This was previously r=jchen and landed, but was backed out because some
Android tests were running on an unrealistically old Linux kernel. These
tests have been retired, so this can reland.
Differential Revision: https://phabricator.services.mozilla.com/D81278
I will admit from the start that this patch is goofy and is not the "proper" fix.
If you're unlucky enough to tickle the current unification in `dom/media`, then there's a conflict between `namespace mozilla::details` of `SPSCQueue.h` and `namespace details` of Chromium's `task.h`. Ultimately, the badness stems from some unified file doing `using namespace mozilla;`, but I'm really not enthusiastic about reworking that when I just want to get unblocked.
I'm going to attempt to justify the yuckiness of this patch on the basis that:
* `SPSCQueue.h` is super self-contained and rarely looked at, I bet nobody will notice this change
* Tons of other code prefers the spelling `mozilla::detail` anyway
Differential Revision: https://phabricator.services.mozilla.com/D81162
The following simplifications are made:
* Unnecessary function template arguments are removed.
* Unnecessary copy constructor definitions are removed (making the types
movable where possible).
* Iterators are moved where possible rather than copied.
* Unnecessary MOZ_IMPLICIT on a constructor with two arguments is removed.
Differential Revision: https://phabricator.services.mozilla.com/D80015
The following simplifications are made:
* Unnecessary function template arguments are removed.
* Unnecessary copy constructor definitions are removed (making the types
movable where possible).
* Iterators are moved where possible rather than copied.
* Unnecessary MOZ_IMPLICIT on a constructor with two arguments is removed.
Differential Revision: https://phabricator.services.mozilla.com/D80015
mozilla::ReverseIterator is required for "stashing iterators", but otherwise
std::reverse_iterator can be used and is potentially more efficient.
Also fix some other glitches in iterator type definitions.
Differential Revision: https://phabricator.services.mozilla.com/D79489
Previously, the tests assumed that the alignment of int and long equals
their size. This commit fixes the tests for targets like m68k that have
sizeof(int) == 4 and alignof(int) == 2. A static helper function sizemax
was introduced as the offset of the second element in Pair<int,long>
might be either determined by its alignment requirement or the size of
the preceding int element and we use the helper function to pick the
larger of the two values.
Differential Revision: https://phabricator.services.mozilla.com/D77289
Currently, MOZ_ALIGNED_DECL uses the order (_type, _align) for its
parameters. However, this order makes the code less readable when
_type is a larger object like a struct because the value for _align
would be at the end of the struct definition. By swapping the order
of _type and _align, the alignment value will always be next to
the type name, regardless how far the definition of _type extends.
Differential Revision: https://phabricator.services.mozilla.com/D77288
Previously, the tests assumed that the alignment of int and long equals
their size. This commit fixes the tests for targets like m68k that have
sizeof(int) == 4 and alignof(int) == 2. A static helper function sizemax
was introduced as the offset of the second element in Pair<int,long>
might be either determined by its alignment requirement or the size of
the preceding int element and we use the helper function to pick the
larger of the two values.
Depends on D77288
Differential Revision: https://phabricator.services.mozilla.com/D77289
Currently, MOZ_ALIGNED_DECL uses the order (_type, _align) for its
parameters. However, this order makes the code less readable when
_type is a larger object like a struct because the value for _align
would be at the end of the struct definition. By swapping the order
of _type and _align, the alignment value will always be next to
the type name, regardless how far the definition of _type extends.
Depends on D77287
Differential Revision: https://phabricator.services.mozilla.com/D77288
There's no need to provide the compiler a bunch of stuff that it's
trivially going to optimize away anyway, and we'd rather the release
assert not mess with the optimizer's view of the code.
Differential Revision: https://phabricator.services.mozilla.com/D77124
The compiler is not always able to remove these (always-successful)
checks, so we might as well help it along.
Depends on D77122
Differential Revision: https://phabricator.services.mozilla.com/D77123
We're going to introduce some friend declarations in the next patch, and
this change makes doing so easier. It's also slightly less confusing.
Depends on D77121
Differential Revision: https://phabricator.services.mozilla.com/D77122
This also uses MakeUniqueForOverwrite* in two places where we immediately copy
over the Buffer from a Span.
Adds move assignment operator as well.
Differential Revision: https://phabricator.services.mozilla.com/D75510
These are the equivalent of std::make_unique_for_overwrite and std::make_unique_for_overwrite with fallible allocation.
Differential Revision: https://phabricator.services.mozilla.com/D75509
The phrase "copy the error" is misleading, as the error type might not be
copyable, and it is never copied but moved. "Propagate" the error seems to
be a good term to describe this.
Differential Revision: https://phabricator.services.mozilla.com/D75093
