Please wait for a second review, I am not authoritative here.

In D27965#625234, @meven wrote:

Please wait for a second review, I am not authoritative here.

OK, thanks.

Looks good to me, I wonder why you turned some to for+iterators.

In D27965#625381, @apol wrote:

Looks good to me, I wonder why you turned some to for+iterators.

Those modified the list as they iterated through the list : authList->removeOne(current);

Having the iterated value change under the hood will eventually break. I'd suggest preferring QList::erase to QList::removeOne.

In D27965#625526, @apol wrote:

Having the iterated value change under the hood will eventually break. I'd suggest preferring QList::erase to QList::removeOne.

Actually, the code also uses 'delete current', now if there's another instance of that AuthInfoContainer* in authList, on a next iteration accessing the pointer "current" would be invalid as the object it had pointed to has been already delete'ed... I think removeAll() would be a better fit; WDYT?

In D27965#625676, @ahmadsamir wrote:

In D27965#625526, @apol wrote:

Having the iterated value change under the hood will eventually break. I'd suggest preferring QList::erase to QList::removeOne.

Actually, the code also uses 'delete current', now if there's another instance of that AuthInfoContainer* in authList, on a next iteration accessing the pointer "current" would be invalid as the object it had pointed to has been already delete'ed... I think removeAll() would be a better fit; WDYT?

authList does not contain duplicates. Otherwise the current code would already have the problem you're mentioning.
So I see no reason to protect against that case.
erase sounds much safer than removeOne which has unclear consequences on the iterators.