* Device: Add a `hasSeamlessWifiToggle` devcap to complement `hasWifiToggle`, to denote platforms where we can toggle WiFi without losing focus, as this has obvious UX impacts, and less obvious technical impacts on some of the NetworkMgr innards...
* Android: Mark as `!hasSeamlessWifiToggle`, as it requires losing focus to the system settings. Moreover, `turnOnWifi` returns *immediately* and we *still* run in the background during that time, for extra spiciness...
* NetworkMgr: Ensure only *one* call to `turnOnWifi` will actually go on when stuff gets re-scheduled by the `beforeWifiAction` framework.
* NetworkMgr: Ensure the `beforeWifiAction` framework will not re-schedule the same thing *ad vitam aeternam* if a previous connection attempt is still ongoing. (i.e., previously, on Android, if you backed out of the system settings, you entered the Benny Hill dimension, as NetworkMgr would keep throwing you back into the system settings ;p). This has a few implications on callbacks requested by subsequent connection attempts, though. Generally, we'll try to honor *explicitly interactive* callbacks, but `beforeWifiAction` stuff will be dropped (only the original cb is preserved). That's what prevents the aforementioned infinite loop, as the `beforeWifiAction` framework was based on the assumption that `turnOnWifi` somewhat guaranteed `isConnected` to be true on return, something which is only actually true on `hasWifiManager` platforms.
* NetworkMgr: In `prompt` mode, the above implies that the prompt will not even be shown for concurrent attempts, as it's otherwise extremely confusing (KOSync on Android being a prime example, as it has a pair of Suspend/Resume handlers, so the initial attempt trips those two because of the focus switch >_<").
* NetworkMgr: Don't attempt to kill wifi when aborting a connection attempt on `!hasSeamlessWifiToggle` (because, again, it'll break UX, and also because it might run at very awkward times (e.g., I managed to go back to KOReader *between* a FM/Reader switch at one point, which promptly caused `UIManager` to exit because there was nothing to show ;p).
* NetworkMgr: Don't drop the connectivity callback when `beforeWifiAction` is set to prompt and the target happens to use a connectivity check in its `turnOnWifi` implementation (e.g., on Kindle).
* Android: Add an `"ignore"` `beforeWifiAction` mode, that'll do nothing but schedule the connectivity check with its callback (with the intent being the system will eventually enable wifi on its own Soon(TM)). If you're already online, the callback will run immediately, obviously. If you followed the early discussions on this PR, this closely matches what happens on `!hasWifiToggle` platforms (as flagging Android that way was one of the possible approaches here).
* NetworkMgr: Bail out early in `goOnlineToRun` if `beforeWifiAction` isn't `"turn_on"`. Prompt cannot work there, and while ignore technically could, it would serve very little purpose given its intended use case.
* KOSync: Neuter the Resume/Suspend handlers early on `CloseDocument`, as this is how focus switches are handled on Android, and if `beforeWifiAction` is `turn_on` and you were offline at the time, we'd trip them because of the swap to system settings to enable wifi.
* KOSync: Allow `auto_sync` to be enabled regardless of the `beforeWifiAction` mode on `!hasSeamlessWifiToggle` platforms. Prompt is still a terrible idea, but given that `goOnlineToRun` now aborts early if the mode is not supported, it's less of a problem.
Move as much of the state tracking as possible inside VirtualKeyboard itself.
InputDialog unfortunately needs an internal tracking of this state because it needs to know about it *before* the VK is shown, so we have to keep a bit of duplication in there, although we do try much harder to keep everything in sync (at least at function call edges), and to keep the damage contained to, essentially, the toggle button's handler.
(Followup to #10803 & #10850)
* Enforce a minimal standby timer for the first standby after a resume.
* On Kobo, sleep a bit before standby.
This aims to alleviate race conditions causing visible refresh glitches on sunxi, especially when using an extremely low standby timer (i.e., below the defaults).
wpa_supplicant returns all non-ASCII SSIDs as raw bytes in the form
\x0a. We interpret these bytes as UTF-8, and make sure that all invalid
characters are replaced with a �.
Fix issue introduced by 976aaf5f: with full screen
text editor and hiding the keyboard, a tap in the
text area would show a new keyboard hiding the
buttons allowing to save/close the text editor,
getting us stuck there.
An attempt was made in the original code, but the whole thing was designed in the hope of actually switching to turbo, so it was super janky without it.
Anyway, we now actually have a sane way to set socket timeouts, so, use that, and set them very tight for now.
This is fairly critical right now, because the server is down, and the default timeouts are ~30s. That happens to be *above* the debounce threshold, so you can't even hope for that to help you. Meaning, right now, you get a 2 * 30s block on resume with auto sync. That's... Very Not Good(TM).
That becomes a single 2s one after this.
Makes more sense this way.
Re: #10828
Also clarify the UIManager "no dialogs left" message, and drop the
return value, as it's meaningless, we just want to break & return.
Assume the host system does things right otherwise.
Should be sane on Kindle
On PocketBook, who knows, but assuming the device actually suspends,
that should effectively kill our keepalive
Irrelevant for Android, as we skipped it because the wifi toggling
methods are interactive.
Fix#10823
(For reference, we *enable* wifi_was_on no matter *how* wifi is enabled,
but we only toggle it off when it's killed by a *direct* user interaction,
the intent being that if *something* non-interactive enabled wifi,
you'll probably silently need it on resume too).
As early as turnOnWifi.
Implement it on hasWifiManager platforms, preventing useless
connectivity checks to run when they're obviously never going to work
because you're out of range of your AP.
Also implemented a flag to notify the backend if the connection attempt
was interactive or not.
Right now, interactive is extremely restricted, it basically means the
menu checkmark, or a gesture.
The intent being that for stuff like the beforeWifiAction framework, we
don't want the backend to spawn extra UI.
Specifically, for hasWifiManager platforms, we no longer spawn the AP
scan list on failure unless the caller was interactive.
TL;DR: beforeWifiAction is now much less obnoxious when you're obviously
not able to connect.
onCloseDocument is way too early, if another, later onCloseDocument
handler trips a ReaderUI repaint, a new task will be scheduled, and that
one will never be cancelled, resulting in it running post-teardown,
crashing horribly.
This was for instance very easy to trigger via KOSync.
Fix#10806
MultiInputDialog does a close -> show in the same callback,
so if we don't actually keep it in sync with the actual state, we lose
the keyboard and essentially softlock the UI, which is Very Bad(TM).
NOTE: InputDialog has its own keyboard_hidden flag, and it looks...
fairly nightmarish.
This is semantically more correct, and should prevent platform-specific
weirdness if an ordering concern ever comes up (e.g., like it did for
the viewport stuff).
Make this a real boy, with a transient lipc handle.
And get rid of the insane 1s sleep on affected ReaderView paints,
because ouchy.
This is completely deprecated anyway, so this is entirely pointless,
and mainly to prevent implementation details from creeping into
reader.lua.
Re #10743
Note that this only makes faulty switches slightly less annoying: for a
stuck switch, instead of a string of page turns, you'll get a single
missed page turn on the tap that actually releases the stuck contact...
The Kindle swipe animations & physical key inversion apply *everywhere*,
so we need this accessible in the FM, too.
Move it to Navigation, which is where it already was on !Touch devices.
we use it as text anyway and test if it is the empty string also.
this fixes a crash in viewhtml when holding a selector with an empty clipboard, we try to contcate getClipboardText which is nil (on !SDL & !Android)
And add an OTM block to do a cleanup pass on existing DBs (which might take a while if you're severely affected, because we've seen reports of DBs north of 2GB).
The ToC is no longer in the HTML we get from the
Wikipedia API. So, add it ourselves.
Also, as we can't get anywhere the Wikipedia localized
string for "Contents" (ie. "Sommaire" in French), use
thick <HR> to mark the start and end of this ToC.
* Notification: Drop the fencing from #10083; it never actually helped, and had subtle side-effects we could do without.
* VirtualKeyBoard: Flash on close, otherwise, some of the fast refresh glitches may be burned into the working buffer until a flash. Making sure we flash ourselves prevent it from sticking around on the page ;).
* util: Move `writeToSysfs` to base (i.e., `ffi/util`), as we need it there (and it actually makes more sense there anyway ;p).
* Bump base for https://github.com/koreader/koreader-base/pull/1645, which is where the actual workaround (hopefully) lives.
Re #8414, #9806, #10558
The HTML we get from Wikipedia has recently changed,
which has caused images to not be detected and shown
in DictQuickLookup, and to be missing or ugly in
saved EPUBs. Update our code and CSS to display them
again.
On long-press on the "Show matched stylesheets rules"
buttons, ignore those from our epub.css/html5.css and
from styletweaks, which may make investigating a book
style simpler without their noise.
It's been made redundant by the RPATH changes
The only platform that gets the dubious honor of actually needing an
LD_LIBRARY_PATH is PocketBook, because of InkView.
Co-authored-by: Benoit Pierre <benoit.pierre@gmail.com>
Bump base to pull in the aforementioned RPATH changes ;).
https://github.com/koreader/koreader-base/pull/1638
Includes:
- bump LunaSVG: intermediate upstream bump, cleanup
- bump LunaSVG: minor upstream tweaks
crengine:
- DocX: add support for similar DocM format
- LVStyleSheet: fix LVCssDeclaration::getHash()
- CSS parsing: accept Unicode values for ID and classnames
- update for Harfbuzz 8, fix some compiler warning
Also fix input not restored when loading failed,
and KOReader not able to exit.
- Add a toggle in Font settings> allowing showing font
ordered by most recently selected (long-press on
it allows clearing this history).
- Keep in G_reader_settings a list of known fonts, so
we can notice newly added user fonts, and put them
at the start of the most recently selected. Show
these new fonts with a symbol in the menu.
- TouchMenu: allows for a flag to trigger menu
refresh when going up.
* Enable before_wifi_action & after_wifi_action on hasWifiToggle platforms (which is basically all of 'em except naked SDL).
* Decouple restoreWifiAsync from hasWifiManger, because we can do that on other platforms (namely, Kindle. Probably PB, too, but WiFi is already a mess there, and I can't test it).
* Implement restoreWifiAsync on Kindle.
* Properly flag rM as hasWifiManager & hasFastWifiStatusQuery, because it is actually both of those (it uses our wpa_supplicant backend).
* Update the KOSync checks to take these changes into account, to properly disable auto_sync if necessary.
* Really made the Network* event signaling consistent. For realz this time.
* In an effort to make the whole beforeWifiAction framework somewhat usable there, we now assume connectivity is always available on !hasWifiToggle platforms...
Namely, that swipes report the gesture's *starting* coordinates as `pos` (for hit-detection purposes), while pans (and basically everything else) use the *lift* coordinates (as well as provide a relative vector).
In each case, we now also provide a separate `start_pos`/`end_pos`, in case handlers need finer-grained data (e.g., anchoring UI elements to a swipe's lift coordinates).
Extended addition of PocketBook InkPad 4 with firmware later U743g.6.8.885 (PB743G) because its device ID was corrected from PB743g to PB743G by firmware U743g.6.8.1719.
Instead of opening the same font multiple times for each different
size (multiple face instances), share one face instance and create
multiple size instances.
- Properly parse input text for words (the previous
code wasn't working with Greek letters)
- With multiple words search, don't allow "substring
matching" for words in the middle
- Remove support for Lua pattens, so to get proper
substring matching (as we have with cre text search)
So that Unicode NFC normalization can combine diacritics
entered via our keyboard with their base into their
canonical form (we can expect books text to be normalized).
Fix: #10539, and for context #6489, #6733, #6534
Reorganize and reword most of the settings to make it clear what actually ties into auto sync, and what doesn't. (Specifically, what happens when a pull attempts to sync forward or backward has nothing to do with auto sync, it applies in all cases; while the periodic sync *does* require auto sync).
The main point of contention, though, is that auto sync will now *always* attempt to setup network connectivity (i.e., on resume/suspend/close). Periodic sync will *not* though (the intent being that, if you use periodic sync, you're relying on the activity check to actually keep wifi on at all times)).
Since this may lead to a large amount of nagging about wifi toggles on devices w/ NetworkManager support, it is now *disabled* by default on those devices. (And given that it wouldn't have worked because of the lack of connectivity, that doesn't really make any practical difference ;p).
Additionally, given the fact that there's no way to make this behavior viable if the "before wifi" action is left at its default of "prompt", this feature now *requires* that to be set to "turn_on" (on devices where it can, of course); attempting to toggle it on will warn about that if necessary.
This change is retroactive (OTM).
Includes an assortment of fixes and cleanups, including migrating to the new LuaSettings API, which is why there's no longer a smattering of superfluous flushes.
Fix#10588
Regression since #10305
While we're there, rejig the FL toggle callback shenanigans so that implementation details don't leak through to *other* implementations.
(i.e., leave the Kobo mess in Kobo land, with only a minimal impact on the public API and its implementation).
Add a restricted but convenient mode showing BookMap like
in initial view, while still allowing chapter levels to be
tweaked. This allows getting back to this view with another
gesture to see the overall progress in the book, while
still having the normal BookMap in flat mode acting as
an alternative ToC.
Available as an action to associate to a gesture, and
with long-press on the "Book map" menu item.
Allows hairy-boxes amateurs to get small markers (under
the baseline) every 10 pages (with -/+ to get small or
medium markers, with optionally smaller ones every 5 pages).
Align page numbers (the vertical ones, on the left of
a BookMapRow) on the baseline of the BookMapRow, where
page slots rise from.
Also tweak the page number spike below the baseline used
in PageBrowser.
sometimes my kindle fires the rtc alarm a tad too late so use a more generous timeout
the real fix would be to have validateWakeupAlarmByProximity return differently if the alarm is past or in the future, but i am too lazy
Notification is a toast, so it doesn't stop event popagation.
If we don't disable thoses handlers inside Notification, we get spurious
duplicate handlers being fired ;).
Fix#10461
Export highlights for selected files.
Having a button "Select all files in folder", it is easy to export the whole folder.
So, closes#10402.
To keep even number of buttons, added a feature "Show selected files list". May be useful to check selections before an operation. Just a sorted list, no titlebar or popup menu, tapping a file jumps to its folder.
Make sure we only send Suspend/Resume events when we *actually* suspend/resume. This is done via the Device `_beforeSuspend`/`_afterResume` methods, and those were called by the *input handlers*, not the PM logic; which means they would fire, while the PM logic could actually take a smarter decision and *not* do what the event just sent implied ;).
(i.e., sleep with a cover -> suspend + actual suspend, OK; but if you then resume with a button -> input assumes resume, but PM will actually suspend again!).
Existing design issue made more apparent by #9448 ;).
Also fixes/generalizes a few corner-cases related to screen_saver_lock handling (e.g., don't allow USBMS during a lock).
And deal with the fallout of the main change to the Kobo frontlight ramp behavior ;).
Updating the selection queries the same page tens of times per second. This prevents KOReader from keeping up with high-frequency input event streams.
Fixes#10443. Relies on koreader/koreader-base#1612.
