From 74057385bfc3baae81b3e35266596d031a196067 Mon Sep 17 00:00:00 2001 From: Henrik Rydgard Date: Wed, 2 Jan 2013 20:56:09 +0100 Subject: [PATCH] Initial UTF-8 support --- Android.mk | 3 +- gfx_es2/draw_buffer.cpp | 13 +- native.vcxproj | 2 + native.vcxproj.filters | 8 +- ui/ui.cpp | 11 ++ ui/ui.h | 1 + util/text/utf8.cpp | 405 ++++++++++++++++++++++++++++++++++++++++ util/text/utf8.h | 38 ++++ 8 files changed, 478 insertions(+), 3 deletions(-) create mode 100644 util/text/utf8.cpp create mode 100644 util/text/utf8.h diff --git a/Android.mk b/Android.mk index a8588cbc6e..ca569fa5e7 100644 --- a/Android.mk +++ b/Android.mk @@ -54,7 +54,8 @@ LOCAL_SRC_FILES :=\ ui/ui_context.cpp \ ui/screen.cpp \ ui/virtual_input.cpp \ - util/random/perlin.cpp + util/random/perlin.cpp \ + util/text/utf8.cpp LOCAL_CFLAGS := -O2 -DGL_GLEXT_PROTOTYPES -DARM -DUSING_GLES2 -fsigned-char -fno-strict-aliasing LOCAL_CPPFLAGS := -fno-exceptions -fno-rtti -std=gnu++0x diff --git a/gfx_es2/draw_buffer.cpp b/gfx_es2/draw_buffer.cpp index 59db3ec0f2..cbe8309ec8 100644 --- a/gfx_es2/draw_buffer.cpp +++ b/gfx_es2/draw_buffer.cpp @@ -10,6 +10,7 @@ #include "gfx/texture_atlas.h" #include "gfx/gl_debug_log.h" #include "gfx/gl_common.h" +#include "util/text/utf8.h" enum { // Enough? @@ -295,6 +296,12 @@ void DrawBuffer::DoAlign(int flags, float *x, float *y, float *w, float *h) { } } +// TODO: Unicode support. +// U+4E00–U+9FBF Kanji +// U+3040–U+309F Hiragana +// U+30A0–U+30FF Katakana + + // ROTATE_* doesn't yet work right. void DrawBuffer::DrawText(int font, const char *text, float x, float y, Color color, int flags) { const AtlasFont &atlasfont = *atlas->fonts[font]; @@ -312,7 +319,11 @@ void DrawBuffer::DrawText(int font, const char *text, float x, float y, Color co else y += atlasfont.ascend*fontscaley; float sx = x; - while ((cval = *text++) != '\0') { + UTF8 utf(text); + while (true) { + if (utf.end()) + break; + cval = utf.next(); if (cval == '\n') { y += atlasfont.height * fontscaley; x = sx; diff --git a/native.vcxproj b/native.vcxproj index de6389110c..df06267bb1 100644 --- a/native.vcxproj +++ b/native.vcxproj @@ -248,6 +248,7 @@ + @@ -338,6 +339,7 @@ AssemblyAndSourceCode AssemblyAndSourceCode + diff --git a/native.vcxproj.filters b/native.vcxproj.filters index ad1206c6d1..d047cd88a6 100644 --- a/native.vcxproj.filters +++ b/native.vcxproj.filters @@ -233,6 +233,9 @@ gfx + + util + @@ -414,6 +417,9 @@ gfx + + util + @@ -474,4 +480,4 @@ {4515306f-4664-46bf-a89b-abfec5520a15} - + \ No newline at end of file diff --git a/ui/ui.cpp b/ui/ui.cpp index da57aa777b..4c209d88be 100644 --- a/ui/ui.cpp +++ b/ui/ui.cpp @@ -108,6 +108,17 @@ void UIText(int font, int x, int y, const char *text, uint32_t color, float scal ui_draw2d.SetFontScale(1.0f, 1.0f); } +void UIText(int font, const LayoutManager &layout, const char *text, uint32_t color, float scale, int align) +{ + ui_draw2d.SetFontScale(scale, scale); + float w, h; + ui_draw2d.MeasureText(font, text, &w, &h); + float x, y; + layout.GetPos(&w, &h, &x, &y); + UIText(font, x, y, text, color, scale, 0); + ui_draw2d.SetFontScale(1.0f, 1.0f); +} + int UIButton(int id, const LayoutManager &layout, float w, const char *text, int button_align) { float h = themeAtlas->images[theme.buttonImage].h; diff --git a/ui/ui.h b/ui/ui.h index 6fdfdb334b..aad8fa0675 100644 --- a/ui/ui.h +++ b/ui/ui.h @@ -203,6 +203,7 @@ int UIHSlider(int id, int x, int y, int w, int max, int *value); // Draws static text, that does not participate in any focusing scheme etc, it just is. void UIText(int font, int x, int y, const char *text, uint32_t color, float scale = 1.0f, int align = ALIGN_TOPLEFT); void UIText(int x, int y, const char *text, uint32_t color, float scale = 1.0f, int align = ALIGN_TOPLEFT); +void UIText(int font, const LayoutManager &layout, const char *text, uint32_t color, float scale = 1.0f, int align = ALIGN_TOPLEFT); // Slide choice, like the Angry Birds level selector. Not yet working. void UISlideChoice(int id, int y, const SlideItem *items, int numItems, UISlideState *state); diff --git a/util/text/utf8.cpp b/util/text/utf8.cpp new file mode 100644 index 0000000000..10b239ea52 --- /dev/null +++ b/util/text/utf8.cpp @@ -0,0 +1,405 @@ +/* + Basic UTF-8 manipulation routines + by Jeff Bezanson + placed in the public domain Fall 2005 + + This code is designed to provide the utilities you need to manipulate + UTF-8 as an internal string encoding. These functions do not perform the + error checking normally needed when handling UTF-8 data, so if you happen + to be from the Unicode Consortium you will want to flay me alive. + I do this because error checking can be performed at the boundaries (I/O), + with these routines reserved for higher performance on data known to be + valid. +*/ +#include +#include +#include +#include + +#include "base/basictypes.h" +#include "utf8.h" + +// is start of UTF sequence +inline bool isutf(char c) { + return (c & 0xC0) != 0x80; +} + +static const uint32_t offsetsFromUTF8[6] = { + 0x00000000UL, 0x00003080UL, 0x000E2080UL, + 0x03C82080UL, 0xFA082080UL, 0x82082080UL +}; + +static const uint8_t trailingBytesForUTF8[256] = { + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, + 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, + 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,4,4,4,4,5,5,5,5, +}; + +/* returns length of next utf-8 sequence */ +int u8_seqlen(const char *s) +{ + return trailingBytesForUTF8[(unsigned int)(unsigned char)s[0]] + 1; +} + +/* conversions without error checking + only works for valid UTF-8, i.e. no 5- or 6-byte sequences + srcsz = source size in bytes, or -1 if 0-terminated + sz = dest size in # of wide characters + + returns # characters converted + dest will always be L'\0'-terminated, even if there isn't enough room + for all the characters. + if sz = srcsz+1 (i.e. 4*srcsz+4 bytes), there will always be enough space. +*/ +int u8_toucs(uint32_t *dest, int sz, const char *src, int srcsz) +{ + uint32_t ch; + const char *src_end = src + srcsz; + int nb; + int i=0; + + while (i < sz-1) { + nb = trailingBytesForUTF8[(unsigned char)*src]; + if (srcsz == -1) { + if (*src == 0) + goto done_toucs; + } + else { + if (src + nb >= src_end) + goto done_toucs; + } + ch = 0; + switch (nb) { + /* these fall through deliberately */ + case 3: ch += (unsigned char)*src++; ch <<= 6; + case 2: ch += (unsigned char)*src++; ch <<= 6; + case 1: ch += (unsigned char)*src++; ch <<= 6; + case 0: ch += (unsigned char)*src++; + } + ch -= offsetsFromUTF8[nb]; + dest[i++] = ch; + } + done_toucs: + dest[i] = 0; + return i; +} + +/* srcsz = number of source characters, or -1 if 0-terminated + sz = size of dest buffer in bytes + + returns # characters converted + dest will only be '\0'-terminated if there is enough space. this is + for consistency; imagine there are 2 bytes of space left, but the next + character requires 3 bytes. in this case we could NUL-terminate, but in + general we can't when there's insufficient space. therefore this function + only NUL-terminates if all the characters fit, and there's space for + the NUL as well. + the destination string will never be bigger than the source string. +*/ +int u8_toutf8(char *dest, int sz, uint32_t *src, int srcsz) +{ + uint32_t ch; + int i = 0; + char *dest_end = dest + sz; + + while (srcsz<0 ? src[i]!=0 : i < srcsz) { + ch = src[i]; + if (ch < 0x80) { + if (dest >= dest_end) + return i; + *dest++ = (char)ch; + } + else if (ch < 0x800) { + if (dest >= dest_end-1) + return i; + *dest++ = (ch>>6) | 0xC0; + *dest++ = (ch & 0x3F) | 0x80; + } + else if (ch < 0x10000) { + if (dest >= dest_end-2) + return i; + *dest++ = (ch>>12) | 0xE0; + *dest++ = ((ch>>6) & 0x3F) | 0x80; + *dest++ = (ch & 0x3F) | 0x80; + } + else if (ch < 0x110000) { + if (dest >= dest_end-3) + return i; + *dest++ = (ch>>18) | 0xF0; + *dest++ = ((ch>>12) & 0x3F) | 0x80; + *dest++ = ((ch>>6) & 0x3F) | 0x80; + *dest++ = (ch & 0x3F) | 0x80; + } + i++; + } + if (dest < dest_end) + *dest = '\0'; + return i; +} + +int u8_wc_toutf8(char *dest, uint32_t ch) +{ + if (ch < 0x80) { + dest[0] = (char)ch; + return 1; + } + if (ch < 0x800) { + dest[0] = (ch>>6) | 0xC0; + dest[1] = (ch & 0x3F) | 0x80; + return 2; + } + if (ch < 0x10000) { + dest[0] = (ch>>12) | 0xE0; + dest[1] = ((ch>>6) & 0x3F) | 0x80; + dest[2] = (ch & 0x3F) | 0x80; + return 3; + } + if (ch < 0x110000) { + dest[0] = (ch>>18) | 0xF0; + dest[1] = ((ch>>12) & 0x3F) | 0x80; + dest[2] = ((ch>>6) & 0x3F) | 0x80; + dest[3] = (ch & 0x3F) | 0x80; + return 4; + } + return 0; +} + +/* charnum => byte offset */ +int u8_offset(const char *str, int charnum) +{ + int offs=0; + + while (charnum > 0 && str[offs]) { + (void)(isutf(str[++offs]) || isutf(str[++offs]) || + isutf(str[++offs]) || ++offs); + charnum--; + } + return offs; +} + +/* byte offset => charnum */ +int u8_charnum(const char *s, int offset) +{ + int charnum = 0, offs=0; + + while (offs < offset && s[offs]) { + (void)(isutf(s[++offs]) || isutf(s[++offs]) || + isutf(s[++offs]) || ++offs); + charnum++; + } + return charnum; +} + +/* number of characters */ +int u8_strlen(const char *s) +{ + int count = 0; + int i = 0; + + while (u8_nextchar(s, &i) != 0) + count++; + + return count; +} + +/* reads the next utf-8 sequence out of a string, updating an index */ +uint32_t u8_nextchar(const char *s, int *i) +{ + uint32_t ch = 0; + int sz = 0; + + do { + ch <<= 6; + ch += (unsigned char)s[(*i)++]; + sz++; + } while (s[*i] && !isutf(s[*i])); + ch -= offsetsFromUTF8[sz-1]; + + return ch; +} + +void u8_inc(const char *s, int *i) +{ + (void)(isutf(s[++(*i)]) || isutf(s[++(*i)]) || + isutf(s[++(*i)]) || ++(*i)); +} + +void u8_dec(const char *s, int *i) +{ + (void)(isutf(s[--(*i)]) || isutf(s[--(*i)]) || + isutf(s[--(*i)]) || --(*i)); +} + +int octal_digit(char c) +{ + return (c >= '0' && c <= '7'); +} + +int hex_digit(char c) +{ + return ((c >= '0' && c <= '9') || + (c >= 'A' && c <= 'F') || + (c >= 'a' && c <= 'f')); +} + +/* assumes that src points to the character after a backslash + returns number of input characters processed */ +int u8_read_escape_sequence(const char *str, uint32_t *dest) +{ + uint32_t ch; + char digs[9]="\0\0\0\0\0\0\0\0"; + int dno=0, i=1; + + ch = (uint32_t)str[0]; /* take literal character */ + if (str[0] == 'n') + ch = L'\n'; + else if (str[0] == 't') + ch = L'\t'; + else if (str[0] == 'r') + ch = L'\r'; + else if (str[0] == 'b') + ch = L'\b'; + else if (str[0] == 'f') + ch = L'\f'; + else if (str[0] == 'v') + ch = L'\v'; + else if (str[0] == 'a') + ch = L'\a'; + else if (octal_digit(str[0])) { + i = 0; + do { + digs[dno++] = str[i++]; + } while (octal_digit(str[i]) && dno < 3); + ch = strtol(digs, NULL, 8); + } + else if (str[0] == 'x') { + while (hex_digit(str[i]) && dno < 2) { + digs[dno++] = str[i++]; + } + if (dno > 0) + ch = strtol(digs, NULL, 16); + } + else if (str[0] == 'u') { + while (hex_digit(str[i]) && dno < 4) { + digs[dno++] = str[i++]; + } + if (dno > 0) + ch = strtol(digs, NULL, 16); + } + else if (str[0] == 'U') { + while (hex_digit(str[i]) && dno < 8) { + digs[dno++] = str[i++]; + } + if (dno > 0) + ch = strtol(digs, NULL, 16); + } + *dest = ch; + + return i; +} + +/* convert a string with literal \uxxxx or \Uxxxxxxxx characters to UTF-8 + example: u8_unescape(mybuf, 256, "hello\\u220e") + note the double backslash is needed if called on a C string literal */ +int u8_unescape(char *buf, int sz, char *src) +{ + int c=0, amt; + uint32_t ch; + char temp[4]; + + while (*src && c < sz) { + if (*src == '\\') { + src++; + amt = u8_read_escape_sequence(src, &ch); + } + else { + ch = (uint32_t)*src; + amt = 1; + } + src += amt; + amt = u8_wc_toutf8(temp, ch); + if (amt > sz-c) + break; + memcpy(&buf[c], temp, amt); + c += amt; + } + if (c < sz) + buf[c] = '\0'; + return c; +} + +const char *u8_strchr(const char *s, uint32_t ch, int *charn) +{ + int i = 0, lasti=0; + uint32_t c; + + *charn = 0; + while (s[i]) { + c = u8_nextchar(s, &i); + if (c == ch) { + return &s[lasti]; + } + lasti = i; + (*charn)++; + } + return NULL; +} + +const char *u8_memchr(const char *s, uint32_t ch, size_t sz, int *charn) +{ + int i = 0, lasti=0; + uint32_t c; + int csz; + + *charn = 0; + while (i < sz) { + c = csz = 0; + do { + c <<= 6; + c += (unsigned char)s[i++]; + csz++; + } while (i < sz && !isutf(s[i])); + c -= offsetsFromUTF8[csz-1]; + + if (c == ch) { + return &s[lasti]; + } + lasti = i; + (*charn)++; + } + return NULL; +} + +int u8_is_locale_utf8(const char *locale) +{ + /* this code based on libutf8 */ + const char* cp = locale; + + for (; *cp != '\0' && *cp != '@' && *cp != '+' && *cp != ','; cp++) { + if (*cp == '.') { + const char* encoding = ++cp; + for (; *cp != '\0' && *cp != '@' && *cp != '+' && *cp != ','; cp++) + ; + if ((cp-encoding == 5 && !strncmp(encoding, "UTF-8", 5)) + || (cp-encoding == 4 && !strncmp(encoding, "utf8", 4))) + return 1; /* it's UTF-8 */ + break; + } + } + return 0; +} + +bool UTF8StringHasNonASCII(const char *utf8string) { + UTF8 utf(utf8string); + while (!utf.end()) { + int c = utf.next(); + if (c > 127) + return true; + } + return false; +} \ No newline at end of file diff --git a/util/text/utf8.h b/util/text/utf8.h new file mode 100644 index 0000000000..84c2ae8c97 --- /dev/null +++ b/util/text/utf8.h @@ -0,0 +1,38 @@ +/* + Basic UTF-8 manipulation routines + by Jeff Bezanson + placed in the public domain Fall 2005 + + This code is designed to provide the utilities you need to manipulate + UTF-8 as an internal string encoding. These functions do not perform the + error checking normally needed when handling UTF-8 data, so if you happen + to be from the Unicode Consortium you will want to flay me alive. + I do this because error checking can be performed at the boundaries (I/O), + with these routines reserved for higher performance on data known to be + valid. +*/ + +// Further modified, and C++ stuff added, by hrydgard@gmail.com. + +#include "base/basictypes.h" + +uint32_t u8_nextchar(const char *s, int *i); +int u8_strlen(const char *s); + +class UTF8 { +public: + UTF8(const char *c) : c_(c), index_(0) {} + bool end() const { return c_[index_] == 0; } + uint32_t next() { + return u8_nextchar(c_, &index_); + } + int length() const { + return u8_strlen(c_); + } + +private: + const char *c_; + int index_; +}; + +bool UTF8StringHasNonASCII(const char *utf8string); \ No newline at end of file