Not doing so will eventually lead to issues with prebuilt contribs when
the pipeline doesn't run in the videolan/vlc namespace

Alexandre Janniaux authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Comments in Makefile recipees are shell comments and not Makefile
comments. Because of that, the #~ is printed when executing the
.sum-dav1d recipee.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

IID_PPV_ARGS would be a lot cleaner though

Some uuidof are not always present in various mingw-w64 so we use their IID directly.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Now FindCaptureDevice always returns p_base_filter.

Fix some dirty casts.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Pass some ComPtr as reference to avoid using .Get() and losing the ComPtr helpers.

Factorize some code to get the tv tuner interface.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

And use ComPtr::As() rather than interface::QueryInterface() to avoid using
GetAddressOf.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

And use ComPtr::As() rather than interface::QueryInterface() to avoid using
GetAddressOf.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Either from direct pointer from the DLL, or via GetProcAddress.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

And use ComPtr::As() rather than interface::QueryInterface() to avoid using
GetAddressOf.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

Steve Lhomme authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's better not to use it when reference counting is not involved. Otherwise
operator& and ReleaseAndGetAddressOf() are better.

And use ComPtr::As() rather than interface::QueryInterface() to avoid using
GetAddressOf.

This is the version that was merged upstream so it makes sense to use it
to facilitate future rebases
upstream PR: https://github.com/taglib/taglib/pull/1040

Rémi Denis-Courmont authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Linux, FreeBSD and OpenBSD have proper checks and do not use this code.
macOS also had tests while it was supported, but it no longer is.

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

The previous behaviour upon filter change was that:
 - If the search string was not empty then everything visible was expanded.
 - If empty, then everything would be collapsed except the selected node.

This is rather annoying IMO, for instance - from a default starting state
where top-level nodes are expanded and 2nd-level collapsed, if you apply a
filter and then subsequently remove it, you annoyingly end up with more
collapsed than you started with.

I feel that it works much better to just leave the expanded/collapsed state
untouched, preserving the initial state and any changes the user has made
before/after filtering, even though this forces users to expand any
2nd-level filter results they are interested in for themselves. It's also
much more simple.

Lyndon Brown authored 6 years ago and Rémi Denis-Courmont committed 3 years ago

the old option was an int type used like a bool. the new options is an
actual bool, simplifying things, including the UX in the GUI and CLI.

Let the compiler pick whichever register it sees fit.

Output the current CPU flags directly into the selected register.
Thus EBX is left alone, and need not be saved/restored.

We do not care about any side effects here, only the output operands.

This gets the EBX value directly from, well, EBX. The compiler knows
that EBX will be clobbered given it is an output operand, so it will
save and restore it if/as needed.

Detection only worked on Linux, but more importantly, this flag fell out of
use with 8512cab6.

This flag fell out of use with 8512cab6.

Detection only worked on Linux, but more importantly, this flag fell out of
use with 8512cab6.

Detection only worked on Linux, but more importantly, this flag fell out of
use with 8512cab6.

Same problem as 0c482e07.

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Unused since b86d55e0.

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Note that in some cases they have been changed to forward declarations, and
in some they have been moved, for instance inclusion of `<QUrl>` was moved
from `qt.hpp` to where it was needed.

Lyndon Brown authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Any tiny build speedup from being able to use forward declarations is
more preferable IMO to having these small functions within the header.