Conversation
| # for one channel default to a white channel | ||
| color_channels = "w" | ||
| elif len(channel_list) == 3: | ||
| # TODO this seems like a bug waiting to happen -- what if the channel list ISNT R G B ordered? |
There was a problem hiding this comment.
I think if the channel list isn't RGB ordered then the user should specify a type to clarify what the channels are. It's a fair assumption that if the user doesn't specify, we have some type of fallback.
Alternatively, we could make type required with a default value of "rgb"
There was a problem hiding this comment.
This case in particular only happens if you dont have a config[channels] or a config[type], so I guess really the user is letting the platform defaults do their thing. Since channel_list is always from a simple number to channel set generation via parse_light_number_to_channels, it's not really a userland bug that could happen here, it's a platform not defining a set of three channels as exactly r g then b.
So maybe the comment is too harsh, and all we need is documentation on the parse_light_number_to_channels interface that it must return an rgb-ordered channel list
bosh
force-pushed
the
lights_and_exp_cleanup
branch
from
649456c to
98efd4f
Compare
bosh
force-pushed
the
lights_and_exp_cleanup
branch
from
98efd4f to
1be2006
Compare
bosh
force-pushed
the
lights_and_exp_cleanup
branch
from
1be2006 to
c21e745
Compare
…h > 1 carry value though this should never happen in practice since we only call loop range(3) times, so +0, +1, and +2. Even a 4-channel light could only carry once, so really this should only matter in the 5+ channel case (or perhaps if you can start with an offset of +3, which really should be a carry+1 with 0 offset
for readability. The config params are generally pulled out inside the initialize in these, rather than from the self config reference within the function -- this is intentional to try and keep config reading inside initialize itself
…rought execution order for readability
to better orient the seams with the data structures and relevancy
bosh
force-pushed
the
lights_and_exp_cleanup
branch
from
c21e745 to
dd67090
Compare
|
2 participants
Nonbreaking refactors and cleanup in the lighting code in prep for FAST EXP ER support