freenode/#shirakumo - IRC Chatlog
Search
9:31:26
phoe
Shinmera: https://github.com/phoe-trash/qtools-ui/commit/69119c3392e21ad8f3cf5623bd91e020dde35b9b
9:31:27
Colleen
github.com/phoe-trash/qtool... Website (HTML), Title: Use DOCUMENTATION-UTILS · phoe-trash/qtools-ui@69119c3 · GitHub
9:38:51
Shinmera
Typically I prefer having every documentation form with the qualified type (so function, where missing)
9:50:18
Colleen
github.com/phoe-trash/qtool... Website (HTML), Title: Add explicit cl:function in define-docs · phoe-trash/qtools-ui@dbc1529 · GitHub
9:54:14
Shinmera
Wrap them at a consistent length, whether that's 80 or 100 I don't particularly care
9:55:09
Shinmera
Alternatively I've been thinking about making Staple wrap things in pres too, so that I can just write text blocks without having to manually wrap at all
9:57:27
Shinmera
the format is always "summary line that can exceed wrapping length\n\nlong description with multiple paragraphs\n\nSee xrefs, one line at a time
10:01:53
phoe
https://cdn.discordapp.com/attachments/297478485000192010/545545184902119434/Screenshot_from_2019-02-14_11-01-00.png
10:02:50
phoe
there is no newline between (type rgb-color-slider ...) and (type hsv-color-slider ...)
10:04:31
Shinmera
If you want you could depend on multilang-documentation-utils and also translate the docstrings into polish :v
10:17:55
Shinmera
because this resampling business, and other potential plugins as well, require me to allow changing the sample count on the fly
10:18:35
Shinmera
mostly since the graph is a DAG, I can't just propagate the sample count as a maximum downwards in the graph
10:19:57
Shinmera
problem is though, no matter what, currently libmixed lacks the infrastructure for this dependence information. You build the graph implicitly through shared buffers
10:21:21
Shinmera
also, the resampling only knows the updated sample count /after/ performing the resampling
10:21:50
Shinmera
the alternate would be to complicate buffer logic a lot and turn them into ring buffers
10:22:49
Shinmera
problem with that approach, aside from complicating buffer access, is that now the buffers can never be shared
10:23:47
Shinmera
or at least I don't think they could be for any segments that have varying sample counts
10:25:32
Shinmera
wait, no, they could never be shared, since now even segments that don't change sample count themselves effectively become ones that do, due to the change becoming infectious
10:55:49
phoe
one more thing, https://cdn.discordapp.com/attachments/297478485000192010/545558738619203604/Screenshot_from_2019-02-14_11-03-29.png