tags:

views:

119

answers:

2

Is the v-table (virtual method table) a part of the C++ specification, or is it up to the compiler to solve the virtual method lookups?

In case it's part of the spec: Why?

(I'd guess that it's compiler dependent, but someone said to me that it's part of the spec.)

(References are very welcome!)

+4  A: 

No, it's not part of the language specification. The standard specifies how calls to virtual functions must be resolved but not the mechanism that compiler implements to achieve the required results.

It's difficult to provide a "negative" reference (i.e. to where the standard doesn't mandate a v-table) but the relevant parts of the standard for virtual function calls are 5.2.2 [expr.call] and 10.3 [class.virtual].

Charles Bailey
+11  A: 

1.7 The C++ memory model *3* [...] Various features of the language, such as references and virtual functions, might involve additional memory locations that are not accessible to programs but are managed by the implementation. [...]

There you have it. It is up to the implementation.

dirkgently
You have a URL for that quote? ([This one](http://www.kuzbass.ru:8086/docs/isocpp/) doesn't say the same thing).
aioobe
No, but you can download n3092.pdf. I am not sure which version of the draft/which Standard that link refers to. (It doesn't even open for me!)
dirkgently
Worth emphasizing that n3092.pdf is the draft for the _next_ version of C++ (C++0x) and not a draft of the current standard.
Charles Bailey