A small round of fixes #9676
A small round of fixes #9676
Conversation
Because of floating point computery math stuff. Regression since koreader#9609 c.f., koreader#9609 (comment)
Which allows us to exit the screensaver on, well, any key press ;o).
And simplify the few we do catch by using aliases instead of duplicated functions;).
|
(Rebase me) |
| for i = 1,5 do | ||
| self:setIntensityHW(math.ceil(self.fl_min + ((self.fl_intensity * (1/5)) * i))) | ||
| for i = 1, 5 do | ||
| self:setIntensityHW(math.ceil(self.fl_min + (self.fl_intensity / 5 * i))) |
There was a problem hiding this comment.
| self:setIntensityHW(math.ceil(self.fl_min + (self.fl_intensity / 5 * i))) | |
| self:setIntensityHW(math.ceil(self.fl_min + (self.fl_intensity / 5 * i))) -- @zwim don't change me :-P |
There was a problem hiding this comment.
I dont know if it is important, but after applying these changes, Toggling frontlight on Kobo Libra 2 works as it should!
Thank You very much! :)
Actually... almost ;o). |
If you've ever enabled the main loop debugging, you'll know that actually dumping the full window stack was *hilarious*. Just print table counts, it's often good enough to debug what's happening in the exceedingly rare cases you need this ;). Also, it'll actually be readable, unlike the previous insanity ^^.
|
Does the frontlight issue fix deserves a v2022.10.1 (not affected, can't judge criticity) ? |
frontend/ui/time.lua
Outdated
| @@ -171,7 +171,7 @@ end | |||
| function time.split_s_us(time_fts) | |||
| if not time_fts then return nil, nil end | |||
| local sec = math.floor(time_fts * FTS2S) | |||
| local usec = math.floor(time_fts - sec * S2FTS) * FTS2US | |||
| local usec = math.floor((time_fts - (sec * S2FTS)) * FTS2US) | |||
There was a problem hiding this comment.
Good catch. The original line is buggy. Your version works. The innermost parentheses are superfluous. So maybe write
local usec = math.floor((time_fts - sec * S2FTS) * FTS2US)
The good news is, that this bug is not noticeable by now, as PRECISION can be stored in a double without precision loss; US2FTS is exactly 1 and FTS2US is exactly 1, too. But for future (if someone wants to tune the thing by using a different PRECISION value (e.g. a power of 2), this buggy line would cause an issue).
That's a good question. It breaks toggle off on a not insignificant number of frontlight values, which is mildly annoying, but can be worked-around by just changing the fl value by +/- 1, so, eeeeh. Might be annoying to field those issues for a month, though ;). |
The us part wasn't actually truncated properly.
|
OK, so it does not prevent anybody from using the software. We mostly did some rerelease when it crashed at start I guess. With v2022.10, they can upgrade to dev channel and get it fixed. Or live with it, they deserve it :) for not helping us fixing stuff by being on the dev channel :) We can remove "Chinese, Japanese and Korean typography: some questions" from the pinned issue if we want a slot to pin the frontlight issue until v2022.11, and may be avoid some duplicates opening. |
There was a problem hiding this comment.
Reviewed 14 of 15 files at r1, 3 of 4 files at r2, 1 of 1 files at r4, 1 of 1 files at r5, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @Frenzie, @pazos, @poire-z, and @zwim)
onAnyKeyPressedhandler (allowing one to close the ScreenSaver with page-turn buttons, for instance). [This is properly disabled when using the child lock pattern mode, because @poire-z rules ;)].This change is