NULL vs. 0

[Peter Kovacs]

As far as know the sugested way to define null pointers is using 0 instead of NULL.

In the source codes there are only a few usage of NULL, [445b86e6bb9f] replaces them with 0.

[Alpar Juttner]

Replying to kpeter:

As far as know the sugested way to define null pointers is using 0 instead of NULL.

IMHO this is a very strange "guideline" in C++. The implicit int -> pointer conversion always "smells" like hacking to me, thus I personally prefer using NULL.

Actually NULL will never be really deprecated, as keeping the backward compatibility with C is a fundamental design rule of C++.

[Peter Kovacs]

Replying to alpar:

Replying to kpeter:

As far as know the sugested way to define null pointers is using 0 instead of NULL.

IMHO this is a very strange "guideline" in C++.

Okay, what should we do then? Do we want to have a convention about this question or both form will be accepted? (E.g. in bfs/dfs/dijkstra in most cases 0 is used, but in only one constructor NULL is used.)

[Alpar Juttner]

Replying to kpeter:

Okay, what should we do then? Do we want to have a convention about this question or both form will be accepted? (E.g. in bfs/dfs/dijkstra in most cases 0 is used, but in only one constructor NULL is used.)

We all know that NULL and 0 is equivalent in this context. Does it give us any benefit to introduce a another convention and force people to follow it? I don't think so.

[Peter Kovacs]

So we should close this ticket with "won't fix" status, shouldn't we?