/
_____________________________|_________________________________________

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

[Matt D. <ma...@co...>]

I think VC7 supports template template parameters.  But partial template
specialisation is not implemented yet.  However, I have successfully
rewritten parts of Loki to work without PTS, although a couple of my
typelist structures do not work 100% (its a bugger debugging them!!) but I
basically did it through nested structs and major use of the Loki Select
structure, which I also had to convert.

Matt Davies
Programmer, Confounding Factor
ma...@co...
www.confounding-factor.com



-----Original Message-----
From: gam...@li...
[mailto:gam...@li...]On Behalf Of
Wesley Hunt
Sent: 28 January 2002 07:13
To: gam...@li...
Subject: Re: [GD-Windows] Visual C++ .Net


> There are other features of the standard that are not supported in Visual
C
> 6 that I can live without, such as template template parameters (which is
> implemented in VC 7 BTW) and differentiation of void, const void, volatile
> void and const volatile void (needed for Boost).  But partial template
> specialisation is extremely important to me and I wish Microsoft would put
> in the effort to ensure that Visual C uses it.  How long did they work on
> Visual C 7?  There are just some things that need partial template
> specialisation and so I've been using compilers such as Borland C++,
> Metrowerks and GCC 3.0.  BTW, am I the only person who sees similarities
of
> meta-programming to writing Prolog programs?
>
> For those developers who don't see the usefulness of partial template
> specialisation, check out the book Modern C++ Design
>
(http://s1.amazon.co.uk/exec/varzea/ts/exchange-glance/Y02Y6828407Y6383067/q
> id=1011868065/sr=1-1/ref=sr_aps_z_2_1/202-0065032-5172623) or the website
> www.moderncppdesign.com and look at the Loki library, which is simpler to
> understand than Boost.  Perhaps more developers will scream out for it and
> perhaps Microsoft will actually listen to this one.
>
Doesn't the Stan Lipmann interview mention MS's commitment to making popular
libaries like blitz, boost, loki, etc. compile in VC7? Partial template
specialization and template template parameters are essential to loki, as a
quick perusing of Modern C++ Design will show. Does that mean I can't
compile loki with the first release of VC7, and will have to wait for the
first service pack?

-Wes



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

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

[Wesley H. <Wes...@ir...>]

> There are other features of the standard that are not supported in Visual
C
> 6 that I can live without, such as template template parameters (which is
> implemented in VC 7 BTW) and differentiation of void, const void, volatile
> void and const volatile void (needed for Boost).  But partial template
> specialisation is extremely important to me and I wish Microsoft would put
> in the effort to ensure that Visual C uses it.  How long did they work on
> Visual C 7?  There are just some things that need partial template
> specialisation and so I've been using compilers such as Borland C++,
> Metrowerks and GCC 3.0.  BTW, am I the only person who sees similarities
of
> meta-programming to writing Prolog programs?
>
> For those developers who don't see the usefulness of partial template
> specialisation, check out the book Modern C++ Design
>
(http://s1.amazon.co.uk/exec/varzea/ts/exchange-glance/Y02Y6828407Y6383067/q
> id=1011868065/sr=1-1/ref=sr_aps_z_2_1/202-0065032-5172623) or the website
> www.moderncppdesign.com and look at the Loki library, which is simpler to
> understand than Boost.  Perhaps more developers will scream out for it and
> perhaps Microsoft will actually listen to this one.
>
Doesn't the Stan Lipmann interview mention MS's commitment to making popular
libaries like blitz, boost, loki, etc. compile in VC7? Partial template
specialization and template template parameters are essential to loki, as a
quick perusing of Modern C++ Design will show. Does that mean I can't
compile loki with the first release of VC7, and will have to wait for the
first service pack?

-Wes

RE: [GD-Windows] (GDI) "Flashing white" in window

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

> I assume you are dealing with WM_ERASEBKGND?

And you would have assumed wrong =) Sweet, that fixed it, thanks!

Brian

RE: [GD-Windows] (GDI) "Flashing white" in window

[Andy G. <an...@mi...>]

Ancient? :-)  seems like only yesterday -sigh-

Maybe spy on your window with Spy+ and see what messages are coming to
you and make sure you handle all of them properly.

I assume you are dealing with WM_ERASEBKGND? You should just return true
to this - this can cause these flashes to the background color - try
altering the background color of windows in the display appearance tab -
see if it alters your flashing color.

Andy Glaister

-----Original Message-----
From: Brian Hook [mailto:bri...@py...]=20
Sent: Saturday, January 26, 2002 12:37 PM
To: gam...@li...
Subject: [GD-Windows] (GDI) "Flashing white" in window


I don't expect to see many responses to this, primarily because I doubt
anyone even bothers with GDI anymore, however I thought I'd ask just in
case some of you folks who've done the GDI bit might have some kind of
primordial memory of the subject.

My current blitter is just a typical DIBSection/BitBlt affair, however
I've noticed that on some occasions I get "white flashes" (maybe one
frame), as if an improper set of bits are being blitted.  To track this
down, I force a clear on my backbuffer to red -- if the culprit is my
own code not filling the backbuffer correctly (e.g. I skip a frame or
something like that) I'd expect that the flash would become a "red
flash".  Alas, this isn't the case -- even if I force the backbuffer to
red right before I do a BitBlt(), I still get the occasional white
flash.

If this was DDraw, I'd think that the surface was being lost and not
restored, but since these are just DIBSections (or, more specifically,
just one DIB section) I'm thinking it has to be something else I'm doing
wrong.  It happens on a wide variety of hardware, so I don't think it's
a driver issue.

Anyone have any ancient memories on what might cause this type of thing?

Thanks,

Brian


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

[GD-Windows] (GDI) "Flashing white" in window

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

I don't expect to see many responses to this, primarily because I doubt
anyone even bothers with GDI anymore, however I thought I'd ask just in
case some of you folks who've done the GDI bit might have some kind of
primordial memory of the subject.

My current blitter is just a typical DIBSection/BitBlt affair, however
I've noticed that on some occasions I get "white flashes" (maybe one
frame), as if an improper set of bits are being blitted.  To track this
down, I force a clear on my backbuffer to red -- if the culprit is my
own code not filling the backbuffer correctly (e.g. I skip a frame or
something like that) I'd expect that the flash would become a "red
flash".  Alas, this isn't the case -- even if I force the backbuffer to
red right before I do a BitBlt(), I still get the occasional white
flash.

If this was DDraw, I'd think that the surface was being lost and not
restored, but since these are just DIBSections (or, more specifically,
just one DIB section) I'm thinking it has to be something else I'm doing
wrong.  It happens on a wide variety of hardware, so I don't think it's
a driver issue.

Anyone have any ancient memories on what might cause this type of thing?

Thanks,

Brian

RE: [GD-Windows] Data alignment...

[Wayne C. <wc...@re...>]

Can I just note for anyone else after checking last night VC7 does indeed
support _aligned_malloc( size, alignment ) and _aligned_free( ptr ).

Thanks again,
Wayne
-Virus scanned and cleared ok

RE: [GD-Windows] Data alignment...

[Wayne C. <wc...@re...>]

> When you install the vc6 proc pack, it installs a CHM called 
> procpack.chm ( Strangly enough ) so see if you have it. 

Well, the actual processor pack is a VC6 upgrade not VC7 so I won't have the
'processor pack' but the new support the processor pack adds is built into
VC7.

__declspec( align(#) ) was added with the processor pack in VC6 but comes
built into VC7, as does support for SSE and SSE2 opcodes and intrinsics.

I certainly have the processor pack installed here at work and a quick look
in the .chm as you suggested has found the functions you mention
(_aligned_malloc, _aligned_free, which would explain why my searches didn't
find them :). Hopefully as with the other features these will exist within
VC7 and I'll look for them tonight.

Thanks again!
Wayne
-Virus scanned and cleared ok

RE: [GD-Windows] Data alignment...

[Gareth L. <GL...@cl...>]

When you install the vc6 proc pack, it installs a CHM called procpack.chm (
Strangly enough ) so see if you have it. 

> -----Original Message-----
> From: Wayne Coles [mailto:wc...@re...]
> Sent: 25 January 2002 10:32
> To: 'gam...@li...'
> Subject: RE: [GD-Windows] Data alignment...
> 
> 
> > I'm not sure if vc7 comes with the proccecor pack built in, 
> but if it
> > doesn't you should download it from Microsoft's site.
> > 
> > Then use the new __declspec (align) tags, and if needed use the
> > align_malloc() and align_free() functions.
> 
> Yes it has the processor pack built in and I've flagged the 
> structures with
> __declspec( align( 16 ) ).
> 
> I couldn't find the align_malloc and align_free functions 
> though, thanks for
> the pointer I'll take a look for them tonight when I get 
> home. I wrote my
> own but I'd prefer to use the supplied routines (typically a 
> search for
> align, alignment, malloc, _mm_malloc didn't find anything).
> 
> Thanks again.
> Wayne
> -Virus scanned and cleared ok
> 
> _______________________________________________
> Gamedevlists-windows mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-windows
> 

RE: [GD-Windows] Data alignment...

[Wayne C. <wc...@re...>]

> I'm not sure if vc7 comes with the proccecor pack built in, but if it
> doesn't you should download it from Microsoft's site.
> 
> Then use the new __declspec (align) tags, and if needed use the
> align_malloc() and align_free() functions.

Yes it has the processor pack built in and I've flagged the structures with
__declspec( align( 16 ) ).

I couldn't find the align_malloc and align_free functions though, thanks for
the pointer I'll take a look for them tonight when I get home. I wrote my
own but I'd prefer to use the supplied routines (typically a search for
align, alignment, malloc, _mm_malloc didn't find anything).

Thanks again.
Wayne
-Virus scanned and cleared ok

RE: [GD-Windows] Data alignment...

[Gareth L. <GL...@cl...>]

Hi.

I'm not sure if vc7 comes with the proccecor pack built in, but if it
doesn't you should download it from Microsoft's site.

Then use the new __declspec (align) tags, and if needed use the
align_malloc() and align_free() functions.

Regards, Gareth Lewin

> -----Original Message-----
> From: Wayne Coles [mailto:wc...@re...]
> Sent: 25 January 2002 09:46
> To: 'gam...@li...'
> Subject: [GD-Windows] Data alignment...
> 
> 
> I've been doing some SSE optimisations lately using VC7. Does 
> anyone know of
> a built-in way of allocating data on 16 byte bounds (or 
> preferably user
> specified alignment)?
> 
> I've flagged the structure to be aligned on 16 bytes which works when
> declaring it, but not after calling new (I wouldn't expect 
> malloc to align,
> as it knows nothing of the type of allocation, but I thought 
> new would, ahhh
> I'm so naive :).
> 
> According to an intel doc, there is an intrinsic function 
> _mm_malloc which
> allows you to specify an alignment but VC7 doesn't have this 
> function nor
> can I find something similar.
> 
> I've resorted to writing my own _mm_malloc and overloading 
> operator new and
> delete. Just wondering if I missed something obvious and 
> wasted my time
> (this a home project, so waste of time isn't as bad but just 
> as irritating
> :).
> 
> Thanks,
> Wayne
> -Virus scanned and cleared ok
> 
> _______________________________________________
> Gamedevlists-windows mailing list
> Gam...@li...
> https://lists.sourceforge.net/lists/listinfo/gamedevlists-windows
> 

[GD-Windows] Data alignment...

[Wayne C. <wc...@re...>]

I've been doing some SSE optimisations lately using VC7. Does anyone know of
a built-in way of allocating data on 16 byte bounds (or preferably user
specified alignment)?

I've flagged the structure to be aligned on 16 bytes which works when
declaring it, but not after calling new (I wouldn't expect malloc to align,
as it knows nothing of the type of allocation, but I thought new would, ahhh
I'm so naive :).

According to an intel doc, there is an intrinsic function _mm_malloc which
allows you to specify an alignment but VC7 doesn't have this function nor
can I find something similar.

I've resorted to writing my own _mm_malloc and overloading operator new and
delete. Just wondering if I missed something obvious and wasted my time
(this a home project, so waste of time isn't as bad but just as irritating
:).

Thanks,
Wayne
-Virus scanned and cleared ok

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

[Rich <leg...@xm...>]

In article <CFE...@co...>,
    "Matt Davies" <ma...@co...>  writes:

> Can't you get a plug-in for Visual Studio that provides an interface to CVS?
> 
> Hang on... (frantic searching of web)
> 
> Here we are:
> 
> http://www.geocities.com/kaczoroj/CvsIn/

Well that's new since the last time I looked.  Frankly, I'm not
bothered by going to the WinCVS GUI to do CVS operations.  I'm not
even bothered by going to the command line to do CVS operations.  Like
I said, its not the UI of the version control tool that is important
to me, it is the correctness.  VSS loses on correctnes and there is no
amount of GUI integration that can make up for that.
--
Ask me about my upcoming book on Direct3D from Addison-Wesley!
  Direct3D Book <http://www.xmission.com/~legalize/book/>
    Don't Support Spammers!  Boycott Fractal Painter 7!
     <http://www.xmission.com/~legalize/spammers.html>

Re: Alternatives to MSVC was RE: [GD-Windows] Visual C++ .Net

[Dave S. <dav...@sd...>]

I doubt it.  How many software monopolies do you know? Ha ha!!

Seriously, about 6 months ago I switched from using SGI Unix and gcc/cvd,
to NT MSVC 6.0(NT was not my choice but eh..)  I love it.  It can do so much
more than what I thought I needed.

Now  I want C++ compliance, and MS libraries to be namespaced.
I hate compiler name conflicts!! Grrrrrr.....Rectangle in windows.h??
Give me a break.  First year programmers must love that one.

However, I would love to get the editor to reckognize VI commands.
Then I would be in a golden age!!  (I hope that already exists. :-)

-DaveS


> So, this leads into the next logical question -- are there
> compilers/environments out there that can compete with MSVC as overall
> packages.  And Emacs/gdb/GCC doesn't count =)

Alternatives to MSVC was RE: [GD-Windows] Visual C++ .Net

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

> and so I've been using compilers such as Borland C++, 
> Metrowerks and GCC 3.0.

So, this leads into the next logical question -- are there
compilers/environments out there that can compete with MSVC as overall
packages.  And Emacs/gdb/GCC doesn't count =)

Brian

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

[Matt D. <ma...@co...>]

Can't you get a plug-in for Visual Studio that provides an interface to CVS?

Hang on... (frantic searching of web)

Here we are:

http://www.geocities.com/kaczoroj/CvsIn/


Matt Davies
Programmer, Confounding Factor
ma...@co...
www.confounding-factor.com



-----Original Message-----
From: gam...@li...
[mailto:gam...@li...]On Behalf Of
Rich
Sent: 23 January 2002 20:19
To: gam...@li...
Subject: Re: [GD-Windows] Visual C++ .Net 



In article <000b01c1a445$cce2aa30$8600a8c0@sboberg6>,
    "Stefan Boberg" <sbo...@te...>  writes:

>   SS definitely works [...]

Depends on your definition of "works".  After we had multiple
developers editing a single file and lost many hours of work due to
SS's brain-dead "merge", we switched to CVS.  I can live without
context menus inside visual studio if the underlying source code
control system isn't going to lose my changes.

I would never recommend SS to anyone based on my experience with it.
--
Ask me about my upcoming book on Direct3D from Addison-Wesley!
  Direct3D Book <http://www.xmission.com/~legalize/book/>
    Don't Support Spammers!  Boycott Fractal Painter 7!
     <http://www.xmission.com/~legalize/spammers.html>

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

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

[Matt D. <ma...@co...>]

Its probably true that a lot of C++ developers don't have a clue what
partial template specialisation but I do and its a very useful feature which
is in the standards.  It allows my code to be more robust and type-safe by
checking type traits etc. and also allows the writing of more efficient
code.

There are other features of the standard that are not supported in Visual C
6 that I can live without, such as template template parameters (which is
implemented in VC 7 BTW) and differentiation of void, const void, volatile
void and const volatile void (needed for Boost).  But partial template
specialisation is extremely important to me and I wish Microsoft would put
in the effort to ensure that Visual C uses it.  How long did they work on
Visual C 7?  There are just some things that need partial template
specialisation and so I've been using compilers such as Borland C++,
Metrowerks and GCC 3.0.  BTW, am I the only person who sees similarities of
meta-programming to writing Prolog programs?

For those developers who don't see the usefulness of partial template
specialisation, check out the book Modern C++ Design
(http://s1.amazon.co.uk/exec/varzea/ts/exchange-glance/Y02Y6828407Y6383067/q
id=1011868065/sr=1-1/ref=sr_aps_z_2_1/202-0065032-5172623) or the website
www.moderncppdesign.com and look at the Loki library, which is simpler to
understand than Boost.  Perhaps more developers will scream out for it and
perhaps Microsoft will actually listen to this one.

Just a personal rant, I'll shut up now.


Matt Davies
Programmer, Confounding Factor
ma...@co...
www.confounding-factor.com



-----Original Message-----
From: gam...@li...
[mailto:gam...@li...]On Behalf Of
Stefan Boberg
Sent: 23 January 2002 19:45
To: gam...@li...
Subject: RE: [GD-Windows] Visual C++ .Net


> 	Microsoft will not simply implement features because they
> 	are specified in the standard.  They will implement them
> 	when people want them because they are beneficial."
>
> I'd believe that a whole lot more if it weren't for the fact that the
> entire Windows C++ community has been screaming for better template
> support for several YEARS now and MSVC 7 doesn't have it.

   M$'s reasoning makes perfect business sense. Even though I'd be
interested to see PTS supported, I can see where they are coming from.
Also, I think most developers using C++ simply don't have a clue what
partial template specialisation is and don't really care about it
either.

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





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

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

[Rich <leg...@xm...>]

[stupid list, always redirecting my public response to a private email]

In article <3C4...@co...>,
    Kent Quirk <ken...@co...>  writes:

> You're obscuring the point. What I'm concerned about is that MS is
> deciding for the rest of the industry what features the standards SHOULD
> have included. The feature was added to the standard. Given that many
> features people wanted *weren't* added to the standard, I have a
> presumption that if it's in the standard, it's a target that compiler
> vendors should strive for. 

Have you ever been involved in any standards process?  The assumption
inherint in the above paragraph is that somehow the "most important"
or "best" features were picked for inclusion in the standard.
Standards bodies are more like legislatures with political factions
arguing to get their favorite pork included (or their most hated pork
removed).  This varies from standard to standard based on the
politically motivations of the participants.  Why, I've seen people
argue for weeks about whether tab stops should be 4 or 8 characters in
source code!  To think that just because its a standards body that its
going to be immune from such things is naive, IMO.

> I don't mind a vendor saying "we couldn't get that one in this time
> around, but we're working on it." I do mind them saying "you don't need
> that."

The obvious answer is to keep telling MS that you "need that".
Personally, I think you're reading -way- too much into the interview.
What I read out of the interview is that they are pursuing the
low-hanging fruit first before calling in the Hook and Ladder Truck to
get that one piece of fruit at the very top of the tree.  MS is
responding to requests for improvements/enhancements to the compiler.
The request I've heard most often is for improvements to the standard
library and the template support.  That is what they've stated that
they are working on providing in VC7.  That interview also mentions
that some things aren't going to be done in time for the product
launch, so there will be a service pack release later in the year to
address those items.

AFAIK, there aren't any compilers, free or otherwise, that implement
the export keyword.
--
Ask me about my upcoming book on Direct3D from Addison-Wesley!
  Direct3D Book <http://www.xmission.com/~legalize/book/>
    Don't Support Spammers!  Boycott Fractal Painter 7!
     <http://www.xmission.com/~legalize/spammers.html>