We use pre-multiplied alpha color channels, but were having bad
rounding errors due to the alpha divider being truncated to an
integer.
The algorithm for pre-multiplying a color channel is:
alpha_divider = 0xffff / alpha
pre_mult_color = color / alpha_divider
In order to fix the rounding errors, we could turn ‘alpha_divider’
into a double.
That however would introduce a performance penalty since now we’d need
to do floating point math for each cell.
The algorithm can be trivially converted to:
pre_mult_color = color * alpha / 0xffff
Since both color and alpa values are < 65536, the multiplication is
“safe”; it will not overflow an uint32_t.
If the user did *not* configured a specific monitor, we prefer the
primary monitor. However, if that monitor is disconnected, yambar
would exit with:
no matching monitor
This patch changes this, to use the *last* connected monitor. It also
improves the error message.
Note: if the user did specify a monitor in the configuration, but
perhaps misspelled it, we will *not* fallback to another monitor, but
instead log an error saying that specific monitor could not be found.
Assume that a closed pipe means the child died. Even if it hasn’t, we
can’t read anymore from it. We’ll end up killing it anyway before
returning from run().
This is mainly to fix a race when river is *not* running; sometimes we
ended up allocating particles for N seats in content(), but then when
iterating the seats, run() had destroyed all, or some of the seats,
causing us to feed NULL pointers to dynlist, which crashed.
* Verify int/float/bool values are that, and nothing else
* Verify tag ranges are integers
* Verify a range tag value is inside its range
* Don’t allow anything but false|true for booleans
User is expected to send ‘false’ or ‘true’. But we were parsing the
value using `strtol()`. This caused all bools to be false, since
`strtol()` would always return 0.
