apio/Luna
1
1
Fork 0
Code Issues Pull Requests Actions Projects Releases 8 Wiki Activity

luna: Make Utf8String{De,En}coders return the number of bytes written

Browse Source 
This means we can avoid a call to code_points() in mbstowcs(),
which would parse a string twice.
This commit is contained in:
apio 2023-01-14 11:55:19 +01:00
parent da805eec83
commit 00ee8314b3
Signed by: apio
GPG Key ID: B8A7D06E42258954
3 changed files with 16 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
libc/src/stdlib.cpp
View File

@ -104,22 +104,15 @@ extern "C"
__builtin_unreachable();
}
// FIXME: This is walking a UTF-8 string twice. Once to decode, and another to count code points.
size_t mbstowcs(wchar_t* buf, const char* src, size_t max)
{
if (max == 0) return 0;
Utf8StringDecoder decoder(src);
auto rc = decoder.decode(buf, max);
if (!buf) { return decoder.code_points().value_or((size_t)-1); }
if (rc.has_error()) return (size_t)-1;
size_t code_points = decoder.code_points().value_or(0);
if (code_points >= max) return max - 1;
return code_points;
return decoder.decode(buf, max).value_or((size_t)-1);
}
void* malloc(size_t size)

8
luna/include/luna/Utf8.h
View File

@ -16,9 +16,9 @@ class Utf8StringDecoder
Result<usize> code_points() const;
// The caller must ensure that 'buf' is at least code_points() + a NULL wide.
Result<void> decode(wchar_t* buf) const;
Result<usize> decode(wchar_t* buf) const;
Result<void> decode(wchar_t* buf, usize max) const;
Result<usize> decode(wchar_t* buf, usize max) const;
private:
const char* m_str;
@ -38,9 +38,9 @@ class Utf8StringEncoder
Result<usize> byte_length() const;
// The caller must ensure that 'buf' is at least byte_length() + a NULL wide.
Result<void> encode(char* buf) const;
Result<usize> encode(char* buf) const;
Result<void> encode(char* buf, usize max) const;
Result<usize> encode(char* buf, usize max) const;
private:
const wchar_t* m_str;

16
luna/src/Utf8.cpp
View File

@ -2,6 +2,8 @@
#include <luna/CString.h>
#include <luna/Utf8.h>
// FIXME: Not enough space for a sequence is not an error. (mbstowcs(3) and wcstombs(3), case 2 when buf is not NULL)
static_assert(WCHAR_MAX > 0x10ffff);
static Result<usize> utf8_char_length(char c)
@ -146,9 +148,10 @@ Result<usize> Utf8StringDecoder::code_points() const
return len;
}
Result<void> Utf8StringDecoder::decode(wchar_t* buf, usize max) const
Result<usize> Utf8StringDecoder::decode(wchar_t* buf, usize max) const
{
const char* it = m_str;
wchar_t* const buf_start = buf;
while ((usize)(it - m_str) < m_byte_length && max--)
{
@ -160,10 +163,10 @@ Result<void> Utf8StringDecoder::decode(wchar_t* buf, usize max) const
*buf = 0;
return {};
return (usize)(buf - buf_start);
}
Result<void> Utf8StringDecoder::decode(wchar_t* buf) const
Result<usize> Utf8StringDecoder::decode(wchar_t* buf) const
{
return decode(buf, (usize)-1);
}
@ -186,9 +189,10 @@ Result<usize> Utf8StringEncoder::byte_length() const
return len;
}
Result<void> Utf8StringEncoder::encode(char* buf, usize max) const
Result<usize> Utf8StringEncoder::encode(char* buf, usize max) const
{
const wchar_t* it = m_str;
char* const buf_start = buf;
while (*it && max > 1)
{
@ -201,10 +205,10 @@ Result<void> Utf8StringEncoder::encode(char* buf, usize max) const
*buf = 0;
return {};
return (usize)(buf - buf_start);
}
Result<void> Utf8StringEncoder::encode(char* buf) const
Result<usize> Utf8StringEncoder::encode(char* buf) const
{
return encode(buf, (usize)-1);
}