renun/koreader
1
0
Fork 0
mirror of https://github.com/koreader/koreader.git synced 2025-08-10 00:52:38 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Update master

Browse Source
This commit is contained in:
NuPogodi
2012-07-06 22:16:31 +03:00
committed by Qingping Hou
parent 333537dfd6
commit c06605e86c
1 changed files with 35 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

49
screen.lua
View File

@@ -107,7 +107,17 @@ function Screen:restoreFromBB(bb)
end
end
-- NuPogodi (02.07.2012): functions to save the fb-content as common graphic files - bmp & pgm.
function Screen:screenshot()
lfs.mkdir("./screenshots")
local start = os.clock()
--showInfoMsgWithDelay("making screenshot... ", 1000, 1)
self:fb2bmp("/dev/fb0", lfs.currentdir().."/screenshots/"..os.date("%Y%m%d%H%M%S")..".bmp", true, "bzip2 ")
--self:fb2pgm("/dev/fb0", lfs.currentdir().."/screenshots/"..os.date("%Y%m%d%H%M%S")..".pgm", "bzip2 ", 4)
showInfoMsgWithDelay(string.format("Screenshot is ready in %.2f(s) ", os.clock()-start), 1000, 1)
end
-- NuPogodi (02.07.2012): added the functions to save the fb-content in common graphic files - bmp & pgm.
-- ToDo: png, gif ?
function Screen:LE(x) -- converts positive upto 32bit-number to a little-endian for bmp-header
@@ -126,10 +136,12 @@ end
--[[ This function saves the 4bpp framebuffer as 4bpp BMP and, if necessary, packes the output by command os.execute(pack..fn).
Since framebuffer enumerates the lines from top to bottom and the bmp-file does it in the inversed order, the process includes
a vertical flip that makes it a bit slower - ~0.16(s) @ Kindle3 & Kindle2.
a vertical flip that makes it a bit slower, namely,
~0.16(s) @ Kindle3 & Kindle2 (600x800) ~0.02s @ without v-flip
~0.36(s) @ Kindle DX (824x1200)
NB: needs free memory of G_width*G_height/2 bytes to manupulate the fb-content! ]]
function Screen:fb2bmp(fin, fout, pack) -- atm, for 4bpp framebuffers only
function Screen:fb2bmp(fin, fout, vflip, pack) -- atm, for 4bpp framebuffers only
local inputf = assert(io.open(fin,"rb"))
if inputf then
local outputf, size = assert(io.open(fout,"wb"))
@@ -137,7 +149,7 @@ function Screen:fb2bmp(fin, fout, pack) -- atm, for 4bpp framebuffers only
outputf:write(string.char(0x42,0x4D,0xF6,0xA9,3,0,0,0,0,0,0x76,0,0,0,40,0),
self:LE(G_width), self:LE(G_height), -- width & height: 4 chars each
string.char(0,0,1,0,4,0,0,0,0,0),
self:LE(G_width*G_height/2), -- raw bytes in image
self:LE(G_height*G_width/2), -- raw bytes in image
string.char(0x87,0x19,0,0,0x87,0x19,0,0), -- 6536 pixel/m = 166 dpi for both x&y resolutions
string.char(16,0,0,0,0,0,0,0)) -- 16 colors
local line, i = G_width/2, 15
@@ -146,23 +158,32 @@ function Screen:fb2bmp(fin, fout, pack) -- atm, for 4bpp framebuffers only
outputf:write(string.char(i*16+i):rep(3), string.char(0))
i=i-1
end
-- read the fb-content line-by-line & fill the content-table in the inversed order
local content = {}
for i=1, G_height do
table.insert(content, 1, inputf:read(line))
end
-- write the v-flipped bmp-data
for i=1, G_height do
outputf:write(content[i])
if vflip then -- flip image vertically to make it bmp-compliant
-- read the fb-content line-by-line & fill the content-table in the inversed order
local content = {}
for i=1, G_height do
table.insert(content, 1, inputf:read(line))
end
-- write the v-flipped bmp-data
for i=1, G_height do
outputf:write(content[i])
end
else -- without v-flip, it takes only 0.02s @ 600x800, 4bpp
outputf:write(inputf:read("*all"))
end
inputf:close()
outputf:close()
-- here one may use either standard archivers (bzip2, gzip)
-- or standalone converters (bmp2png, bmp2gif)
if pack then os.execute(pack..fout) end
end
end
--[[ This function saves the fb-content (both 4bpp and 8bpp) as 8bpp PGM and pack it. It's relatively slow for 4bpp devices
(~2.5s on Kindle3, 600x800, 4bpp), but should be extremely fast (<0.1s) when no color conversion (4bpp>8bpp) is needed. ]]
--[[ This function saves the fb-content (both 4bpp and 8bpp) as 8bpp PGM and pack it.
It's relatively slow for 4bpp devices such as
~2.5s @ K2 and K3 > 600x800, 4bpp
~5.0s @ KDX > 824x1200,
but should be very fast (<<0.1s) when no color conversion (4bpp>8bpp) is needed. ]]
function Screen:fb2pgm(fin, fout, pack, bpp)
local inputf = assert(io.open(fin,"rb"))