Full refresh when highlighting text is quite annoying, I fixed it by
using 'ui' refresh type.
And this patch also fixed a bug that after cleaning temporary highlights
the screen is not refreshed.
This is a larger clean-up of the refresh situation.
The general shift is that refreshes are now mainly triggered by
the (top-level) widgets when they get shown or closed via UIManager.
All refreshes for the widgets when they are in use were handled by
themselves before. This adds the case of showing/closing.
It is the desired result of not having UIManager:show()/:close()
do (full screen) refreshes on its own.
This eliminates the API difference between the extra parameters of
UIManager:show() and setDirty(). They work the same now.
Note that this also eliminates the automatic refresh that took place
before when using show() without refresh options. It always refreshed
the full screen, which led to too big refresh regions all over the
place. Thus, refresh has now explicitly to be asked for, hopefully
encouraging to implement it in the widget that gets shown (and is
aware about the screen region it covers).
Also add an event that is triggered when a widget is closed:
CloseWidget. So a widget can implement "onCloseWidget()" to trigger
actions upon closing - most commonly, this is a refresh for the area
previously taken by the widget. That way, the widget's user does not
have to take measures to ensure that the area is refreshed later.
for now, we have show() automatically call setDirty() for the new
widget, as before. However, now show() takes two arguments for
refresh configuration that will get passed on to setDirty().
For compatibility, the default is here in show() to do a partial
refresh. So if you want no refresh triggered (via this show() call),
add a function that doesn't return anything.
it was set to a full refresh. However, we want to behave as if in
non-scrolling mode and issue a partial refresh. That might get
updated to a full refresh if the full-refresh counter has reached
the limit - which is configurable.
This serves as a good example for the way refreshes are done:
setDirty("all", function() ... end)
* the "all" will have all widgets on screen repainted.
In this case that is needed because the config pane has
different sizes, covering different parts of underlying
widgets. So they need to be repainted every time.
* the function will return the area to refresh and is evaluated
after painting. In this example, we take the area that is covered
by the config pane before switching it (if present at all), and
hand it to the refresh area function as an upvalue.
When the function is called later after painting, it will
use that saved area and combine it with the area that is covered
then by the widget. That way, parts that are covered no more are
included in the refresh area, too.
See documentation in the code.
In short: There is now one single method, setDirty(), that triggers
repaints and/or refreshes.
All variables in UIManager are gone - at least from an external
perspective. Everything is done through setDirty().
This also allows for easier debugging, since all requests come
in via function calls.
when the image (e.g. a PNG) does contain an alpha channel, that can
be honored by ImageWidget. It doesn't do so by default for compatibility
(arguably, we should change that in the future), it has to be enabled
by setting the "alpha" property to "true" (boolean, not string).
external data (and in bad cases our own) can contain invalid byte
sequences in UTF8 strings. A prominent example are file names.
There was a 1-off bug in calculating the allowed length for multibyte
chars, and the iterator was a bit too greedy when stumbling upon
invalid sequences, returning a single "invalid" char for a sequence
up to the point where it became invalid in calculation. Now, we present
one invalid char for the first byte of that sequence and then check
for a valid char starting with the next byte.
This is the remaining gruntwork of #1276. I believe that only leaves networkmgr.lua and filemanagersearch.lua, which will require a little more thought.
our network manager script isn't the beauty of the code base.
However, this fixes a case where it would crash the reader when an
external command fails.
fixes#1279.
in order to have debugging facilities in framebuffer:init(), we hand
over the debug function as soon as we can.
Also, set a viewport for Kobo Mini. Hopefully, it fits most people -
I can only test on my unit.
This allows for Viewport debugging. It is configured via an environment
variable, EMULATE_READER_VIEWPORT. It works with Lua table syntax, e.g.:
EMULATE_READER_VIEWPORT="{x=10,w=550,y=5,h=790}"
the height of filechooser is provided when creating filechooser
widget since there may be no title as in filemanger which needs
a special title widget to contain filemanger menu.
Since the new Screen:getSize modthod returns a regular table
instead of a Geom which is set to the dimen field of readerview,
there is no copy method in the dimen field. And we don't need to
copy dimen any more since the result won't be modified. Even the
screen page area somehow needs to be modified, the caller of this
function should make a copy of that variable by itself.
Events have been passed to the frontend/device (or /input) code before.
Some of them have been handled in the FFI/input code, however. That
seems to be highly critical with regard to timing, though, so we just
let it enqueue the event for our code to process later. That way, a
mutex that locks the input event queue can be freed faster.
Lots of the device-related distinction wandered into
base/ffi/framebuffer_<driver>. This eases the refresh logic in
UI manager, which basically only decides what kind of refresh
to trigger. The device specific configuration in the framebuffer
driver decides how to realize that whish.
screen.lua is gone, in its place is now the framebuffer driver.
The device abstraction decides what framebuffer driver to load.
While looking into #1219 I accidentally ended up refactoring some stuff.
Tested in emulator and on H2O, but be wary because I might have overlooked something.
Don't increase counter for regional updates
Also some workarounds for Kobos
Try to avoid update_regions_func poisoning
Reset it at the end of repaint() even if nothing was found dirty
Ensure regional updates are always PARTIAL, in
case we get a region attached to an automatically triggered refresh, not
marked force_partial [which, hey, shouldn't happen, but apparently does
sometimes ^^]
This adds more documentation about the viewport offset calculations
for refreshing rotated viewports.
The old behaviour also was (still) buggy, even after the latest round
of fixes. Now, the offsets when rotated are hopefully calculated correct.
This will convert any file name to lowercase before doing the comparison.
Note that this will only work for ASCII character range, a full Unicode
aware solution will be much more complicated. And in the end, file names
are byte arrays, not character strings ;-)
fixes#1183.
The strcoll() workaround we had in place for Kobo devices was (or has
become) ineffective. We had set self.strcoll to nil on Kobo devices -
but this was the instance variable. Setting it to nil effectively makes
the instance variable vanish, so when trying to access it later, it
was not there and got looked up via the metatable - which had the original
reference. Setting it to nil had no effect whatsoever.
We simplify that approach and set the replacement function where before we
had set this to nil.
This is a partial fix for issue #1183 (and explains a comment in issue #686
which says that the old fix did not work).
However, to really fix#1183 - if we want to do so - we would need a collate
function that normalizes uppercase/lowercase before compare.
The calculation did not correctly calculate the offset in rotated
modes when the rotation was 90 or 270 degrees. Under these circumstances,
getWidth() and getHeight() did return rotated values for the framebuffer,
while Device.viewport did return not rotated values.
Since commit 12a76fee33, we had a potential
bug on the event mechanism:
It introduced (besides the checkTasks method itself) a second run of the
checkTasks() method. In the second run, however, scheduled events were
not taken into consideration in how long to wait for input events
afterwards.
So when the after the first run of checkTasks() there were new scheduled
tasks added to the task queue, they were not properly scheduled and
and depended on an already existing scheduled event or an input event
to trigger.
This might have led to unexpected order of execution (though the order
is not guaranteed by the task scheduling anyway!) or to events triggering
not at all until the next input event.
And koreader release version is normalized to a 10 digits containing
a four digits year field followed by a two digits month field followed
by a four digits revision field like "2014041079".
Don't do regional footer refreshes when we're in flipping mode.
There's probably a more elegant/correct way to handle this, but this
does the job ;).
Fix#1140
the cache would behave badly when the same item was insert()ed twice:
it would add the size twice to memory consumption, but would never substract
it twice when purging the (actually single) object from cache. So the cache
would seem to fill up while in fact it wasn't.
re #1119
AUTO appears to be doing the right thing...
(even if that baffles me, given the state of the Kernel sources,
unless they flip some switches at compile time
[NTX_WFM_MODE_OPTIMIZED_REAGL / NO_AUTO_REAGL_MODE] ...)
Anyway, that's what nickel does, so follow its lead ;).
and remove the wifi toggler on the footer of each menu page,
so that network status is only checked (currently with the stdout of ip cmd)
when navigating to the "Network settings" submenu instead of checking
on each menu popup.
Don't hijack forced partial updates (i.e., from UI elements) on always
FULL REAGL devices. It doesn't implode if we don't, and it makes for a
snappier UI.
WIP, untested, can probably simplified some more.
Good news is, it should be smarter, and thus provide a smoother user
experience :).
Relies on the relevant changes in base.
FIXME!
Needs actual confirmation from people with the HW!!
Once that's done, some device checks will probably need to be
udated in the codebase to take those new devices into account ;).
We wouldn't want those to become full updates ;).
Also apply an optionally different waveform mode for such partial,
regional updates (right now, only Kindle devices make use of this finer
grained control).
This should fix#1016.
Usage for #1016:
Create a file named "custom.event.map.lua" in the koreader directory
with the following content:
return {
[102] = "LPgFwd",
}
which currently just sets free the limitation of panning gestures
emitting rate. This should fix#1039 when unchecking the
"E-ink optimization" in the "Screen settings".
This is a major overhaul of the hardware abstraction layer.
A few notes:
General platform distinction happens in
frontend/device.lua
which will delegate everything else to
frontend/device/<platform_name>/device.lua
which should extend
frontend/device/generic/device.lua
Screen handling is implemented in
frontend/device/screen.lua
which includes the *functionality* to support device specifics.
Actually setting up the device specific functionality, however,
is done in the device specific setup code in the relevant
device.lua file.
The same goes for input handling.
ConfirmBox and InfoMessage are default to be modal.
Now returning to filamanager after highlighting a PDF page
the confirmbox asking users to save the document will not be hidden by
the filamanager window.
And it's tested on Kindle that #791 is already been solved probably by
out refacorting of MuPDF backend.
And also fix a touchmenu bug that prevent some menu entries
(always at the bottom of the menu) from showing submenus. The reason is
that the onTapSelect method of TouchMenuItem doesn't return true, so that
the tap gesture will propagate to the parent widget and be captured by
touchmenu itself. But at this time the dimen of touchmenu is changed to
the submenu of the tapped menuitem. If the submenu is short enough the
tap will be outside of touchmenu which will cause a TapCloseAllMenus
event and the submenu will never be shown.
so that when document is modified the persistent cache will
be invalidated automatically because the cache key will not
be matched. There is no perfermance overhead here at all. We
even don't need to check the modification time of the cache item
on disk, because the name of the on disk cache is a md5sum of the
cacheitem key, now the filename of the cache files contains the
modification time information.
If the document is modified since one rendered page is cached to disk,
the cache key won't match the cache file. And the cache file will
be discarded without the need to open the cache file or to check
the modification time of the cache file itself.
colors were a mixture of 4bpp integers (0=white, 15=black) and
fractional blackness levels (0=white, 1.0=black) before. This is
now unified to use the color specification of the Blitbuffer API.
The #1011 bug is caused by the `contains` check of link box with
visible area. When the link is at the very right (left?) of
the page, enlarged link box often exceeds the page bound rendering
the mentioned `contains` check failed. This patch uses an `intersectWith`
test to check if the link is in current page.
This will use the new API where the Lua/C interface was used before
and also drop usage of mupdfimg wrapper which is now implemented
within the mupdf interface
rendertext.lua did use addblitFrom() for rendering text - i.e. blitting
the letters to a BlitBuffer. However, it used intensity=1.0, which is
the same as doing a (faster, more efficient) blitFrom(). So use that
instead.
What was probably intented here is a different kind of blitting - using
the bitbuffer of the glyph as a mask.
This plugin implements a simple Calibre Companion protocol that
communicates with Calibre Wireless Server from which users can
send documents to Koreader devices directly with WIFI connection.
Note that Calibre Companion(CC) is a trade mark held by MultiPie Ltd.
The Android app Calibre Companion provided by MultiPie is closed-source.
This plugin only implements a subset function of CC according to the
open-source smart device driver from Calibre source tree.
More details can be found at calibre/devices/smart_device_app/driver.py.