Software / code / prosody
Changeset
8697:c60fdf148118
util.json: Unescape surrogate pairs
| author | Matthew Wild <mwild1@gmail.com> |
|---|---|
| date | Tue, 27 Mar 2018 13:44:40 +0100 |
| parents | 8696:164da3186511 |
| children | 8698:0499f3da0ec4 |
| files | util/json.lua |
| diffstat | 1 files changed, 5 insertions(+), 4 deletions(-) [+] |
line wrap: on
line diff
--- a/util/json.lua Sun Mar 25 23:30:14 2018 +0200 +++ b/util/json.lua Tue Mar 27 13:44:40 2018 +0100 @@ -246,7 +246,7 @@ end end local _unescape_error; -local function _unescape_surrogate_func(x) -- luacheck: ignore +local function _unescape_surrogate_func(x) local lead, trail = tonumber(x:sub(3, 6), 16), tonumber(x:sub(9, 12), 16); local codepoint = lead * 0x400 + trail - 0x35FDC00; local a = codepoint % 64; @@ -260,8 +260,9 @@ local function _unescape_func(x) x = x:match("%x%x%x%x", 3); if x then - --if x >= 0xD800 and x <= 0xDFFF then _unescape_error = true; end -- bad surrogate pair - return codepoint_to_utf8(tonumber(x, 16)); + local codepoint = tonumber(x, 16) + if codepoint >= 0xD800 and codepoint <= 0xDFFF then _unescape_error = true; end -- bad surrogate pair + return codepoint_to_utf8(codepoint); end _unescape_error = true; end @@ -273,7 +274,7 @@ --if s:find("[%z-\31]") then return nil, "control char in string"; end -- FIXME handle control characters _unescape_error = nil; - --s = s:gsub("\\u[dD][89abAB]%x%x\\u[dD][cdefCDEF]%x%x", _unescape_surrogate_func); + s = s:gsub("\\u[dD][89abAB]%x%x\\u[dD][cdefCDEF]%x%x", _unescape_surrogate_func); -- FIXME handle escapes beyond BMP s = s:gsub("\\u.?.?.?.?", _unescape_func); if _unescape_error then return nil, "invalid escape"; end
