diff options
| author | Alexander Yakushev <yakushev.alex@gmail.com> | 2011-10-23 02:03:05 +0300 |
|---|---|---|
| committer | Alexander Yakushev <yakushev.alex@gmail.com> | 2011-10-23 02:03:05 +0300 |
| commit | 25a1b198ae7ad94b020e023b8dc9bfd3e3557785 (patch) | |
| tree | 9f115fff1c80b7daade5f570b2febad26331de35 /jamendo.lua | |
| parent | 445e63ee42c5cee54a930139e71d48ce52745b2f (diff) | |
| download | awesompd-25a1b198ae7ad94b020e023b8dc9bfd3e3557785.tar.gz awesompd-25a1b198ae7ad94b020e023b8dc9bfd3e3557785.tar.bz2 | |
Improve UTF symbol transformation from Jamendo responses
Awesompd was working with 1 or 2-byte symbols (up to \u7ff). This commit fixes that, allowing UTF symbols up to 4 bytes.
Also a change was made to be able to replace on search box with other. It is useful when search crashes (it still happens sometimes) and the searchbox is just hanging there.
Diffstat (limited to 'jamendo.lua')
| -rw-r--r-- | jamendo.lua | 46 |
1 files changed, 35 insertions, 11 deletions
diff --git a/jamendo.lua b/jamendo.lua index 5af6121..57eccb5 100644 --- a/jamendo.lua +++ b/jamendo.lua @@ -307,20 +307,44 @@ end -- Jamendo returns Unicode symbols as \uXXXX. Lua does not transform -- them into symbols so we need to do it ourselves. function utf8_codes_to_symbols (s) +-- print(utf8_codes_to_symbols, s) local hexnums = "[%dabcdefABCDEF]" - local pattern = string.format("\\u(%s%s%s%s?)", - hexnums, hexnums, hexnums, hexnums) + local pattern = string.format("\\u(%s%s%s%s?%s?)", + hexnums, hexnums, hexnums, hexnums, hexnums) local decode = function(code) + print("Look at me! I parse " .. code) code = tonumber(code, 16) - -- Grab high and low byte - local hi = math.floor(code / 256) * 4 + 192 - local lo = math.mod(code, 256) - -- Reduce low byte to 64, add overflow to high - local oflow = math.floor(lo / 64) - hi = hi + oflow - lo = math.mod(code, 64) + 128 - -- Return symbol as \hi\lo - return string.char(hi, lo) + if code < 128 then -- one-byte symbol + return string.char(code) + elseif code < 2048 then -- two-byte symbol + -- Grab high and low bytes + local hi = math.floor(code / 64) + local lo = math.mod(code, 64) + -- Return symbol as \hi\lo + return string.char(hi + 192, lo + 128) + elseif code < 65536 then + -- Grab high, middle and low bytes + local hi = math.floor(code / 4096) + local leftover = code - hi * 4096 + local mi = math.floor(leftover / 64) + leftover = leftover - mi * 64 + local lo = math.mod(leftover, 64) + -- Return symbol as \hi\mi\lo + return string.char(hi + 224, mi + 160, lo + 128) + elseif code < 1114112 then + print("I am actually here") + -- Grab high, highmiddle, lowmiddle and low bytes + local hi = math.floor(code / 262144) + local leftover = code - hi * 262144 + local hm = math.floor(leftover / 4096) + leftover = leftover - hm * 4096 + local lm = math.floor(leftover / 64) + local lo = math.mod(leftover, 64) + -- Return symbol as \hi\hm\lm\lo + return string.char(hi + 240, hm + 128, lm + 128, lo + 128) + else -- It is not Unicode symbol at all + return tostring(code) + end end return string.gsub(s, pattern, decode) end |