[cfe-dev] null pointer literals, warnings, and fixups

[David Blaikie]

>
> Doesn't that make a perfect case for an off-by-default warning, then?
> Rather than users having to employ a separate tool (though, I admit, if a
> clang plugin is easy/lightweight enough I'm not at all averse to doing this
> sort of thing as a plugin. I'll certainly follow up on that) to ensure they
> don't make such mistakes.
>
>
> I have already expressed my reasons for not wanting to introduce yet more
> off-by-default warnings.
>

Right - my point was this seems like a counterargument to that, showing a
good use case for an off-by-default warning.


> The point I'm making is that it's more than style - it has practical
> benefits that the compiler can understand the developers intent & provide
> diagnostics. How does the user know they're missing out on those useful
> diagnostics? Even if they know the diagnostics exist, it'll catch all their
> spurious NULL usage, but not the other way around (0 as null pointer) - that
> will be silently accepted without the use of an external tool.
>
>
> Every purveyor of a convention claims that it's more than style, and it has
> practical benefits if only developers would follow that convention.
>

Yet we already agree that there's practical benefits to this - that we only
get half of. Anyone using 0 everywhere gets nothing, but if people err on
the side of NULL they get something but without much of a guarantee that
they haven't accidentally slipped in a 0 somewhere. 'spose that's not so
different from bracing, though - "hey, brace every block & you'll be less
likely to get bugs over dangling elses, but the compiler won't check that
you haven't accidentally snuck in an unbraced block".


> That's not relevant here: the convention is not widespread enough to codify
> it in a compiler.
>

That was my point about -Wparentheses. In that case the only way it got
codified was by its inclusion in the compiler. I doubt many people had that
as an existing practice prior to that.


> Hence the use of a style checker, which presumably will get run often.
>

Yep. I'll certainly have a go at that.


> Yes, these are among the reasons why -Wparentheses makes sense for Clang
> but complaining about 0 in a pointer context does not. In any case, my
> decision stands: we will not accept a warning about '0' being used in a
> pointer context at this time.
>

Noted.

I'll have a go at the "don't use NULL in non-pointer contexts" fix at some
point, at least.


> IIRC, it does have Fix-Its on the notes.
>

Oh, so it does - not sure how I missed that. But that does answer one of my
previous questions - it is possible to have multiple fixits variations
presented to the user, they just have to be presented in notes, not both in
the same diagnostic message itself. Good to know.

- David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20110826/792dbac0/attachment.html>