zhangyiss/toml11
1
0
Fork 0
mirror of https://github.com/ToruNiina/toml11.git synced 2025-09-18 18:41:43 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

refactor: simplify branches about utf8 codepoint

Browse Source
This commit is contained in:
ToruNiina
2019-03-15 17:48:47 +09:00
parent 514df99e40
commit 76863cb27f
1 changed files with 4 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
toml/parser.hpp
View File

@@ -262,15 +262,8 @@ std::string read_utf8_codepoint(const region<Container>& reg,
character += static_cast<unsigned char>(0x80|(codepoint >> 6 & 0x3F)); character += static_cast<unsigned char>(0x80|(codepoint >> 6 & 0x3F));
character += static_cast<unsigned char>(0x80|(codepoint & 0x3F)); character += static_cast<unsigned char>(0x80|(codepoint & 0x3F));
} }
else if(codepoint < 0x200000) // U+010000 ... U+1FFFFF else if(codepoint < 0x110000) // U+010000 ... U+10FFFF
{ {
if(0x10FFFF < codepoint) // out of Unicode region
{
throw syntax_error(format_underline("[error] "
"toml::read_utf8_codepoint: input codepoint is too large to "
"decode as a unicode character.", {{std::addressof(loc),
"should be in [0x00..0x10FFFF]"}}));
}
// 11110yyy 10yyxxxx 10xxxxxx 10xxxxxx // 11110yyy 10yyxxxx 10xxxxxx 10xxxxxx
character += static_cast<unsigned char>(0xF0| codepoint >> 18); character += static_cast<unsigned char>(0xF0| codepoint >> 18);
character += static_cast<unsigned char>(0x80|(codepoint >> 12 & 0x3F)); character += static_cast<unsigned char>(0x80|(codepoint >> 12 & 0x3F));
@@ -279,9 +272,9 @@ std::string read_utf8_codepoint(const region<Container>& reg,
} }
else // out of UTF-8 region else // out of UTF-8 region
{ {
throw std::range_error(format_underline(concat_to_string("[error] " throw syntax_error(format_underline("[error] toml::read_utf8_codepoint:"
"input codepoint (", str, ") is too large to encode as utf-8."), " input codepoint is too large.",
{{std::addressof(reg), "should be in [0x00..0x10FFFF]"}})); {{std::addressof(loc), "should be in [0x00..0x10FFFF]"}}));
} }
return character; return character;
} }