/**
\page unicode Unicode and UTF-8 Support
FLTK provides comprehensive Unicode support through UTF-8 encoding, allowing
your applications to handle international text and be easily localized for
users worldwide.
\section unicode_overview Overview
Starting with version 1.3.0, FLTK uses UTF-8 as its primary text encoding. This means:
- All text in FLTK is expected to be UTF-8 encoded
- Your application can display text in any language
- File operations work correctly with international filenames
- Most existing ASCII code continues work unchanged
\note Unicode support in FLTK is functional but still evolving. Some advanced
features like bidirectional text and complex script shaping
are not yet implemented.
\section unicode_quick_start Quick Start
International code test for HTML and PDF:
\code_international{"unicode_about.png"}
// This is a test
// 日本語テストテキスト
// 中文测试文本
// Ελληνικό κείμενο δοκιμής
\endcode_international
For most applications, you simply need to ensure your text is UTF-8 encoded:
\code
// These all work automatically with UTF-8:
Fl_Window window(400, 300, "Hallo Römer"); // Mixed ASCII and German
button->label("Café"); // Accented characters
fl_fopen("pièce.txt", "r");
\endcode
\note FLTK supports most Unicode character sets, including Chinese,
Cyrillic, Greek, and many more. However the Doxygen-to-pdf pipeline does not
easily support those. So I limited this document to German and French examples.
\section unicode_background What is Unicode and UTF-8?
__Unicode__ is a standard that assigns a unique number to every character used
in human languages - from Latin letters to Chinese characters to emoji. Each
character has a "code point" like U+0041 for 'A' or U+00DF for 'ß'.
Note that some characters may be composed of multiple consecutive codepoints.
FLTK supports only a subset of the more complex Unicode concepts.
__UTF-8__ is a way to store Unicode codepoints as a sequence of bytes,
called an octet. UTF-8 is backward-compatible with ASCII and efficient
for most text:
- ASCII characters (like 'A') use 1 byte
- European accented characters use 2 bytes
- Most other characters (Chinese, Arabic, etc.) use 3 bytes
- Rare characters and emoji may use 4 bytes
FLTK chose UTF-8 because it works well with existing C string functions and
doesn't break legacy ASCII code.
\section unicode_functions Unicode Functions in FLTK
\subsection unicode_validation Text Validation and Analysis
Functions to check and analyze UTF-8 text:
fl_utf8test() - Check if a string contains valid UTF-8
\code_international{"unicode_hello.png"}
const char* text = "Hello 世界";
int result = fl_utf8test(text, strlen(text));
// Returns: 0=invalid, 1=ASCII, 2=2-byte chars, 3=3-byte chars, 4=4-byte chars
\endcode_international
fl_utf8len() - Get the length in bytes of a UTF-8 octet
\code
char ch = '\xE4'; // First byte of a 3-byte UTF-8 sequence
int len = fl_utf8len(ch); // Returns 3 (or -1 if invalid)
\endcode
fl_utf8locale() - Check if system uses UTF-8 encoding
\code
if (fl_utf8locale()) {
// System uses UTF-8, no conversion needed
} else {
// May need to convert from local encoding
}
\endcode
fl_utf_nb_char() - Count Unicode codepoints by decoding UTF-8 octets
\code
const char* text = "Téléphone";
int char_count = fl_utf_nb_char((unsigned char*)text, strlen(text));
// Returns 9 (number of characters, not bytes)
\endcode
fl_utf8bytes() / fl_unichar_to_utf8_size() - Get number of bytes needed
to create a UTF-8 octet for a Unicode codepoint
\code
unsigned int unicode_char = 0x4E2D; // Chinese character 'zhong'
int bytes_needed = fl_utf8bytes(unicode_char); // Returns 3
\endcode
fl_nonspacing() - Check if codepoint is non-spacing (combining character)
Non-spacing codepoints typically combine this codepoint with the folowing
codepoint into a single glyph. Don't line-break after a nonspacing codepoint.
A more detailed description and links to the Unicode standard can be found
in the fl_nonspacing() docs.
\subsection unicode_conversion Text Conversion
Functions to convert between encodings:
fl_utf8decode() / fl_utf8encode() - Convert one UTF-8 octet into a Unicode
codepoint or a Unicode codepoint into a UTF-8 octet.
\code
// Decode UTF-8 to Unicode code point
const char* utf8_char = "中";
int len;
unsigned int unicode = fl_utf8decode(utf8_char, utf8_char + 3, &len);
// unicode = 0x4E2D, len = 3
// Encode Unicode codepoint to UTF-8
char buffer[5];
int bytes = fl_utf8encode(0x4E2D, buffer); // Returns 3
buffer[bytes] = '\0'; // Now buffer contains "中"
\endcode
fl_utf8froma() / fl_utf8toa() - Convert between UTF-8 and single-byte
ISO-8859-1 encoding
\code
// Convert ISO-8859-1 to UTF-8
char utf8_buffer[200];
fl_utf8froma(utf8_buffer, sizeof(utf8_buffer), "café", 4);
// Convert UTF-8 to ISO-8859-1 (non-representable chars become '?')
char text_buffer[100];
fl_utf8toa("café", 5, text_buffer, sizeof(text_buffer));
\endcode
fl_utf8fromwc() / fl_utf8towc() - Convert between UTF-8 and wide characters
\code
// Convert wide string to UTF-8
wchar_t wide_text[] = L"Hello 世界";
char utf8_buffer[100];
fl_utf8fromwc(utf8_buffer, sizeof(utf8_buffer), wide_text, wcslen(wide_text));
// Convert UTF-8 to wide string
const char* utf8_text = "Hello 世界";
wchar_t wide_buffer[50];
fl_utf8towc(utf8_text, strlen(utf8_text), wide_buffer, 50);
\endcode
fl_utf8toUtf16() - Convert UTF-8 to UTF-16
\code
const char* utf8_text = "Hello 世界";
unsigned short utf16_buffer[100];
unsigned int result = fl_utf8toUtf16(utf8_text, strlen(utf8_text),
utf16_buffer, 100);
// Converts to UTF-16, handling surrogate pairs on Windows
\endcode
fl_utf2mbcs() - Convert UTF-8 to local multibyte encoding
\code
const char* utf8_text = "Hello 世界";
char* local_text = fl_utf2mbcs(utf8_text);
// Converts to system's local encoding (Windows CP, etc.)
// Remember to free the returned pointer
free(local_text);
\endcode
fl_utf8from_mb() / fl_utf8to_mb() - Convert between UTF-8 and local multibyte
\code
// Convert from local multibyte to UTF-8
char utf8_buffer[200];
fl_utf8from_mb(utf8_buffer, sizeof(utf8_buffer), local_text, strlen(local_text));
// Convert from UTF-8 to local multibyte
char local_buffer[200];
fl_utf8to_mb(utf8_text, strlen(utf8_text), local_buffer, sizeof(local_buffer));
\endcode
\subsection unicode_navigation Text Navigation
Functions to move through UTF-8 text safely:
fl_utf8back() / fl_utf8fwd() - Find codepoint boundaries
\code
const char* text = "Café";
const char* start = text;
const char* end = text + strlen(text);
const char* e_pos = text + 3; // Points to 'é'
// Move to previous codepoint
const char* c_pos = fl_utf8back(e_pos, start, end); // Points to 'f'
// Move to next codepoint
const char* next_pos = fl_utf8fwd(e_pos, start, end); // Points after 'é'
\endcode
\subsection unicode_string_ops String Operations
UTF-8 aware string functions:
fl_utf8strlen() - Count Unicode codepoints in a UTF-8 encoded text
\code
const char* text = "Café"; // 5 bytes, 4 characters
int chars = fl_utf8strlen(text); // Returns 4
int bytes = strlen(text); // Returns 5
\endcode
fl_utf_strcasecmp() / fl_utf_strncasecmp() - Compare strings ignoring case.
Please note the list of limitations in FLTK's Unicode support at the bottom of
this page.
\code
int result = fl_utf_strcasecmp("Café", "CAFÉ"); // Returns 0 (equal)
int result2 = fl_utf_strncasecmp("Café", "CAFÉ", 2); // Compare first 2 chars
\endcode
fl_tolower() / fl_toupper() - Convert case for individual Unicode characters
\code
unsigned int lower_a = fl_tolower(0x41); // 'A' -> 'a' (0x61)
unsigned int upper_e = fl_toupper(0xE9); // 'é' -> 'É' (0xC9)
\endcode
fl_utf_tolower() / fl_utf_toupper() - Convert case for UTF-8 strings
\code
const char* text = "Café";
char lower_buffer[20];
fl_utf_tolower((unsigned char*)text, strlen(text), lower_buffer);
// lower_buffer now contains "café"
\endcode
\subsection unicode_file_ops File Operations
Cross-platform file functions that handle UTF-8 filenames:
__Basic file operations:__
\code
// These work with international filenames on all platforms:
FILE* f = fl_fopen("sœur.txt", "r"); // Open file
int fd = fl_open("gummibär.bin", O_RDONLY); // Open with file descriptor
int result = fl_stat("français.dat", &stat_buf); // Get file info
\endcode
__File access and properties:__
\code
fl_access("château.txt", R_OK); // Check if file is readable
fl_chmod("fête.dat", 0644); // Change file permissions
fl_unlink("Küche.tmp"); // Delete file
fl_rename("Äpfel.txt", "Birnen.txt"); // Rename file
\endcode
__Directory operations:__
\code
fl_mkdir("Straßen", 0755); // Create directory
fl_rmdir("éléphant"); // Remove directory
char current_dir[1024];
fl_getcwd(current_dir, sizeof(current_dir)); // Get current directory
\endcode
__Path operations:__
\code
fl_make_path("animaux/éléphant"); // Create directory path
fl_make_path_for_file("Tiere/Mäuse/Stuart.txt"); // Create path for file
\endcode
__Process and system operations:__
\code
fl_execvp("löschen", argv); // Execute program
fl_system("echo 'Hallo Rüdiger!'"); // Execute system command
char* value = fl_getenv("crème"); // Get environment variable
\endcode
\section unicode_best_practices Best Practices
\subsection unicode_practices_files File Handling
- Always use fl_fopen(), fl_open(), etc. for file operations with
international names
- Test with international filenames during development
\subsection unicode_practices_strings String Processing
- Use fl_utf8strlen() instead of strlen() for character counts
- Use fl_utf8fwd()/fl_utf8back() when iterating through text character
by character
- Validate user input with fl_utf8test() if accepting external data
- Be careful when truncating strings - use character boundaries,
not arbitrary byte positions
\subsection unicode_practices_display Display and UI
- Test your interface with text in various languages (especially long
German words or wide Asian characters)
- Consider that text length varies greatly between languages when
designing layouts
- Ensure your chosen fonts support the characters you need to display
\subsection unicode_practices_performance Performance Notes
- ASCII text has no performance overhead compared to single-byte encodings
- UTF-8 functions are optimized for common cases (ASCII and Western European text)
- File operations may be slightly slower on Windows due to UTF-16 conversion
\section unicode_troubleshooting Common Issues and Solutions
\subsection unicode_problem_display "My international text shows up as question marks"
__Solution:__ Ensure your text is UTF-8 encoded and your font supports
the characters. If reading from files, verify they're saved as UTF-8.
\subsection unicode_problem_files "File operations fail with international names"
__Solution:__ Use FLTK's Unicode file functions instead of standard C functions:
\code
// Instead of:
FILE* f = fopen("Löwe.txt", "r"); // May fail on Windows
// Use:
FILE* f = fl_fopen("Löwe.txt", "r"); // Works correctly
\endcode
\subsection unicode_problem_length "String length calculations are wrong"
__Solution:__ Use UTF-8 aware functions:
\code
// Wrong - counts bytes, not characters:
int len = strlen("Café"); // Returns 5
// Correct - counts characters:
int len = fl_utf8strlen("Café"); // Returns 4
\endcode
\subsection unicode_problem_truncation "Text gets corrupted when I truncate it"
__Solution:__ Don't truncate UTF-8 strings at arbitrary byte positions:
\code
// Wrong - may cut in middle of character:
char truncated[10];
strncpy(truncated, utf8_text, 9);
// Correct - find proper character boundary:
const char* end = utf8_text;
int char_count = 0;
while (char_count < max_chars && *end) {
end = fl_utf8fwd(end, utf8_text, utf8_text + strlen(utf8_text));
char_count++;
}
int safe_length = end - utf8_text;
\endcode
\section unicode_error_handling Error Handling
FLTK handles invalid UTF-8 sequences gracefully using configurable behavior:
__Error handling modes (compile-time configuration):__
- __ERRORS_TO_CP1252__ (default): Treats bytes 0x80-0x9F as
CP1252 characters
- __STRICT_RFC3629__: Strict UTF-8 validation according to RFC 3629
- __ERRORS_TO_ISO8859_1__ (default): Invalid bytes returned as-is,
otherwise returns Unicode replacement character (U+FFFD)
\note You can configure these with compiler flags like -DERRORS_TO_CP1252=0
This design allows FLTK to handle legacy text files that mix encodings,
making it more robust in real-world scenarios.
\section unicode_limitations Current Limitations
FLTK's Unicode support covers most common use cases but has some limitations:
__Text Processing:__
- No automatic text normalization (combining characters are treated separately)
- No complex script shaping (may affect some Arabic, Indic scripts)
- No bidirectional text support (right-to-left languages like Arabic/Hebrew)
__Character Range:__
- Full Unicode range supported (U+000000 to U+10FFFF)
- Some legacy APIs may be limited to 16-bit characters (Basic Multilingual Plane)
__Sorting and Comparison:__
- String comparison is byte-based, not linguistically correct
- Use system locale functions for proper collation when needed for sorting
__Composed Characters:__
- Composed characters (base + combining accents) are treated as separate characters
- No automatic character composition or decomposition
Most applications won't encounter these limitations in practice. The Unicode
support in FLTK is sufficient for displaying and processing international
text in the majority of real-world scenarios.
\htmlonly