To declare an object as deprecated in C/C++ under Visual Studio, you have three solutions:
#pragma deprecated(X)__pragma(deprecated(X))__declspec(deprecated(X))The first two ones are the same, except that only the second one can be used inside a macro; I've put the first one only for completeness. The third one seems to be the most used in the dev community.
I wonder what is the difference between the last two ones. According to the documentation on MSDN here and here, I understand that there is no difference. What is strange in that case is that a different warning code is raised depending on what you're using: C4995 for the pragma-case, C4996 for the declspec-case.
So does somebody knows if there is actually a difference (any tiny one), or why these directives don't raise the same warning code?
See deprecated (C++):
(Microsoft specific) With the exceptions noted below, the deprecated declaration offers the same functionality as the deprecated pragma:
- The deprecated declaration lets you specify particular forms of function overloads as deprecated, whereas the pragma form applies to all overloaded forms of a function name.
- The deprecated declaration lets you specify a message that will display at compile time. The text of the message can be from a macro.
- Macros can only be marked as deprecated with the deprecated pragma.
For #pragma vs. __pragma, see Pragma Directives and the __Pragma Keyword:
The __pragma() Keyword
Microsoft specific
The compiler also supports the __pragma keyword, which has the same functionality as the #pragma directive, but can be used inline in a macro definition.
It makes sense to note, as @Deduplicator mentioned, that C++14 introduces the [[deprecated]] attribute.
7.6.5 Deprecated attribute [dcl.attr.deprecated]
The attribute-token
deprecatedcan be used to mark names and entities whose use is still allowed, but is discouraged for some reason. [ Note: in particular,deprecatedis appropriate for names and entities that are deemed obsolescent or unsafe. —end note ]