Done with ad hoc scripting hacks processing unused imports found by pyflakes:
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Logs.* but unused' | cut -d: -f1 | while read f; do sed -i 's/^import waflib.Logs as Logs,/import/g' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Options.* but unused' | cut -d: -f1 | while read f; do sed -i 's/import waflib.Options as Options, /import /g' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Options.* but unused' | cut -d: -f1 | while read f; do sed -i 's/^from waflib import Options,/from waflib import/g' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep ' imported but unused$' | sed "s/^\([^:]*\):[0-9]*:[0-9]* '\(.*\)'.*/\1 \2/g" | while read f lib; do sed -i "/^import $lib$/d" $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Options.* but unused' | cut -d: -f1 | while read f; do sed -i '/from waflib import Options$/d' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.TaskGen.* but unused' | cut -d: -f1 | while read f; do sed -i '/from waflib import TaskGen$/d' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Task.Task.* but unused' | cut -d: -f1 | while read f; do sed -i '/^from waflib.Task import Task$/d' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Tools.winres.* but unused' | cut -d: -f1 | while read f; do sed -i '/^from waflib.Tools import winres$/d' $f; done
for f in $( find * -name wscript ); do echo; pyflakes $f; done | grep 'waflib.Utils.* but unused' | cut -d: -f1 | while read f; do sed -i '/^import waflib.Utils as Utils$/d' $f; done
As was noted in 88ee3af3ea it is unsafe/undefined behavior if two threads
sleep on the JACK request file descriptor, since there is no way to control
which one will wake and process the request. Since each thread may have
sent a different request, this can lead to a thread misinterpreting the
response because it is reading the wrong response.
This may (or may not) solve some subtle problems with JACK, but was
revealed by having a control surface (LaunchPad Pro) that registers
three ports from the butler thread at about the same as the GUI
thread is registering the auditioner. One thread read the wrong
response, and because of some slightly weird code/design, it attempts
to rename the port from within the response handler, which in JACK1
leads to deadlock (and later, zombification).
Without this, two threads can both sleep on the same communication channel, and the wake order
is non-determinate, so the wrong thread may process the response to the other thread's request.
See also 976e03c15c which does this for `route_list_to_control_list`.
Fix crashes with empty route-lists e.g. momentary solo which was
introduced in 03105aa760.
this defines how the grid interacts with other snap targets (ph, etc)
* Grid: we ignore other snap targets when the grid is enabled
* Both: we snap to both grid and other snap-targets when grid is enabled
* Other: we only snap to other snap-targets and ignore the grid, even
though the grid is enabled
Previously,
0 -> no swing (1:1, 50%)
50 -> triplet swing (2:1, 66%)
75 -> hard swing (3:1, 75%)
100 -> sextuplet swing (5:1, 83%) (default!)
150 -> absolute maximum (inf:1, 100%)
This is rather confusing...
One common interpretation uses percentages of the beat, where triplet
swing is 66%. However, that causes precision issues since it's really
66.666...
Since we already default to 100 and take "no swing" as zero, let's make
that reference point triplet swing. Then the scale becomes:
0 -> no swing (1:1)
100 -> triplet swing (2:1)
150 -> hard swing (3:1)
200 -> sextuplet swing (5:1)
300 -> absolute maximum (inf:1)
300 doesn't make any sense, so let's change the range to -250 .. 250
which covers all useful values.
Also remove the division through 100 and back, to avoid rounding issues.
Signed-off-by: Asahi Lina <lina@asahilina.net>
- Fix selection of what beats to swing (was always done)
- Fix swing strength (lack of precision rounded to 50% or 100%)
- Fix model offset not being applied properly
Signed-off-by: Asahi Lina <lina@asahilina.net>
When drawing the outline of a rectangle, the bounding
box must cover the whole pixel of the line. Otherwise
the line would be left behind when the rectangle shrinks.
The read_index is adjusted in the loop, which means that the calculation of how much
data can be delivered to the stretcher must also be inside the loop
