2
0
mirror of https://github.com/koreader/koreader synced 2024-11-04 12:00:25 +00:00
koreader/frontend/ui/widget/footnotewidget.lua

423 lines
16 KiB
Lua
Raw Normal View History

local BD = require("ui/bidi")
local Blitbuffer = require("ffi/blitbuffer")
local BottomContainer = require("ui/widget/container/bottomcontainer")
local CenterContainer = require("ui/widget/container/centercontainer")
local Device = require("device")
local Event = require("ui/event")
local FrameContainer = require("ui/widget/container/framecontainer")
local Geom = require("ui/geometry")
local GestureRange = require("ui/gesturerange")
local HorizontalGroup = require("ui/widget/horizontalgroup")
local HorizontalSpan = require("ui/widget/horizontalspan")
local InputContainer = require("ui/widget/container/inputcontainer")
local LineWidget = require("ui/widget/linewidget")
local ScrollHtmlWidget = require("ui/widget/scrollhtmlwidget")
local Size = require("ui/size")
The great Input/GestureDetector/TimeVal spring cleanup (a.k.a., a saner main loop) (#7415) * ReaderDictionary: Port delay computations to TimeVal * ReaderHighlight: Port delay computations to TimeVal * ReaderView: Port delay computations to TimeVal * Android: Reset gesture detection state on APP_CMD_TERM_WINDOW. This prevents potentially being stuck in bogus gesture states when switching apps. * GestureDetector: * Port delay computations to TimeVal * Fixed delay computations to handle time warps (large and negative deltas). * Simplified timed callback handling to invalidate timers much earlier, preventing accumulating useless timers that no longer have any chance of ever detecting a gesture. * Fixed state clearing to handle the actual effective slots, instead of hard-coding slot 0 & slot 1. * Simplified timed callback handling in general, and added support for a timerfd backend for better performance and accuracy. * The improved timed callback handling allows us to detect and honor (as much as possible) the three possible clock sources usable by Linux evdev events. The only case where synthetic timestamps are used (and that only to handle timed callbacks) is limited to non-timerfd platforms where input events use a clock source that is *NOT* MONOTONIC. AFAICT, that's pretty much... PocketBook, and that's it? * Input: * Use the <linux/input.h> FFI module instead of re-declaring every constant * Fixed (verbose) debug logging of input events to actually translate said constants properly. * Completely reset gesture detection state on suspend. This should prevent bogus gesture detection on resume. * Refactored the waitEvent loop to make it easier to comprehend (hopefully) and much more efficient. Of specific note, it no longer does a crazy select spam every 100µs, instead computing and relying on sane timeouts, as afforded by switching the UI event/input loop to the MONOTONIC time base, and the refactored timed callbacks in GestureDetector. * reMarkable: Stopped enforcing synthetic timestamps on input events, as it should no longer be necessary. * TimeVal: * Refactored and simplified, especially as far as metamethods are concerned (based on <bsd/sys/time.h>). * Added a host of new methods to query the various POSIX clock sources, and made :now default to MONOTONIC. * Removed the debug guard in __sub, as time going backwards can be a perfectly normal occurrence. * New methods: * Clock sources: :realtime, :monotonic, :monotonic_coarse, :realtime_coarse, :boottime * Utility: :tonumber, :tousecs, :tomsecs, :fromnumber, :isPositive, :isZero * UIManager: * Ported event loop & scheduling to TimeVal, and switched to the MONOTONIC time base. This ensures reliable and consistent scheduling, as time is ensured never to go backwards. * Added a :getTime() method, that returns a cached TimeVal:now(), updated at the top of every UI frame. It's used throughout the codebase to cadge a syscall in circumstances where we are guaranteed that a syscall would return a mostly identical value, because very few time has passed. The only code left that does live syscalls does it because it's actually necessary for accuracy, and the only code left that does that in a REALTIME time base is code that *actually* deals with calendar time (e.g., Statistics). * DictQuickLookup: Port delay computations to TimeVal * FootNoteWidget: Port delay computations to TimeVal * HTMLBoxWidget: Port delay computations to TimeVal * Notification: Port delay computations to TimeVal * TextBoxWidget: Port delay computations to TimeVal * AutoSuspend: Port to TimeVal * AutoTurn: * Fix it so that settings are actually honored. * Port to TimeVal * BackgroundRunner: Port to TimeVal * Calibre: Port benchmarking code to TimeVal * BookInfoManager: Removed unnecessary yield in the metadata extraction subprocess now that subprocesses get scheduled properly. * All in all, these changes reduced the CPU cost of a single tap by a factor of ten (!), and got rid of an insane amount of weird poll/wakeup cycles that must have been hell on CPU schedulers and batteries..
2021-03-30 00:57:59 +00:00
local TimeVal = require("ui/timeval")
local UIManager = require("ui/uimanager")
local VerticalGroup = require("ui/widget/verticalgroup")
local VerticalSpan = require("ui/widget/verticalspan")
local _ = require("gettext")
local Screen = Device.screen
local T = require("ffi/util").template
-- If we wanted to use the default font set for the book,
-- we'd need to add a few functions to crengine and cre.cpp
-- to get the font files paths (for each font, regular, italic,
-- bold...) so we can pass them to MuPDF with:
-- @font-face {
-- font-family: 'KOReader Footnote Font';
-- src: url("%1");
-- }
-- @font-face {
-- font-family: 'KOReader Footnote Font';
-- src: url("%2");
-- font-style: italic;
-- }
-- body {
-- font-family: 'KOReader Footnote Font';
-- }
-- But it looks quite fine if we use "Noto Sans": the difference in font look
-- (Sans, vs probably Serif in the book) will help noticing this is a KOReader
-- UI element, and somehow justify the difference in looks.
-- Note: we can't use < or > in comments in the CSS, or MuPDF complains with:
-- error: css syntax error: unterminated comment.
-- So, HTML tags in comments are written upppercase (eg: <li> => LI)
-- Independent string for @page, so we can T() it individually,
-- without needing to escape % in DEFAULT_CSS
local PAGE_CSS = [[
@page {
margin: %1 %2 %3 %4;
font-family: '%5';
}
%6
]]
-- Make default MuPDF styles (source/html/html-layout.c) a bit
-- more similar to our epub.css ones, and more condensed to fit
-- in a small bottom pannel
local DEFAULT_CSS = [[
body {
margin: 0; /* MuPDF: margin: 1em */
padding: 0;
line-height: 1.3;
text-align: justify;
}
/* We keep left and right margin the same so it also displays as expected in RTL */
h1, h2, h3, h4, h5, h6 { margin: 0; } /* MuPDF: margin: XXem 0 , vary with level */
blockquote { margin: 0 1em; } /* MuPDF: margin: 1em 40px */
p { margin: 0; } /* MuPDF: margin: 1em 0 */
ol { margin: 0; } /* MuPDF: margin: 1em 0; padding: 0 0 0 30pt */
ul { margin: 0; } /* MuPDF: margin: 1em 0; padding: 0 0 0 30pt */
dl { margin: 0; } /* MuPDF: margin: 1em 0 */
dd { margin: 0 1em; } /* MuPDF: margin: 0 0 0 40px */
pre { margin: 0.5em 0; } /* MuPDF: margin: 1em 0 */
a { color: black; } /* MuPDF: color: #06C; */
/* MuPDF has no support for text-decoration, so we can't underline links,
* which is fine as we don't really need to provide link support */
/* MuPDF draws the bullet for a standalone LI outside the margin.
* Avoid it being displayed if the first node is a LI (in
* Wikipedia EPUBs, each footnote is a LI */
body > li { list-style-type: none; }
/* MuPDF always aligns the last line to the left when text-align: justify,
* which is wrong with RTL. So cancel justification on RTL elements: they
* will be correctly aligned to the right */
*[dir=rtl] { text-align: initial; }
/* Remove any (possibly multiple) backlinks in Wikipedia EPUBs footnotes */
.noprint { display: none; }
/* Style some FB2 tags not known to MuPDF */
strike, strikethrough { text-decoration: line-through; }
underline { text-decoration: underline; }
emphasis { font-style: italic; }
small { font-size: 80%; }
big { font-size: 130%; }
empty-line { display: block; padding: 0.5em; }
image { display: block; padding: 0.4em; border: 0.1em solid black; width: 0; }
date { display: block; font-style: italic; }
epigraph { display: block; font-style: italic; }
poem { display: block; }
stanza { display: block; font-style: italic; }
v { display: block; text-align: left; hyphenate: none; }
text-author { display: block; font-weight: bold; font-style: italic; }
/* Attempt to display FB2 footnotes as expected (as crengine does, putting
* the footnote number on the same line as the first paragraph via its
* support of "display: run-in" and a possibly added autoBoxing element) */
body > section > autoBoxing > *,
body > section > autoBoxing > title > *,
body > section > title,
body > section > title > p,
body > section > p {
display: inline;
}
body > section > autoBoxing + p,
body > section > p + p {
display: block;
}
body > section > autoBoxing > title,
body > section > title {
font-weight: bold;
}
]]
-- Add this if needed for debugging:
-- @page { background-color: #cccccc; }
-- body { background-color: #eeeeee; }
-- Widget to display footnote HTML content
local FootnoteWidget = InputContainer:new{
html = nil,
css = nil,
-- font_face can't really be overriden, it needs to be known by MuPDF
font_face = "Noto Sans",
-- For the doc_* values, we expect to be provided with the real
-- (already scaled) sizes in screen pixels
doc_font_size = Screen:scaleBySize(18),
doc_margins = {
left = Screen:scaleBySize(20),
right = Screen:scaleBySize(20),
top = Screen:scaleBySize(10),
bottom = Screen:scaleBySize(10),
},
follow_callback = nil,
on_tap_close_callback = nil,
close_callback = nil,
dialog = nil,
}
function FootnoteWidget:init()
-- Set widget size
self.width = Screen:getWidth()
self.height = math.floor(Screen:getHeight() * 1/3) -- will be decreased when content is smaller
if Device:isTouchDevice() then
local range = Geom:new{
x = 0, y = 0,
w = Screen:getWidth(),
h = Screen:getHeight(),
}
self.ges_events = {
TapClose = {
GestureRange:new{
ges = "tap",
range = range,
}
},
SwipeFollow = {
GestureRange:new{
ges = "swipe",
range = range,
}
},
HoldStartText = {
GestureRange:new{
ges = "hold",
range = range,
},
},
HoldPanText = {
GestureRange:new{
ges = "hold",
range = range,
},
},
HoldReleaseText = {
GestureRange:new{
ges = "hold_release",
range = range,
},
-- callback function when HoldReleaseText is handled as args
args = function(text, hold_duration)
if self.dialog then
The great Input/GestureDetector/TimeVal spring cleanup (a.k.a., a saner main loop) (#7415) * ReaderDictionary: Port delay computations to TimeVal * ReaderHighlight: Port delay computations to TimeVal * ReaderView: Port delay computations to TimeVal * Android: Reset gesture detection state on APP_CMD_TERM_WINDOW. This prevents potentially being stuck in bogus gesture states when switching apps. * GestureDetector: * Port delay computations to TimeVal * Fixed delay computations to handle time warps (large and negative deltas). * Simplified timed callback handling to invalidate timers much earlier, preventing accumulating useless timers that no longer have any chance of ever detecting a gesture. * Fixed state clearing to handle the actual effective slots, instead of hard-coding slot 0 & slot 1. * Simplified timed callback handling in general, and added support for a timerfd backend for better performance and accuracy. * The improved timed callback handling allows us to detect and honor (as much as possible) the three possible clock sources usable by Linux evdev events. The only case where synthetic timestamps are used (and that only to handle timed callbacks) is limited to non-timerfd platforms where input events use a clock source that is *NOT* MONOTONIC. AFAICT, that's pretty much... PocketBook, and that's it? * Input: * Use the <linux/input.h> FFI module instead of re-declaring every constant * Fixed (verbose) debug logging of input events to actually translate said constants properly. * Completely reset gesture detection state on suspend. This should prevent bogus gesture detection on resume. * Refactored the waitEvent loop to make it easier to comprehend (hopefully) and much more efficient. Of specific note, it no longer does a crazy select spam every 100µs, instead computing and relying on sane timeouts, as afforded by switching the UI event/input loop to the MONOTONIC time base, and the refactored timed callbacks in GestureDetector. * reMarkable: Stopped enforcing synthetic timestamps on input events, as it should no longer be necessary. * TimeVal: * Refactored and simplified, especially as far as metamethods are concerned (based on <bsd/sys/time.h>). * Added a host of new methods to query the various POSIX clock sources, and made :now default to MONOTONIC. * Removed the debug guard in __sub, as time going backwards can be a perfectly normal occurrence. * New methods: * Clock sources: :realtime, :monotonic, :monotonic_coarse, :realtime_coarse, :boottime * Utility: :tonumber, :tousecs, :tomsecs, :fromnumber, :isPositive, :isZero * UIManager: * Ported event loop & scheduling to TimeVal, and switched to the MONOTONIC time base. This ensures reliable and consistent scheduling, as time is ensured never to go backwards. * Added a :getTime() method, that returns a cached TimeVal:now(), updated at the top of every UI frame. It's used throughout the codebase to cadge a syscall in circumstances where we are guaranteed that a syscall would return a mostly identical value, because very few time has passed. The only code left that does live syscalls does it because it's actually necessary for accuracy, and the only code left that does that in a REALTIME time base is code that *actually* deals with calendar time (e.g., Statistics). * DictQuickLookup: Port delay computations to TimeVal * FootNoteWidget: Port delay computations to TimeVal * HTMLBoxWidget: Port delay computations to TimeVal * Notification: Port delay computations to TimeVal * TextBoxWidget: Port delay computations to TimeVal * AutoSuspend: Port to TimeVal * AutoTurn: * Fix it so that settings are actually honored. * Port to TimeVal * BackgroundRunner: Port to TimeVal * Calibre: Port benchmarking code to TimeVal * BookInfoManager: Removed unnecessary yield in the metadata extraction subprocess now that subprocesses get scheduled properly. * All in all, these changes reduced the CPU cost of a single tap by a factor of ten (!), and got rid of an insane amount of weird poll/wakeup cycles that must have been hell on CPU schedulers and batteries..
2021-03-30 00:57:59 +00:00
local lookup_target = hold_duration < TimeVal:new{ sec = 3 } and "LookupWord" or "LookupWikipedia"
self.dialog:handleEvent(
Event:new(lookup_target, text)
)
end
end
},
}
end
if Device:hasKeys() then
self.key_events = {
Close = { {"Back"}, doc = "cancel" },
Follow = { {"Press"}, doc = "follow link" },
}
end
-- Workaround bugs in MuPDF:
-- There is something with its handling of <BR>:
-- <div>abc<br/>def</div> : 2 lines, no space in between
-- <div>abc<br anyattr="anyvalue"/>def</div> : 3 lines, empty line in between
-- Remove any attribute, let a <br/> be a plain <br/>
self.html = self.html:gsub([[<br[^>]*>]], [[<br/>]])
-- Elements with a id= attribute get a line above them (by some internal MuPDF
-- code, possibly generate_anchor() in html-layout.c).
-- Working around it with the following does not work: *[id] {margin-top: -1em;}
-- So, just rename the id= attribute, as we don't follow links in this popup.
self.html = self.html:gsub([[(<[^>]* )[iI][dD]=]], [[%1disabledID=]])
-- We may use a font size a bit smaller than the document one (because
-- footnotes are usually smaller, and because NotoSans is a bit on the
-- larger size when compared to other fonts at the same size)
local font_size = G_reader_settings:readSetting("footnote_popup_absolute_font_size")
if font_size then
font_size = Screen:scaleBySize(font_size)
else
font_size = self.doc_font_size + (G_reader_settings:readSetting("footnote_popup_relative_font_size") or -2)
end
-- We want to display the footnote text with the same margins as
-- the document, but keep the scrollbar in the right margin, so
-- both small footnotes (without scrollbar) and longer ones (with
-- scrollbar) have their text aligned with the document text.
-- MuPDF, when rendering some footnotes, may put list item
-- bullets in its own left margin. To get a chance to have them
-- shown, we let MuPDF handle our left margin.
local html_left_margin = self.doc_margins.left .. "px"
local html_right_margin = "0"
if BD.mirroredUILayout() then
html_left_margin, html_right_margin = html_right_margin, html_left_margin
end
local css = T(PAGE_CSS, "0", html_right_margin, "0", html_left_margin, -- top right bottom left
self.font_face, DEFAULT_CSS)
if self.css then -- add any provided css
css = css .. "\n" .. self.css
end
-- require("logger").dbg("CSS:", css)
-- require("logger").dbg("HTML:", self.html)
-- Scrollbar on the right: the document may have quite large margins,
-- and we don't want a scrollbar that large.
-- Ensute a not too large scrollbar, and bring it closer to the screen
-- edge, leaving the remaining available room between it and the text.
local item_width = math.min(math.ceil(self.doc_margins.right * 2/5), Screen:scaleBySize(10))
local scroll_bar_width = item_width
local padding_right = item_width
local text_scroll_span = self.doc_margins.right - scroll_bar_width - padding_right
if text_scroll_span < padding_right then
-- With small doc margins, space between text and scrollbar may get
-- too small: switch it with right padding
text_scroll_span, padding_right = padding_right, text_scroll_span
end
local htmlwidget_width = self.width - padding_right
-- Top and bottom padding: we'd rather not use document margins:
-- they can be large, which makes sense at screen edges, but
-- would be too large for our top padding. Also, MuPDF will often
-- let some blank area at bottom with its line layout and page
-- breaking algorithms, which will visually add to the bottom padding.
local padding_top = Size.padding.large
local padding_bottom = Size.padding.large
local htmlwidget_height = self.height - padding_top - padding_bottom
self.htmlwidget = ScrollHtmlWidget:new{
html_body = self.html,
css = css,
default_font_size = font_size,
width = htmlwidget_width,
height = htmlwidget_height,
scroll_bar_width = scroll_bar_width,
text_scroll_span = text_scroll_span,
dialog = self.dialog,
}
-- We only want a top border, so use a LineWidget for that
local top_border_size = Size.line.thick
local vgroup = VerticalGroup:new{
LineWidget:new{
dimen = Geom:new{
w = self.width,
h = top_border_size,
}
},
VerticalSpan:new{ width = padding_top },
HorizontalGroup:new{
self.htmlwidget,
HorizontalSpan:new{ width = padding_right },
},
VerticalSpan:new{ width = padding_bottom },
}
-- If htmlwidget contains only one page (small footnote content),
-- display only the valuable area: push the blank area down the screen
-- (we can do that because no scroll bar will be displayed)
local single_page_height = self.htmlwidget:getSinglePageHeight() -- nil if multi-pages
if single_page_height then
local added_bottom_pad = 0
-- See if needed:
-- Add a bit to bottom padding, as getSinglePageHeight() cut can be rough
-- added_bottom_pad = math.floor(font_size * 0.2)
local reduced_height = single_page_height + top_border_size + padding_top + padding_bottom + added_bottom_pad
vgroup = CenterContainer:new{
dimen = Geom:new{
h = reduced_height,
w = self.width,
},
ignore = "height",
vgroup,
}
self.height = reduced_height -- for close_callback
end
-- Needed only to set a white background
self.container = FrameContainer:new{
background = Blitbuffer.COLOR_WHITE,
bordersize = 0,
margin = 0,
padding = 0,
vgroup,
}
self[1] = BottomContainer:new{
dimen = Screen:getSize(),
self.container
}
end
function FootnoteWidget:onShow()
UIManager:setDirty(self.dialog, function()
return "ui", self.container.dimen
end)
end
function FootnoteWidget:onCloseWidget()
UIManager:setDirty(self.dialog, function()
return "partial", self.container.dimen
end)
end
function FootnoteWidget:onClose()
UIManager:close(self)
if self.close_callback then
self.close_callback(self.height)
end
return true
end
function FootnoteWidget:onFollow()
if self.follow_callback then
if self.close_callback then
self.close_callback(self.height)
end
return self.follow_callback()
end
end
function FootnoteWidget:onTapClose(arg, ges)
if ges.pos:notIntersectWith(self.container.dimen) then
UIManager:close(self)
-- Allow ReaderLink to check if our dismiss tap
-- was itself on another footnote, and display
-- it. This avoids having to tap 2 times to
-- see another footnote.
if self.on_tap_close_callback then
self.on_tap_close_callback(arg, ges, self.height)
elseif self.close_callback then
self.close_callback(self.height)
end
return true
end
return false
end
function FootnoteWidget:onSwipeFollow(arg, ges)
local direction = BD.flipDirectionIfMirroredUILayout(ges.direction)
if direction == "west" then
if self.follow_callback then
if self.close_callback then
self.close_callback(self.height)
end
return self.follow_callback()
end
elseif direction == "south" or direction == "east" then
UIManager:close(self)
-- We can close with swipe down. If footnote is scrollable,
-- this event will be eaten by ScrollHtmlWidget, and it will
-- work only when started outside the footnote.
-- Also allow closing with swipe east (like we do to go back
-- from link)
if self.close_callback then
self.close_callback(self.height)
end
return true
elseif direction == "north" then
-- no use for now
do end -- luacheck: ignore 541
else -- diagonal swipe
-- trigger full refresh
UIManager:setDirty(nil, "full")
-- a long diagonal swipe may also be used for taking a screenshot,
-- so let it propagate
end
return false
end
return FootnoteWidget