The macros
HAVE_PTHREAD_MUTEX_TIMEDLOCK
HAVE_PTHREAD_COND_TIMEDWAIT
are still being used in the Pthreads threading implementation, but the
CMake build system never sets them, so the timed versions aren't used.
This ended up making WorkerThread heartbeats not beat. </3
IIDX hides inactive sections, but unlike the current behavior of our
OnlyShowActiveSection metric (based on certain Pumps according to
DaisukeMaster's comment), it retains the section title so that the
section can be closed normally.
New metric: MusicWheel.HideActiveSectionTitle, defaulting to true in
_fallback. A bit unfortunate, but keeps backward compatibility with the
existing behavior of OnlyShowActiveSection.
If neither metric is true, wheel displays all titles plus the contents
of the active section (if any).
If OSAS is true but HAST is false, wheel displays the active section
title plus its contents, or all titles if no section is active.
If both are true, wheel displays only the active section's contents, or
all titles if no section is active.
Recommended by Travis docs:
http://docs.travis-ci.com/user/installing-dependencies/
We may eventually want to look into whether the before_install/install
hooks are supposed to be used for this rather than before_script.
Moves all of the wheel-entry-hiding logic into SetOpenSection rather
than prematurely throwing out section headings. Now when we close the
active section with OnlyShowActiveSection in effect, the headings are
shown rather than an empty wheel.
The mappings in this driver are specific to the piuio kernel driver, but
once lights mapping is possible this could be made into a generic leds
class driver.
The mappings in this driver are specific to the piuio kernel driver, but
once lights mapping is possible this could be made into a generic leds
class driver.
This reverts commit 54933034e8.
We might have to figure out a more selective way to do this. For now,
let's not have people on all the other OSes getting stuck-modified files
like hg did.
This function, oddly, used to take a 1 or 2 for the player number where
other nearby functions (e.g. IsPlayerEnabled) would take the proper enum
value, 0 or 1. If the old-style syntax (raw number) is used, also use
the old-style interpretation.
The XML patchset includes more items in the FGCHANGES line, which is
stored in cache. Previous caches wouldn't include these entries.
Sorry 'bout that folks.
For backwards compatibility, broadcast the old StepP1/2 message in
addition to the new Step message. New themes, charts, etc. should of
course use the new message with parameters.
We don't need the ending beat until the change has already started, so
just use the legitimate On to calculate things rather than doing
anything hacky.
We load foreground changes at the beginning of the screen, but they
shouldn't be issued an OnCommand until it's time for them to start. The
Foreground layer, however, should get the OnCommand right away. So
don't propagate it, since we already issue On manually to changes.
The m_iSelection array could probably be made unsigned... IF we were
confident we never try to assign it a negative number, of which fact I'm
not convinced.
Themes can use these to implement custom per-profile data, such as
SpeedMods.txt. For memory cards, they will be executed while the card is
mounted and should read/write everything they will need at that point.
The current placement of the call to GetLuaInformation misses a bunch of our
"singletons." Moving it causes problems, but we can at least include them in
the documentation as if it worked properly. :)
The sort was in place for this purpose. Let's use it on the modifier keys, but
not sort the regular keys, since the order matters for those.
This isn't really the right way to do this from the start, but that'll be
something that could be rewritten at another time given that our mappings are
still pretty much hardcoded.
When there were tap notes on the same row as the start of a roll, autoplaying
the roll would result in the receptors on other tracks to glow. This was
because the Step() call would use the roll's head row for a parameter, which
included the other notes. This changes that parameter to -1 to automatically
use the current row in the song, which makes more sense anyway.
Fixes bug 241.
Devin J. Pohly