Just curious, why do you define end as opposed tho this:
for (auto it = m_buttons.constBegin(); it != m_buttons.constEnd(); it++) {

}

Another route you can go which looks much cleaner imho (requires Qt 5.7 because of qAsConst):
for (auto entry, qAsConst(m_buttons)) {

}

Just my 2 cents.

Just curious, why do you define end as opposed tho this:

Because @broulik tends to point out that it is not cached.

Another route you can go which looks much cleaner imho (requires Qt 5.7 because of qAsConst):

does that work in a sensible way for a QHash? The most lean way would have been:

if (std::any_of(m_buttons.constBegin(), m_buttons.constEnd(), [] ...))

But that doesn't work with QHash. So I kind of doubt QHash and qAsConst do something sensible.

This is where STL and Qt apparently diverge a bit.

for (auto value : qAsConst(m_buttons)) {
// ...
}

Gives you the values in the map. Not the keys. And that is because it's a QHash container which apparently behaves like that.
If it were an std::unordered_map it would have given you a std::pair where first would be the key, second would be the value.

However, in *this* case you are only using the value thus you are fine when using:
for (auto value : qAsConst(m_buttons)) {
// ...
}

Since you seem to mention optimization (you shouldn't have done that ;)), this is the most efficient version, and i looked it up [1], hehe:
for (cont auto &value : qAsConst(m_buttons)) {
// ...
}

In fact. You likely always want this version of the range-based-for loop. And "for (auto&& value: ...)" if you want to modify them in place. You nearly never want a plain range-based-for without a reference symbol in there because that makes a copy.

Hope that helps :)

[1] Read https://blog.petrzemek.net/2016/08/17/auto-type-deduction-in-range-based-for-loops/ for the details