A bunch of fixes#9902
Conversation
|
(Draft because I still have others things planned for tonight, but wanted to braindump a few things first) |
|
I Just debugged this, (I had suspected that commit) and came to comment when i saw this :) My fix was just to add a line: self._exit_code = nilto the beginning of |
|
it is the |
|
Yup! I'd rather be overly explicit to document this and avoid a future regression if we lost this to the sands of time, though ;). (And also to actually skip the :quit() call, because it might not always stay as mostly harmless as it is today). |
d650e3b to
9b72f37
Compare
Confirmed when testing the Kindle stuff on a slow PW1: it works just fine. |
|
Kindle t1 timeout trickery seems to behave, yay. |
Well, does that mean it won't complain when we have unused: But mhhh, what does it complain about if it can assume they are unused ? :) |
Weeeeell, I have no idea what's happening then :D. |
Mystery solved: it's when function arguments are prefixed. local function foo(_bar, baz)
return _bar + baz
end |
Fix koreader#9887, likely a regression since koreader#9669
Good to know: it now defaults to assuming that function arguments prefixed with an underscore would be unused, which is... nice in theory ;).
e.g., on the Sage, only the power_dev stuff changes between HW revs, the rest of the hardcoded stuff is still relevant, so skip probing those.
the main one *always* has priority, regardless of mtime. Should help workaround weird local/UTC issues when USBMS is involved, c.f., koreader#9227 (comment)
Fix koreader#9883, regression since c7f5bfb
Use the same mtime (the newest) for both entries, so that priority order wins the tie, giving priority to the primary. Avoids preferring an older legacy sidecar in case of issues with the new files.
our own timeout can actually be honored. Fix koreader#7874, fix koreader#9868
NiLuJe
left a comment
There was a problem hiding this comment.
Reviewed 2 of 5 files at r1, 8 of 8 files at r2, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion
Regression since koreader#9902, because I'd forgotten that the default Kobo block *does* set those... Reported by @ptrm on Gitter, thanks!
Regression since koreader#9902, because I'd forgotten that the default Kobo block *does* set those... Reported by @ptrm on Gitter, thanks! Take two, essentially revert 5974425 We have defaults for most of those in the Kobo superclass, so these can never really be `nil`.
(Rebase me)
This change is