Some Hold and move (hold on title and move away from it) would not work.
Pan (=swipe with hold at end) while selecting text would move the
window (it now does nothing: proper text selection still needs Hold
on word at start).
Also increase hold duration from 2s to 3s (for switching lookup
between dict/wikipedia) to be consistent with the 3s duration in
readerhighlight.
Some big chunks that compressed well, could, when uncompressed,
not fit in some buffer. This would cause invalidation of the full
cache, and provoke a full reload and re-rendering.
Remove the subprocess management functions from xutil.lua, as they were
moved into base/ffi/util.lua, and use them from there.
Also use the cre_storage_size_factor setting when processing credocuments,
to avoid CRE WARNING while indexing too.
Fix dealing with MuPDF document opening failures (previously, these
were not noticed and indexing was retried each time).
Bump crengine for cache re-use fix and warnings when
cache sizes reached.
Allow for increasing some crengine cache sizes with a new
setting: "cre_storage_size_factor" = 1.5 (or 2, or 4...)
Holding on a word for more than 3s should bring the highlight menu.
It crashed on readerpaging documents because self.hold_pos is no more
a Geom object with them.
Also inverted selected_word/selected_text logic, as with PDF, selected_word
stays set when selected_text is available.
If the device supports natural light (currently only KA1), extend the
frontlight widget to also allow for changing the 'warmth' value. We
more or less duplicate the widget for 'brightness', with the exception
that we drop the 'toggle' button, which does not make much sense for
the frontlight's warmth. Also, we add captions 'Brightness' and
'Warmth' for clarity.
New zoom configuration available among the others
Allows to easily navigate in two-column documents
In order to work properly, **scroll mode must be deactivated**
The best zoom is obtained with proper cropping around the two columns (should be the default one for content-width)
Fixes#501
With ["cre_compress_cached_data"] = false in settings.reader.lua.
Not using compression for cache files does indeed take more space,
but it does speed up opening, rendering, page turns, and closing,
with big documents.
Bump base and crengine.
ReaderLink: make all links be a table (they were a table for PDF,
but a string for CRE) for clearer code. Also have location_stack
store them as tables, with additional properties.
Get original position of link source (and verify it is valid)
so we can show a marker there.
Also:
Hold on "Go back to previous location" to clear location stack.
Resists "Swipe to go back" when previous locations stack has just
become empty, and show a notification.
Fix wrong links with Swipe to follow nearest link on PDF documents.
If the device has natural light support (currently only KA1), use the
sysfs interface instead of the ioctl-based one. It might also be
usable on other Kobo devices for setting brightness without natural
light, but since I cannot test this, this is currently restricted to
the KA1.
By default, the ioctl interface is used for accessing the
frontlight. However, the ioctl interface has two problems: the overall
brightness seems to be lower (issue #2321), and it does not support
the natural light support from the Kobo Aura One (issue #2285). This
interface is the first step to fix both these issues.
From observing Nickel's behavior, it appears the white/green/red
intensities are roughly linear dependent on the 4th root of brightness
and warmth values. This leads to a very quick increase of the
intensities for lower brightness/warmth values, which then saturate
for larger ones. The values for gain/offset were determined by a
simple linear regression, but the plan is to create a widget so that
users may adapt these values to their liking.
First item would be underlined if the device is both Touch and Keys
capable (emulator, android, pocketboot...), which is not really
pretty. Delay that till Keys for focusing are really used.
Fix various inconsitencies: some occurences not highlighted
on displayed page; different occurences highlighted on same
page wheter we went there searching backward or forward; pages
with occurences simply just skipped when searching in one
direction, and not in the other...
To avoid edge cases, crengine will now give back results on up
to 3 pages, that we need to filter.
Bump base and crengine
