Basically:
* Use `extend` for class definitions
* Use `new` for object instantiations
That includes some minor code cleanups along the way:
* Updated `Widget`'s docs to make the semantics clearer.
* Removed `should_restrict_JIT` (it's been dead code since https://github.com/koreader/android-luajit-launcher/pull/283)
* Minor refactoring of LuaSettings/LuaData/LuaDefaults/DocSettings to behave (mostly, they are instantiated via `open` instead of `new`) like everything else and handle inheritance properly (i.e., DocSettings is now a proper LuaSettings subclass).
* Default to `WidgetContainer` instead of `InputContainer` for stuff that doesn't actually setup key/gesture events.
* Ditto for explicit `*Listener` only classes, make sure they're based on `EventListener` instead of something uselessly fancier.
* Unless absolutely necessary, do not store references in class objects, ever; only values. Instead, always store references in instances, to avoid both sneaky inheritance issues, and sneaky GC pinning of stale references.
* ReaderUI: Fix one such issue with its `active_widgets` array, with critical implications, as it essentially pinned *all* of ReaderUI's modules, including their reference to the `Document` instance (i.e., that was a big-ass leak).
* Terminal: Make sure the shell is killed on plugin teardown.
* InputText: Fix Home/End/Del physical keys to behave sensibly.
* InputContainer/WidgetContainer: If necessary, compute self.dimen at paintTo time (previously, only InputContainers did, which might have had something to do with random widgets unconcerned about input using it as a baseclass instead of WidgetContainer...).
* OverlapGroup: Compute self.dimen at *init* time, because for some reason it needs to do that, but do it directly in OverlapGroup instead of going through a weird WidgetContainer method that it was the sole user of.
* ReaderCropping: Under no circumstances should a Document instance member (here, self.bbox) risk being `nil`ed!
* Kobo: Minor code cleanups.
* Allow doing away with CacheItem
Now that we have working FFI finalizers on BBs, it's mostly useless overhead.
We only keep it for DocCache, because it's slightly larger, and memory pressure might put us in a do or die situation where waiting for the GC might mean an OOM kill.
* Expose's LRU slot-only mode
And use it for CatalogCache, which doesn't care about storage space
* Make GlyphCache slots only (storage space is insignificant here, it was
always going to be evicted by running out of slots).
* More informative warning when we chop the cache in half
Ought to be faster than our naive array-based approach.
Especially for the glyph cache, which has a solid amount of elements,
and is mostly cache hits.
(There are few things worse for performance in Lua than
table.remove @ !tail and table.insert @ !tail, which this was full of :/).
DocCache: New module that's now an actual Cache instance instead of a
weird hack. Replaces "Cache" (the instance) as used across Document &
co.
Only Cache instance with on-disk persistence.
ImageCache: Update to new Cache.
GlyphCache: Update to new Cache.
Also, actually free glyph bbs on eviction.
* Minor updates to the min & max cache sizes (16 & 64MB). Mostly to satisfy my power-of-two OCD.
* Purge broken on-disk cache files
* Optimize free RAM computations
* Start dropping LRU items when running low on memory before pre-rendring (hinting) pages in non-reflowable documents.
* Make serialize dump the most recently *displayed* page, as the actual MRU item is the most recently *hinted* page, not the current one.
* Use more accurate item size estimations across the whole codebase.
TileCacheItem:
* Drop lua-serialize in favor of Persist.
KoptInterface:
* Drop lua-serialize in favor of Persist.
* Make KOPTContext caching actually work by ensuring its hash is stable.
* Modernize jit syntax
It's built-in, no need for a require anymore.
* Flag Android should_restrict_JIT again
It's mildly helpful at putting off the inevitable.
(very, very, very, very mildly).
* Bump android-luajit-launcher
https://github.com/koreader/android-luajit-launcher/pull/281
- Add IconWidget, use it for icons instead of ImageWidget.
Specify icons by name only, look for them (with either
.svg or .png suffixes) in multiple directories (including
koreader/settings/icons/ to allow customizing them).
Don't crash when icon name not found, shown a black
background warning icon instead.
- Don't trust the icons' native sizes: replace
scale_for_dpi=true with width/height=DGENERIC_ICON_SIZE,
so all icons get the same (tunable) size - except in
a few specific use cases.
- Top and bottom menu bars: normalize, and have icons
properly centered in them, extend vertical line
separators up to the edges.
- TOC: adjust expand/collapse icons size to items size
A few fixes and enhancement related to bold text:
- When using bold=true with a regular font, use its bold
variant if one exists (can be prevented by manually
adding a setting: "use_bold_font_for_bold" = false).
- When using a bold font without bold=true, promote bold
to true, so fallback fonts are drawn bold too.
- Whether using a bold font, or using bold=true, ensure
fallback fonts are drawn bold, with their available bold
variant if one exists, or with synthetized bold.
- When using a bold variant of a fallback font, keep using
the regular variant as another fallback (as bold fonts
may contain less glyphs than their regular counterpart).
- Allow providing bold=Font.FORCE_SYNTHETIZED_BOLD to
get synth bold even when a bold font exists (might be
interesting to get text in bold the same width as the
same text non-bold).
- Use the font realname in the key when caching glyphs,
instead of our aliases (cfont, infont...) to avoid
duplication and wasting memory.
bump base for libkoreader-xtext.so:
- ffi/pic.lua: fix memory leak with some unsupported PNG files
- FreeType ffi: add methods for use with Harfbuzz shaping
- thirdparty: adds libunibreak
- Adds libkoreader-xtext.so: enhanced text shaping
Add a getFallbackFont(N) to each Lua font object to instantiate
(if not yet done) and return the Nth fallback font for the font.
Fallback fonts list: add NotoSansArabicUI-Regular.ttf
Add RenderText:getGlyphByIndex() to get a glyph bitmap by glyph
index, which is what we'll get from Harfbuzz shaping.
(RenderText:getGlyph() works with unicode charcode).
Lots of code was doing some renderText calls to get the size
of some text string, and truncate it to some width if needed,
with or without an added ellipsis, before instantiating
a TextWidget with that tweaked text string.
This PR fixes/adds some properties and methods to TextWidget
so all that can be done by it. It makes the calling code
simpler, as they don't need to use RenderText directly.
(Additionally, when we go at using Harfbuzz for text rendering,
we'll just have to update or replace textwidget.lua without
the need to update any higher level code.)
Also:
- RenderText: removed the space added by truncateTextByWidth
after the ellipsis, as it doesn't feel needed, and break
right alignment of the ellipsis with other texts.
- KeyValuePage: fix some subtle size and alignment issues.
- NumberPickerWidget: fix font size (provided font size was
not used)
This commit standardizes the various todos around the code a bit in a manner recognized by LDoc.
Besides drawing more attention by being displayed in the developer docs, they're also extractable with LDoc on the command line:
```sh
ldoc --tags todo,fixme *.lua
```
However, whether that particular usage offers any advantage over other search tools is questionable at best.
* and some random beautification
InputText: checks whether provided content can be given
back unaltered, which may not be the case after it is
splitted to UTF8 chars if the text is binary content.
Prevent editing text if that is the case.
Adds InputText and InputDialog :isEditable() and :isEdited()
methods.
Also accounts for the scrollbar width when measuring text
to prevent it from being displayed when not needed.
Also ensure a minimal size of the scrollbar thumb so it is
rendered when huge text with many lines is displayed.
Virtual keyboard: Hold on Backspace: delete from cursor
to start of line instead of clearing all text content.
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.
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 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.
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.
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.
