Thread: warning: C4786 identifier was truncated to '255' characters in the debug information

I really REALLY hate VC++ 6.0!! But I'm stuck with it.

From what I understand, there's no way to silence the C4786 warning, even if I add:

Code:

#pragma warning(disable: 4786)	// identifier was truncated to '255' characters in the debug information.

before including the <utility> header...

I have my project set to Treat Warnings As Errors; so once it hits this warning the build stops. I'd like to keep that setting if possible, so I was wondering if there's a way to temporarily turn off that setting (kind of like the #pragma warning(error:...) statements, except with the opposite effect)?

Originally Posted by Bleech

I think you're probably looking for the /IGNORE switch. I usually just add it as a parameter in my build settings (vs2005, but vc6 has the same command line options list in the project settings menu), but you could try using a #pragma comment.

Code:

#pragma comment(linker, "/IGNORE:4786")

I tried that, but it doesn't help.

Originally Posted by matsp

Or use shorter symbols??

--
Mats

Unfortunately I didn't write the <utility> STL header, so that's not an option.

Are you using PCH?
I've never had a problem shutting them up. There is supposedly a real MS bug with disabling this warning (167355), but it doesn't seem to always be triggered.

For non-PCH, just putting the pragma at the top of the file works for me. You can also try the "/wd n" compiler command line option.

gg

I think you should not try to block this warning. It is warning of a legitimate problem, which although it does not affect execution, might make debugging difficult.

I believe in eliminating warnings by fixing them. If they can't be fixed, just let them be. Silencing the warning seems like the wrong thing. Just add a sentence to the build notes saying that this warning is expected.

Apparently VC++ 6.0 doesn't recognize the /wd4786 switch.
I'm not using PCH.

Originally Posted by brewbuck

I think you should not try to block this warning. It is warning of a legitimate problem, which although it does not affect execution, might make debugging difficult.

I believe in eliminating warnings by fixing them. If they can't be fixed, just let them be. Silencing the warning seems like the wrong thing. Just add a sentence to the build notes saying that this warning is expected.

I like to fix warnings also, at least when the warning isn't caused by one of the standard header files from the ancient broken compiler.

I'd like to let the warning appear and then just ignore it, but since I have "Treat Warnings As Errors" enabled, if I get even 1 warning, the build stops and I get no .exe...
I like to use the "Treat Warnings As Errors" in addition to the highest warning level, because it forces people to fix warnings immediately instead of just ignoring it. Plus when you rebuild, you see all the warnings instead of just new ones in the files that were changed.

Well I'll be... adding /IGNORE:4786 to the Linker tab seems to have fixed it (at least for now). But then again this problem only started appearing today out of nowhere, even though the code it's complaining about wasn't changed...

BTW, why would you need to ignore that warning in the Linker tab instead of the Compiler tab? This is a Compiler warning after all...