1139993 - Add VectorBase::ConstRange that only gives out `const T&` entries

Status: RESOLVED FIXED

(Core :: MFBT, defect)

Description

[Nick Fitzgerald [:fitzgen] [⏰PST; UTC-8]]

VectorBase::Range gives out `T&`, but for cases where consumers shouldn't be able to modify entries, only read them, it doesn't provide any nice guarantees. We should have a VectorBase::ConstRange that only gives out `const T&` entries.

Is there a better (standard-ish for this type of thing?) name than ConstRange?

Comment 1

[Nick Fitzgerald [:fitzgen] [⏰PST; UTC-8]]

Attachment: Add VectorBase::ConstRange that gives out |const T&| entries

Comment 2

[Jeff Walden [:Waldo]]

Comment on attachment 8573439 [details] [diff] [review]
Add VectorBase::ConstRange that gives out |const T&| entries

Review of attachment 8573439 [details] [diff] [review]:
-----------------------------------------------------------------

::: mfbt/Vector.h
@@ +492,5 @@
> +    {
> +      MOZ_ASSERT(aCur <= aEnd);
> +    }
> +
> +  public:

Either include ConstRange() {}, or (more likely) remove Range() {} from the class just above.

@@ +496,5 @@
> +  public:
> +    bool empty() const { return mCur == mEnd; }
> +    size_t remain() const { return PointerRangeSize(mCur, mEnd); }
> +    const T& front() const { MOZ_ASSERT(!empty()); return *mCur; }
> +    void popFront() { MOZ_ASSERT(!empty()); ++mCur; }

Why no popCopyFront?  Seems like it should be fine to include, it's not mutating any members, just making copies of them.

Comment 3

[Nick Fitzgerald [:fitzgen] [⏰PST; UTC-8]]

Attachment: Add VectorBase::ConstRange that gives out |const T&| entries

Comment 4

[Nick Fitzgerald [:fitzgen] [⏰PST; UTC-8]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=f18fa193e90f

Comment 5

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/9da55484cc2d

Comment 6

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/9da55484cc2d