vieter-v
/
v
2
0
Fork 0
Code Issues Pull Requests Releases Activity

removed bits C code inside utf8 builtin

pull/691/head
Henrixounez 2019-06-25 17:22:01 +02:00 committed by Alexander Medvednikov
parent d24be8cf6c
commit f6a401aa02
1 changed files with 53 additions and 63 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

116
builtin/utf8.v
View File

@ -251,71 +251,63 @@ fn (s string) runes() []string {
// Convert utf32 to utf8 // Convert utf32 to utf8
// utf32 == Codepoint // utf32 == Codepoint
fn utf32_to_str(code u32) string { fn utf32_to_str(code u32) string {
// println('code = $code') icode := int(code) //Prevents doing casts everywhere
buffer := malloc(5) mut buffer := malloc(5)
# if (code <= 0x7F) { if icode <= 127 /* 0x7F */ {
// println('!!!!!!!1') buffer[0] = icode
# buffer[0] = code; return tos(buffer, 1)
# return tos(buffer, 1); }
# } if (icode <= 2047 /* 0x7FF */) {
# if (code <= 0x7FF) { buffer[0] = 192 /*0xC0*/ | (icode >> 6) /* 110xxxxx */
// println('!!!!!!!2') buffer[1] = 128 /*0x80*/ | (icode & 63 /*0x3F*/) /* 10xxxxxx */
# buffer[0] = 0xC0 | (code >> 6); /* 110xxxxx */ return tos(buffer, 2)
# buffer[1] = 0x80 | (code & 0x3F); /* 10xxxxxx */ }
# return tos(buffer, 2); if (icode <= 65535 /* 0xFFFF */) {
# } buffer[0] = 224 /*0xE0*/ | (icode >> 12) /* 1110xxxx */
# if (code <= 0xFFFF) { buffer[1] = 128 /*0x80*/ | ((icode >> 6) & 63 /*0x3F*/) /* 10xxxxxx */
// println('!!!!!!!3') buffer[2] = 128 /*0x80*/ | (icode & 63 /*0x3F*/) /* 10xxxxxx */
# buffer[0] = 0xE0 | (code >> 12); /* 1110xxxx */ return tos(buffer, 3)
# buffer[1] = 0x80 | ((code >> 6) & 0x3F); /* 10xxxxxx */ }
# buffer[2] = 0x80 | (code & 0x3F); /* 10xxxxxx */ if (icode <= 1114111 /* 0x10FFFF */) {
# return tos(buffer, 3); buffer[0] = 240 /*0xF0*/ | (icode >> 18) /* 11110xxx */
# } buffer[1] = 128 /*0x80*/ | ((icode >> 12) & 63 /*0x3F*/) /* 10xxxxxx */
# if (code <= 0x10FFFF) { buffer[2] = 128 /*0x80*/ | ((icode >> 6) & 63 /*0x3F*/) /* 10xxxxxx */
# buffer[0] = 0xF0 | (code >> 18); /* 11110xxx */ buffer[3] = 128 /*0x80*/ | (icode & 63 /*0x3F*/) /* 10xxxxxx */
# buffer[1] = 0x80 | ((code >> 12) & 0x3F); /* 10xxxxxx */ return tos(buffer, 4)
# buffer[2] = 0x80 | ((code >> 6) & 0x3F); /* 10xxxxxx */ }
# buffer[3] = 0x80 | (code & 0x3F); /* 10xxxxxx */
# return tos(buffer, 4);
# }
return '' return ''
} }
// TODO copypasta // TODO copypasta
fn utf32_to_str_no_malloc(code u32, buf voidptr) string { fn utf32_to_str_no_malloc(code u32, buffer byteptr) string {
// println('code = $code') icode := int(code) //Prevents doing casts everywhere
# char* buffer = buf; if icode <= 127 /* 0x7F */ {
# if (code <= 0x7F) { buffer[0] = icode
// println('!!!!!!!1') return tos(buffer, 1)
# buffer[0] = code; }
# return tos(buffer, 1); if (icode <= 2047 /* 0x7FF */) {
# } buffer[0] = 192 /*0xC0*/ | (icode >> 6) /* 110xxxxx */
# if (code <= 0x7FF) { buffer[1] = 128 /*0x80*/ | (icode & 63 /*0x3F*/) /* 10xxxxxx */
// println('!!!!!!!2') return tos(buffer, 2)
# buffer[0] = 0xC0 | (code >> 6); /* 110xxxxx */ }
# buffer[1] = 0x80 | (code & 0x3F); /* 10xxxxxx */ if (icode <= 65535 /* 0xFFFF */) {
# return tos(buffer, 2); buffer[0] = 224 /*0xE0*/ | (icode >> 12) /* 1110xxxx */
# } buffer[1] = 128 /*0x80*/ | ((icode >> 6) & 63 /*0x3F*/) /* 10xxxxxx */
# if (code <= 0xFFFF) { buffer[2] = 128 /*0x80*/ | (icode & 63 /*0x3F*/) /* 10xxxxxx */
// println('!!!!!!!3') return tos(buffer, 3)
# buffer[0] = 0xE0 | (code >> 12); /* 1110xxxx */ }
# buffer[1] = 0x80 | ((code >> 6) & 0x3F); /* 10xxxxxx */ if (icode <= 1114111 /* 0x10FFFF */) {
# buffer[2] = 0x80 | (code & 0x3F); /* 10xxxxxx */ buffer[0] = 240 /*0xF0*/ | (icode >> 18) /* 11110xxx */
# return tos(buffer, 3); buffer[1] = 128 /*0x80*/ | ((icode >> 12) & 63 /*0x3F*/) /* 10xxxxxx */
# } buffer[2] = 128 /*0x80*/ | ((icode >> 6) & 63 /*0x3F*/) /* 10xxxxxx */
# if (code <= 0x10FFFF) { buffer[3] = 128 /*0x80*/ | (icode & 63 /*0x3F*/) /* 10xxxxxx */
# buffer[0] = 0xF0 | (code >> 18); /* 11110xxx */ return tos(buffer, 4)
# buffer[1] = 0x80 | ((code >> 12) & 0x3F); /* 10xxxxxx */ }
# buffer[2] = 0x80 | ((code >> 6) & 0x3F); /* 10xxxxxx */
# buffer[3] = 0x80 | (code & 0x3F); /* 10xxxxxx */
# return tos(buffer, 4);
# }
return '' return ''
} }
// Convert utf8 to utf32 // Convert utf8 to utf32
fn (_rune string) utf32_code() int { fn (_rune string) utf32_code() int {
// println('utf 32 of $rune len=$rune.len')
if _rune.len == 0 { if _rune.len == 0 {
return 0 return 0
} }
@ -323,20 +315,18 @@ fn (_rune string) utf32_code() int {
if _rune.len == 1 { if _rune.len == 1 {
return int(_rune[0]) return int(_rune[0])
} }
b := byte(int(_rune[0])) mut b := byte(int(_rune[0]))
// TODO should be // TODO should be
// res := int( rune[0] << rune.len) // res := int( rune[0] << rune.len)
# b <<= _rune.len; b <<= _rune.len
res := int(b) mut res := int(b)
mut shift := 6 - _rune.len mut shift := 6 - _rune.len
for i := 1; i < _rune.len; i++ { for i := 1; i < _rune.len; i++ {
// println('c=$res')
c := int(_rune[i]) c := int(_rune[i])
# res <<= shift; res <<= shift
# res |= c & 0x3f; res |= c & 63 /* 0x3f */
shift = 6 shift = 6
} }
// println('!!!!!!!! utf32 $rune res = $res')
return res return res
} }