Enable range based looping of sf::VertexArray

[ChrisThrasher]

Description

begin and end iterators are all that is required of a container to supported C++11 ranged-for loops. This PR also uses this feature as much as possible internally. Let me know if I missed any more places we can be using this.

[coveralls]

Pull Request Test Coverage Report for Build 12536923590

Details

• 25 of 27 (92.59%) changed or added relevant lines in 3 files are covered.
• 4 unchanged lines in 1 file lost coverage.
• Overall coverage decreased (-0.004%) to 57.02%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/SFML/Graphics/Text.cpp	2	4	50.0%

Files with Coverage Reduction	New Missed Lines	%
src/SFML/Window/WindowImpl.cpp	4	52.29%

Totals
Change from base Build 12509336161:	-0.004%
Covered Lines:	12068
Relevant Lines:	20002

---

💛 - Coveralls

[kimci86]

VertexArray::operator[] is not covered by tests now.

[ChrisThrasher]

VertexArray::operator[] is not covered by tests now.

This operator is still used indirectly by some other tests.

[kimci86]

This operator is still used indirectly by some other tests.

Which one? Coveralls report does not see it.

[ChrisThrasher]

This operator is still used indirectly by some other tests.

Which one? Coveralls report does not see it.

SFML/test/Graphics/VertexArray.test.cpp

Line 78 in 3c084bf

CHECK(vertexArray[0].position == vertex.position);

It’s used in a few places. It’s weird that coverage doesn’t catch this.

[kimci86]

Ok, this is because the non-const variant is used by tests but the const variant is not.

[ChrisThrasher]

Ok, this is because the non-const variant is used by tests but the const variant is not.

Ah okay this makes sense. I added tests explicitly for this case to make sure it remains covered. Thanks for catching this!

[eXpl0it3r]

Am I understanding it correctly, that this opens up additional opens to manipulate the VertexArray's vertices? Anything that takes a begin()-end() "range", right? Is there anything we need to consider in those cases?

What I mean, the PR focuses on ranged-based looping, but the feature itself goes beyond that. So I just want to make sure when people do manipulate the vertices through that interface, nothing breaks. I currently don't see anything, but maybe someone else can think of some potential problem.

[ZXShady]

Am I understanding it correctly, that this opens up additional opens to manipulate the VertexArray's vertices? Anything that takes a begin()-end() "range", right? Is there anything we need to consider in those cases?

What I mean, the PR focuses on ranged-based looping, but the feature itself goes beyond that. So I just want to make sure when people do manipulate the vertices through that interface, nothing breaks. I currently don't see anything, but maybe someone else can think of some potential problem.

they can't break anything

begin and end are just equalivent semanticly to &vertex_array[0] and &vertex_array[vertex_array.size()] respectivly but less annoying to type.

[ChrisThrasher]

As far as I’m aware, this is purely syntactic sugar. It doesn’t enable any functionality that wasn’t previously possible through an existing interface.

[eXpl0it3r]

Seems like a useful addition 👍