other-licenses/7zstub/src/Common/UTFConvert.cpp@6474c204b198 (annotated)
other-licenses/7zstub/src/Common/UTFConvert.cpp
Wed, 31 Dec 2014 06:09:35 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 06:09:35 +0100
- changeset 0
- 6474c204b198
- permissions
- -rw-r--r--
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
| michael@0 | 1 | // UTFConvert.cpp |
| michael@0 | 2 | |
| michael@0 | 3 | #include "StdAfx.h" |
| michael@0 | 4 | |
| michael@0 | 5 | #include "UTFConvert.h" |
| michael@0 | 6 | #include "Types.h" |
| michael@0 | 7 | |
| michael@0 | 8 | static Byte kUtf8Limits[5] = { 0xC0, 0xE0, 0xF0, 0xF8, 0xFC }; |
| michael@0 | 9 | |
| michael@0 | 10 | // These functions are for UTF8 <-> UTF16 conversion. |
| michael@0 | 11 | |
| michael@0 | 12 | bool ConvertUTF8ToUnicode(const AString &src, UString &dest) |
| michael@0 | 13 | { |
| michael@0 | 14 | dest.Empty(); |
| michael@0 | 15 | for(int i = 0; i < src.Length();) |
| michael@0 | 16 | { |
| michael@0 | 17 | Byte c = (Byte)src[i++]; |
| michael@0 | 18 | if (c < 0x80) |
| michael@0 | 19 | { |
| michael@0 | 20 | dest += (wchar_t)c; |
| michael@0 | 21 | continue; |
| michael@0 | 22 | } |
| michael@0 | 23 | if(c < 0xC0) |
| michael@0 | 24 | return false; |
| michael@0 | 25 | int numAdds; |
| michael@0 | 26 | for (numAdds = 1; numAdds < 5; numAdds++) |
| michael@0 | 27 | if (c < kUtf8Limits[numAdds]) |
| michael@0 | 28 | break; |
| michael@0 | 29 | UInt32 value = (c - kUtf8Limits[numAdds - 1]); |
| michael@0 | 30 | do |
| michael@0 | 31 | { |
| michael@0 | 32 | if (i >= src.Length()) |
| michael@0 | 33 | return false; |
| michael@0 | 34 | Byte c2 = (Byte)src[i++]; |
| michael@0 | 35 | if (c2 < 0x80 || c2 >= 0xC0) |
| michael@0 | 36 | return false; |
| michael@0 | 37 | value <<= 6; |
| michael@0 | 38 | value |= (c2 - 0x80); |
| michael@0 | 39 | numAdds--; |
| michael@0 | 40 | } |
| michael@0 | 41 | while(numAdds > 0); |
| michael@0 | 42 | if (value < 0x10000) |
| michael@0 | 43 | dest += (wchar_t)(value); |
| michael@0 | 44 | else |
| michael@0 | 45 | { |
| michael@0 | 46 | value -= 0x10000; |
| michael@0 | 47 | if (value >= 0x100000) |
| michael@0 | 48 | return false; |
| michael@0 | 49 | dest += (wchar_t)(0xD800 + (value >> 10)); |
| michael@0 | 50 | dest += (wchar_t)(0xDC00 + (value & 0x3FF)); |
| michael@0 | 51 | } |
| michael@0 | 52 | } |
| michael@0 | 53 | return true; |
| michael@0 | 54 | } |
| michael@0 | 55 | |
| michael@0 | 56 | bool ConvertUnicodeToUTF8(const UString &src, AString &dest) |
| michael@0 | 57 | { |
| michael@0 | 58 | dest.Empty(); |
| michael@0 | 59 | for(int i = 0; i < src.Length();) |
| michael@0 | 60 | { |
| michael@0 | 61 | UInt32 value = (UInt32)src[i++]; |
| michael@0 | 62 | if (value < 0x80) |
| michael@0 | 63 | { |
| michael@0 | 64 | dest += (char)value; |
| michael@0 | 65 | continue; |
| michael@0 | 66 | } |
| michael@0 | 67 | if (value >= 0xD800 && value < 0xE000) |
| michael@0 | 68 | { |
| michael@0 | 69 | if (value >= 0xDC00) |
| michael@0 | 70 | return false; |
| michael@0 | 71 | if (i >= src.Length()) |
| michael@0 | 72 | return false; |
| michael@0 | 73 | UInt32 c2 = (UInt32)src[i++]; |
| michael@0 | 74 | if (c2 < 0xDC00 || c2 >= 0xE000) |
| michael@0 | 75 | return false; |
| michael@0 | 76 | value = ((value - 0xD800) << 10) | (c2 - 0xDC00); |
| michael@0 | 77 | } |
| michael@0 | 78 | int numAdds; |
| michael@0 | 79 | for (numAdds = 1; numAdds < 5; numAdds++) |
| michael@0 | 80 | if (value < (((UInt32)1) << (numAdds * 5 + 6))) |
| michael@0 | 81 | break; |
| michael@0 | 82 | dest += (char)(kUtf8Limits[numAdds - 1] + (value >> (6 * numAdds))); |
| michael@0 | 83 | do |
| michael@0 | 84 | { |
| michael@0 | 85 | numAdds--; |
| michael@0 | 86 | dest += (char)(0x80 + ((value >> (6 * numAdds)) & 0x3F)); |
| michael@0 | 87 | } |
| michael@0 | 88 | while(numAdds > 0); |
| michael@0 | 89 | } |
| michael@0 | 90 | return true; |
| michael@0 | 91 | } |
