Including ATL 7.1 makes the linker to generate almost eight hundred LNK4254 and LNK4078 warnings. T…

ATL 8.0 included in WDK 7.1 makes the linker to generate almost eight hundred LNK4254 and LNK4078 warnings. This CL disables both warnings when compiling using "Express" versions of Visual Studio.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=128082

[alexeypa (please no reviews), 2012-03-19 01:33:36]

PTAL

[alexeypa (please no reviews), 2012-03-20 15:48:45]

ping

[alexeypa (please no reviews), 2012-03-21 17:40:03]

Marc-Antoine, could you please take a look?

This, probably, is worth some explanation. ATL 7.1 has a bug that cause the
linker to generate LNK4254 warnings. Here is other folks seeing it:
http://social.msdn.microsoft.com/Forums/sr-Latn-CS/vcmfcatl/thread/4763bc2d-0....
There are a few other stories if you google for "LNK4254".

These warnings are useless, since nobody is going to fix them (unless a fixed
ATL will come out). We have almost 800 of them when compiling all.sln. This CL
tells the linker to shut up.

[M-A Ruel, 2012-03-21 18:22:44]

On 2012/03/21 17:40:03, alexeypa wrote:
> Marc-Antoine, could you please take a look?
> 
> This, probably, is worth some explanation. ATL 7.1 has a bug that cause the
> linker to generate LNK4254 warnings. Here is other folks seeing it:
>
http://social.msdn.microsoft.com/Forums/sr-Latn-CS/vcmfcatl/thread/4763bc2d-0....
> There are a few other stories if you google for "LNK4254".
> 
> These warnings are useless, since nobody is going to fix them (unless a fixed
> ATL will come out). We have almost 800 of them when compiling all.sln. This CL
> tells the linker to shut up.

Didn't we officially drop Visual Studio 2005 support?

[alexeypa (please no reviews), 2012-03-21 18:33:10]

On 2012/03/21 18:22:44, Marc-Antoine Ruel wrote:
> Didn't we officially drop Visual Studio 2005 support?

This is for VS 2010. ATL 7.1. is the latest version of ATL shipped with VS 2010
and WDK 7.1.

[M-A Ruel, 2012-03-21 19:00:03]

?

From C:\Program Files (x86)\Microsoft Visual Studio
9.0\VC\atlmfc\include\atldef.h
#define _ATL_VER 0x0900 // Active Template Library version 9.00

From C:\Program Files (x86)\Microsoft Visual Studio
10.0\VC\atlmfc\include\atldef.h:
#define _ATL_VER 0x0A00 // Active Template Library version 10.00

So 7.1 is like VS 2003 IIRC?

[alexeypa (please no reviews), 2012-03-21 19:15:42]

On 2012/03/21 19:00:03, Marc-Antoine Ruel wrote:
> ?
> 
> From C:\Program Files (x86)\Microsoft Visual Studio
> 9.0\VC\atlmfc\include\atldef.h
> #define _ATL_VER 0x0900 // Active Template Library version 9.00
> 
> From C:\Program Files (x86)\Microsoft Visual Studio
> 10.0\VC\atlmfc\include\atldef.h:
> #define _ATL_VER 0x0A00 // Active Template Library version 10.00
> 
> So 7.1 is like VS 2003 IIRC?

Oops. You are right.

ATL 7.1 that comes with WDK 7.1 is actually 8.0:

> #define _ATL_VER 0x0800 // Active Template Library version 8.00

I was sure it it the same one that is included to Visual Studio but, apparently,
I was wrong.

I use ATL from WDK because I'm using the express edition of VS2010. I think most
of other people building Chrome use full size VS editions.

Should we punt this? What do you think?

[M-A Ruel, 2012-03-21 19:25:33]

On 2012/03/21 19:15:42, alexeypa wrote:
> ATL 7.1 that comes with WDK 7.1 is actually 8.0:
> 
> > #define _ATL_VER 0x0800 // Active Template Library version 8.00

Please fix the patch and description.

> I use ATL from WDK because I'm using the express edition of VS2010. I think
most
> of other people building Chrome use full size VS editions.

My preferred way of supporting non-Pro users is to drop the ATL requirement
entirely. Nobody has yet taken up the challenge in 4 years. I'm quite saddened
by that. I had expected crazier people out there.

But I'm fine to disable the warning, only when the express version is detected.

M-A

[alexeypa (please no reviews), 2012-03-21 21:49:23]

On 2012/03/21 19:25:33, Marc-Antoine Ruel wrote:
> My preferred way of supporting non-Pro users is to drop the ATL requirement
> entirely. Nobody has yet taken up the challenge in 4 years. I'm quite saddened
> by that. I had expected crazier people out there.

:-) Yeah, there is only 112 files including atlbase.h.

> But I'm fine to disable the warning, only when the express version is
detected.

Done.

[alexeypa (please no reviews), 2012-03-21 21:49:23]

On 2012/03/21 19:25:33, Marc-Antoine Ruel wrote:
> My preferred way of supporting non-Pro users is to drop the ATL requirement
> entirely. Nobody has yet taken up the challenge in 4 years. I'm quite saddened
> by that. I had expected crazier people out there.

:-) Yeah, there is only 112 files including atlbase.h.

> But I'm fine to disable the warning, only when the express version is
detected.

Done.

[M-A Ruel, 2012-03-21 21:50:44]

lgtm

[commit-bot: I haz the power, 2012-03-21 21:52:20]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/alexeypa@chromium.org/9718030/8001

[commit-bot: I haz the power, 2012-03-21 22:59:37]

Change committed as 128082

[scottmg1, 2012-03-22 17:54:22]

On 2012/03/21 21:49:23, alexeypa wrote:
> On 2012/03/21 19:25:33, Marc-Antoine Ruel wrote:
> > My preferred way of supporting non-Pro users is to drop the ATL requirement
> > entirely. Nobody has yet taken up the challenge in 4 years. I'm quite
saddened
> > by that. I had expected crazier people out there.
> 
> :-) Yeah, there is only 112 files including atlbase.h.
> 
> > But I'm fine to disable the warning, only when the express version is
> detected.
> 
> Done.

Apologies, was OoO.

Is there a bug for killing ATL entirely?