* Only switch to ondemand when we actually can, and when it's better than the current governor...
This potentially leaves Mk.5 in the lurch, but there's no perfect solution there :/.
* Switch to UI for the unmark event of navigation hints markers
Might help on non-REAGL devices, and doesn't hurt there (re #3983).
* Switch SQLite DBs to WAL
Sounds nice in theory, behaves fine in practice.
* Switch to ondemand CPUFreq governor on Kobo
My H2O boots with the userspace (!!) governor...
Note that the hardware appears to blissfully ignore any of this, because
the stats still show the CPU changing state on its own...
* Handle OTA updates during the restart loop on Kobo/Kindle
So we can actually process OTA updates without ever truly exiting
KOReader (i.e., via the "Restart KOReader" menu entry) ;).
* With a bonus version normalizer fix.
* And a zsync bump/fix to avoid softlocks with OpenStack HTTP frontends (also, pull those from the OTA mirror list).
* Better KOA2 support
* Fix FL
* Add physical pageturn keys
* Add accelerometer support
* Possibly better NaturalLight handling on the Clara
* Fix WiFi module name used in some(?) Mk7 Kobos
* A bunch of kodev fixes
* Link to the WiKi, it's a bit more explainy ;).
* More WiFi trickery for some HW revs
* Fix Rev2 detection under KSM
That's to future-proof it, the other end of this hasn't made it into
current KSM build yet.
* Resync nickel.sh w/ current rcS
We don't have CPU, but we have PLATFORM, which is based on CPU, so,
that'll do :).
* Go back to using rmmod instead of modprobe -r
Functionally identical on current FW, will potentially avoid getting
outsmarted by modprobe if Kobo ever deigns shipping an actually usable
Linux system one day...
* Don't crash if screensavers directory doesn't exist
c.f., https://www.mobileread.com/forums/showpost.php?p=3706979&postcount=2919
* Switch all initial highlights to "fast" update
i.e., everything that does an invert
Plus a few other things that refresh small UI elements onTap
Re #3130
* Tweak refreshtype for a number of widgets:
* Fix iconbutton dimen
* Make touchmenu flash on close & initial menu popup. Full-screen on close.
* Use flashing updates when opening/closing dictionary popup. Full-screen on close.
* Switch FileManager to partial.
It's mostly text, and we want flash promotion there.
* Make configdialog & menu flash on exit
* Make FLWidget flash on close
* virtualkeyboard: flash on layout change & popup.
* Potentially not that great workaround to ensure we actually see the
highlights in the FM's chevrons
* Flash when closing BookStatus Widget
* Optimize away a quirk of the dual "fast" update in touchmenu
* Promote updates to flashing slightly more agressively.
* Document what each refreshtype actually does.
With a few guidelines on their optimal usecases.
* Switch remaining scheduleIn(0.0) to nextTick()
* Tighter scheduling timers
Shaving a hundred ms off UI callbacks...
* Cache FFI C Library namespace
* Ask MuPDF to convert pixmaps to BGR on Kobo
Fix#3949
* Mention koxtoolchain in the README
re #3972
* Kindle: Handle *all* fonts via EXT_FONT_DIR instead of bind mounts insanity
* Make black flashes in UI elements user-configurable
(All or nothing).
* Jot down some random KOA2 sysfs path
* Refresh Kindle model ID routines
* Pickup current OTA packages
We stopped shipping files w/ the full .tar.gz extension a looooong time
ago.
* And actually generally handle current packages properly
* Kindle screensaver handling experiment
WIP, because there's a fair bit of insanity left in there.
Namely, USBMS is anathema. We simply shouldn't do that, at all,
but the system allows us to do it and basically shoot ourselves in the
head one way or another.
* Don't try to handle the insanity that would be USBMS on Kindles
* Yay, one less thing to worry about :).
* Okay, that should be much saner...
Since the whole deal w/ letting the WM handle stuff was for SO, restrict
that to SO devices.
The other concern was USBMS, but we can't support it.
* Reword that
* And move that comment inside the branch, like its counterpart
* Trim unneeded stuff from startup script
I was somehow convinced I'd already done that...
While we're there, explain why we need to siphon those specific vars
* Fix a stray eth0
-> $INTERFACE
* Be very very sure we have INTERFACE set in our env
re #3936
* Make getFirmwareVersion less fragile on Kobo
Not that we actually use it right now, but, still. :D
* Use the same syntax as the PRODUCT check
* Actually implement getProductId
Instead of a stray c/p ^^
* Properly identify the Rev2/Mark7 variants of existing devices
Namely, the H2O² and Aura SE
Not that the H2O²r2 support is still broken, this just allows us to
implement it cleanyl without breaking handling of the original H2O²
re #3925
* Tweak sleeps a bit around Kobo WiFi modules...
See if that jog things up (re #3936)
* Try harder not to suspend with WiFi on on Kobos
Because otherwise, things go boom. (re #3936)
Regression after bdb82d09c1
TL;DR: That thing is race-y as hell.
The more I play with Kobos, the more it feels like the whole thing's
held together by pieces of fraying strings...
* Flag those scripts as executable in git
Somewhat irrelevant because we'll end up living on a FAT32 drive, but,
still.
* Cleanup Kobo startup script
Support KFMon >= 0.9.5
Don't siphon PRODUCT from nickel, it's exported by rcS, so fmon/KFMon
already inherit it.
Siphon NICKEL_HOME, on the off-chance nickel fails to figure that one
out for itself on restart.
Siphon LANG (This may be a terrible idea, rcS sets LANG to en_US.UTF-8,
while we set LC_ALL to en_US.UTF-8, but I don't know if nickel itself
ever updates LANG, since I have mine set to en_US also ;)).
My secret hope if that everything's working as it should and this ensures
we default to Nickel's locale on fresh installs?
Remove the extra sync on startup, both @Baskerville's fmon and KFMon are
smart enough not to do anything overly stupid, and recent FW versions
have a slightly more resilient DB anyway (rollback journal -> WAL).
This effectively shaves over a second off our startup time.
* Slimmed down the nickel restart script.
Based on current rcS
Scrapped a bucketload of irrelevant & legacy crap, since we're never
bootstrapping nickel, only restarting it.
* Update Kobo install instrcutions.
Point to KSM09, @Baskerville's fmon, and KFMon.
Deprecate legacy fmon, and as such, stop shipping a useless KoboRoot
tarball.
You'll still have to call it with `ANDROID_ARCH=x86 ./kodev build/release/run android`.
Don't forget to `./mk-luajit.sh clean` in luajit-launcher when changing architectures.
* Bump android-luajit-launcher
This includes the fix for Android 8. Fixes#3126.
* Bump base
Also improve SDK installation for ease of setting up development system.
This also puts in place most of the missing pieces to finish the intent in #3064.
* Travis: speed up by caching base and running luacheck earlier
* ignore bin and install for git status change detection
* skip coverage except on official master branch. It adds 3 whole minutes and does nothing to prevent regressions
* also cache ~/.luarocks. It evens out but would generally prevent remote timeout shenenigans
* remove base cache dir before caching with verbose remove to see what's going on
* more inclusive shell code quality analysis
* fixed more shellcheck issues
* better shellcheck/shfmt debugging info
Disable wakeup_count handling, that's often causing more harm than
anything else...
Nickel doesn't even use it, that should have been a hint ;).
Make the logging slightly less confusing.
Add commented out debug features (dumping relevant dmesg output).
Re #2188
When using KFMon, make the reboot on exit user-configurable.
The user can now choose to try to restart Nickel directly, like with
fmon.
For now, we still do a full reboot by default.
to make koreader on Android more stable
and with these opt params:
```
require("jit.opt").start("sizemcode=64","maxmcode=64", "hotloop=10000")
```
The strategy here is that we only use precious mcode memory (jitting)
on deep loops like the several blitting methods in blitbuffer.lua and
the pixel-copying methods in mupdf.lua. So that a small amount of mcode
memory (64KB) allocated when koreader is launched in the android.lua
is enough for the program and it won't need to jit other parts of lua
code and thus won't allocate mcode memory any more which by our
observation will be harder and harder as we run koreader.
