power_play/src/string.c

#include "string.h"
#include "arena.h"
#include "memory.h"
#include "scratch.h"
#include "math.h"
#include "uni.h"

/*
 * NOTE: Strings should be considered ~IMMUTABLE~
 *
 * All string functions return a new string as a result. Any strings used as
 * an argument (IE: in string_cat) will not be modified.
 *
 * Use the LIT macro to create strings from string literals
 *
 * NOTE: It is valid for a string to have len 0 but a non-NULL text pointer.
 * Always check string.len rather than string.text for string presence.
 * (If we want to change this behavior then we need to check for length = 0 in
 * our functions that return a pointer from arena_dry_push, or guarantee that
 * all functions returning an arena_dry_push do allocate.)
 */

/* ========================== *
 * Conversion
 * ========================== */

#define INT_CHARS ("0123456789abcdef")

struct string string_from_char(struct arena *arena, char c)
{
    u8 *dest = arena_push(arena, u8);
    *dest = c;
    return (struct string) {
        .len = 1,
        .text = dest
    };
}

struct string string_from_uint(struct arena *arena, u64 n, u64 base, u64 zfill)
{
    /* Base too large */
    ASSERT(base <= (ARRAY_COUNT(INT_CHARS) - 1));

    struct temp_arena scratch = scratch_begin(arena);

    /* Build backwards text starting from least significant digit */
    u64 len = 0;
    u8 *backwards_text = arena_dry_push(scratch.arena, u8);
    do {
        string_from_char(scratch.arena, INT_CHARS[n % base]);
        ++len;
        n /= base;
    } while (n > 0);

    while (len < zfill) {
        string_from_char(scratch.arena, '0');
        ++len;
    }

    /* Reverse text into final string */
    u8 *final_text = arena_push_array(arena, u8, len);
    for (u64 i = 0; i < len; ++i) {
        final_text[i] = backwards_text[len - i - 1];
    }

    scratch_end(scratch);

    return (struct string) {
        .len = len,
        .text = final_text
    };
}

struct string string_from_int(struct arena *arena, i64 n, u64 base, u64 zfill)
{
    u8 *final_text = arena_dry_push(arena, u8);
    u8 len = 0;
    if (n < 0) {
        /* Push sign */
        string_from_char(arena, '-');
        len = 1;
        n = -n;
    }
    /* Push unsigned number */
    struct string uint_str = string_from_uint(arena, n, base, zfill);
    return (struct string) {
        .len = len + uint_str.len,
        .text = final_text
    };
}

struct string string_from_ptr(struct arena *arena, void *ptr)
{
    struct string prepend = string_copy(arena, LIT("0x"));
    struct string uint_str = string_from_uint(arena, (u64)ptr, 16, sizeof(ptr));
    return (struct string) {
        .len = prepend.len + uint_str.len,
        .text = prepend.text
    };
}

struct string string_from_float(struct arena *arena, f64 f, u32 precision)
{
    struct temp_arena scratch = scratch_begin(arena);
    u8 *final_text = arena_dry_push(arena, u8);
    u64 final_len = 0;

    if (F32_IS_NAN(f)) {
        final_len += string_copy(arena, LIT("NaN")).len;
    } else if (f == F64_INFINITY) {
        final_len += string_copy(arena, LIT("inf")).len;
    } else if (f == -F64_INFINITY) {
        final_len += string_copy(arena, LIT("-inf")).len;
    } else {
        if (f < 0) {
            string_from_char(arena, '-');
            f = -f;
            ++final_len;
        }

        /* Add one half of next precision level to round up */
        f += 0.5 / (f64)math_pow_u64(10, (u8)precision);

        f64 part_whole = math_trunc64(f);
        f64 part_decimal = f - part_whole;

        /* Print whole part */
        {
            /* Build backwards text starting from least significant digit */
            u8 *backwards_text = arena_dry_push(scratch.arena, u8);
            u64 backwards_text_len = 0;
            do {
                u64 digit = (u64)math_round_to_int64(math_fmod64(part_whole, 10.0));
                string_from_char(scratch.arena, INT_CHARS[digit % 10]);
                ++backwards_text_len;
                part_whole = math_trunc64(part_whole / 10.0);
            } while (part_whole > 0);

            /* Reverse text into final string */
            arena_push_array(arena, u8, backwards_text_len);
            for (u64 i = backwards_text_len; i-- > 0;) {
                final_text[final_len++] = backwards_text[i];
            }
        }

        /* Print decimal part */
        if (precision > 0) {
            string_from_char(arena, '.');
            for (u64 i = 0; i < precision; ++i) {
                part_decimal *= 10.0;
                u64 digit = (u64)part_decimal;
                part_decimal -= digit;
                string_from_char(arena, INT_CHARS[digit % 10]);
            }
            final_len += (u64)precision + 1;
        }
    }

    scratch_end(scratch);

    return (struct string) {
        .len = final_len,
        .text = final_text
    };
}

struct string string_from_handle(struct arena *arena, u64 v0, u64 v1)
{
    struct string res = ZI;
    res.text = arena_dry_push(arena, u8);
    res.len += string_copy(arena, LIT("h")).len;
    res.len += string_from_uint(arena, v0, 16, 0).len;
    res.len += string_copy(arena, LIT("x")).len;
    res.len += string_from_uint(arena, v1, 16, 0).len;
    return res;
}

struct string string_from_uid(struct arena *arena, struct uid uid)
{
    struct string res = ZI;
    res.text = arena_dry_push(arena, u8);
    res.len += string_from_uint(arena, (uid.hi >> 32), 16, 8).len;
    return res;
}

/* ========================== *
 * String operations
 * ========================== */

struct string string_copy(struct arena *arena, struct string src)
{
    struct string str = {
        .len = src.len,
        .text = arena_push_array(arena, u8, src.len)
    };
    MEMCPY(str.text, src.text, src.len);
    return str;
}

struct string string_copy_to_string(struct string dst, struct string src)
{
    struct string res = ZI;
    res.len = min_u64(dst.len, src.len);
    res.text = dst.text;
    MEMCPY(res.text, src.text, res.len);
    return res;
}

struct string string_repeat(struct arena *arena, struct string src, u64 count)
{
    u64 final_len = src.len * count;
    u8 *final_text = arena_push_array(arena, u8, final_len);
    for (u64 i = 0; i < count; ++i) {
        MEMCPY(final_text + (src.len * i), src.text, src.len);
    }
    return (struct string) {
        .text = final_text,
        .len = final_len
    };
}

struct string string_cat(struct arena *arena, struct string str1, struct string str2)
{
    struct string new_str = ZI;
    new_str.len = str1.len + str2.len;
    new_str.text = arena_push_array(arena, u8, new_str.len);
    MEMCPY(new_str.text, str1.text, str1.len);
    MEMCPY(new_str.text + str1.len, str2.text, str2.len);
    return new_str;
}

/* `arena` is where pieces items will be allocated. These strings point
 * into the existing supplied string and do not allocate any new text. */
struct string_array string_split(struct arena *arena, struct string str, struct string delim)
{
    struct string_array pieces = {
        .count = 0,
        .strings = arena_dry_push(arena, struct string)
    };

    struct string piece = {
        .len = 0,
        .text = str.text
    };

    for (u64 i = 0; i <= str.len - delim.len; ++i) {
        /* Clamp comparison string so we don't overflow. */
        struct string comp_str = {
            .len = delim.len,
            .text = &str.text[i]
        };

        b32 is_delimiter = string_eq(comp_str, delim);
        b32 is_end = i == str.len - 1;

        if (!is_delimiter || is_end) {
            ++piece.len;
        }

        if (is_delimiter || is_end) {
            /* Delimiter found */
            struct string *piece_pushed = arena_push(arena, struct string);
            *piece_pushed = piece;
            ++pieces.count;
            piece.text = piece.text + piece.len + delim.len;
            piece.len = 0;
        }
    }

    return pieces;
}

/* NOTE: Really slow */
struct string string_indent(struct arena *arena, struct string str, u32 indent)
{
    struct temp_arena scratch = scratch_begin(arena);

    u64 final_len = 0;
    u8 *final_text = arena_dry_push(arena, u8);

    struct string_array split = string_split(scratch.arena, str, LIT("\n"));
    for (u64 i = 0; i < split.count; ++i) {
        struct string piece = split.strings[i];
        for (u32 j = 0; j < indent; ++j) {
            string_from_char(arena, ' ');
            ++final_len;
        }
        string_copy(arena, piece);
        final_len += piece.len;
        if (i < split.count - 1) {
            string_from_char(arena, '\n');
            ++final_len;
        }
    }

    scratch_end(scratch);

    return (struct string) {
        .len = final_len,
        .text = final_text
    };
}

struct string string_lower(struct arena *arena, struct string str)
{
    struct string res = ZI;
    res.text = arena_push_array(arena, u8, str.len);
    res.len = str.len;

    for (u64 i = 0; i < str.len; ++i) {
        u8 c = str.text[i];
        if (65 <= c && c <= 90) {
            c += 32;
        }
        res.text[i] = c;
    }

    return res;
}

b32 string_eq(struct string str1, struct string str2)
{
    b32 eq = true;
    if (str1.len == str2.len) {
        for (u64 i = 0; i < str1.len; ++i) {
            if (str1.text[i] != str2.text[i]) {
                eq = false;
                break;
            }
        }
    } else {
        eq = false;
    }
    return eq;
}

i32 string_cmp(struct string str1, struct string str2)
{
    i32 res = 0;
    for (u64 i = 0; i < min_u64(str1.len, str2.len); ++i) {
        res = str1.text[i] - str2.text[i];
        if (res != 0) {
            break;
        }
    }
    if (res == 0) {
        if (str1.len > str2.len) {
            res = str1.text[str2.len];
        } else if (str2.len > str1.len) {
            res = str2.text[str1.len];
        }
    }
    return res;
}

b32 string_contains(struct string str, struct string substring)
{
    if (substring.len > str.len) {
        return false;
    }

    for (u64 i = 0; i <= str.len - substring.len; ++i) {
        b32 match = true;
        for (u64 j = 0; j < substring.len; ++j) {
            if (str.text[i + j] != substring.text[j]) {
                match = false;
                break;
            }
        }
        if (match) {
            return true;
        }
    }

    return false;
}

b32 string_starts_with(struct string str, struct string substring)
{
    if (str.len >= substring.len) {
        for (u64 i = 0; i < substring.len; ++i) {
            if (str.text[i] != substring.text[i]) {
                return false;
            }
        }
        return true;
    }
    return false;
}

b32 string_ends_with(struct string str, struct string substring)
{
    if (str.len >= substring.len) {
        u64 start = str.len - substring.len;
        for (u64 i = 0; i < substring.len; ++i) {
            if (str.text[start + i] != substring.text[i]) {
                return false;
            }
        }
        return true;
    }
    return false;
}

/* ========================== *
 * Format
 * ========================== */

/* String formatting only has one format specifier: "%F". All specifier info is
 * included in the arguments (instead of w/ the specifier like in printf). This
 * is safer.
 *
 * Example:
 * string_format(arena,
 *               LIT("Hello there %F. You are %F feet %F inches tall!"),
 *               FMT_STR(LIT("George")),
 *               FMT_UINT(6),
 *               FMT_FLOAT(5.375));
 *
 * NOTE: FMT_END must be passed as the last arg in the va_list (This is
 * done automatically by the `string_format` macro).
 *
 * Format arguments:
 * FMT_CHAR: Format a single u8 character
 * FMT_STR: Format a `string` struct
 * FMT_UINT: Format a u64
 * FMT_SINT: Format an i64
 * FMT_FLOAT: Format an f64 with DEFAULT_FMT_PRECISION
 * FMT_FLOAT_P: Format an f64 with specified precision
 * FMT_HEX: Format a u64 in hexadecimal notation
 * FMT_PTR: Format a pointer in hexadecimal notation prefixed by "0x"
 * FMT_HANDLE: Format a 128 bit handle
 *
 * FMT_END (internal): Denote the end of the va_list
 *
 * TODO:
 * %n equivalent? (nothing)
 * %e/%E equivalent? (scientific notation of floats)
 * %o equivalent? (octal representation)
 */
struct string string_formatv(struct arena *arena, struct string fmt, va_list args)
{
    __prof;

    u64 final_len = 0;
    u8 *final_text = arena_dry_push(arena, u8);

    u8 *end = fmt.text + fmt.len;
    b32 no_more_args = false;
    for (u8 *c = fmt.text; c < end; ++c) {
        u8 *next = ((c + 1) < end) ? (c + 1) : (u8 *)"\0";

        /* Escape '%%' */
        b32 escape = !no_more_args && *c == '%' && *next == '%';
        if (escape) {
            /* Skip the escape '%' char from parsing */
            ++c;
        }

        if (!no_more_args && !escape && *c == '%' && *next == 'F') {
            struct string parsed_str = ZI;
            /* Detect arg type and parse to string */
            struct fmt_arg arg = va_arg(args, struct fmt_arg);
            switch (arg.type) {
                case FMT_TYPE_CHAR: {
                    parsed_str = string_from_char(arena, arg.value.c);
                } break;

                case FMT_TYPE_STR: {
                    parsed_str = string_copy(arena, arg.value.string);
                } break;

                case FMT_TYPE_UINT: {
                    parsed_str = string_from_uint(arena, arg.value.uint, 10, arg.zfill);
                } break;

                case FMT_TYPE_SINT: {
                    parsed_str = string_from_int(arena, arg.value.sint, 10, arg.zfill);
                } break;

                case FMT_TYPE_HEX: {
                    parsed_str = string_from_uint(arena, arg.value.sint, 16, arg.zfill);
                } break;

                case FMT_TYPE_PTR: {
                    parsed_str = string_from_ptr(arena, arg.value.ptr);
                } break;

                case FMT_TYPE_FLOAT: {
                    parsed_str = string_from_float(arena, arg.value.f, arg.precision);
                } break;

                case FMT_TYPE_HANDLE: {
                    parsed_str = string_from_handle(arena, arg.value.handle.h64[0], arg.value.handle.h64[1]);
                } break;

                case FMT_TYPE_UID: {
                    parsed_str = string_from_uid(arena, arg.value.uid);
                } break;

                case FMT_TYPE_END: {
                    /* Unexpected end. Not enough FMT args passed to function. */
                    ASSERT(false);
                    parsed_str = string_copy(arena, LIT("<?>"));
                    no_more_args = true;
                } break;

                default: {
                    /* Unknown format type */
                    ASSERT(false);
                    parsed_str = string_copy(arena, LIT("<?>"));
                    no_more_args = true;
                } break;
            }
            /* Update final string len / start */
            final_len += parsed_str.len;
            /* Skip 'F' from parsing */
            ++c;
        } else {
            /* Parse character normally */
            string_from_char(arena, *c);
            ++final_len;
        }
    }

#if RTC
    if (!no_more_args) {
        struct fmt_arg last_arg = va_arg(args, struct fmt_arg);
        /* End arg not reached. Too many FMT values passed to function. */
        ASSERT(last_arg.type == FMT_TYPE_END);
    }
#endif

    return (struct string) {
      .len = final_len,
      .text = final_text
    };
}

struct string _string_format(struct arena *arena, struct string fmt, ...)
{
    va_list args;
    va_start(args, fmt);
    struct string new_str = string_formatv(arena, fmt, args);
    va_end(args);
    return new_str;
}

/* ========================== *
 * Unicode
 * ========================== */

/* Codepoint iteration */

struct string_codepoint_iter string_codepoint_iter_begin(struct string str)
{
    return (struct string_codepoint_iter) {
        .src = str
    };
}

b32 string_codepoint_iter_next(struct string_codepoint_iter *iter)
{
    if (iter->pos < iter->src.len) {
        struct string str_remaining = { .len = (iter->src.len - iter->pos), .text = iter->src.text + iter->pos };
        struct uni_decode_utf8_result decoded = uni_decode_utf8(str_remaining);
        iter->pos += decoded.advance8;
        iter->codepoint = decoded.codepoint;
        return true;
    } else {
        return false;
    }
}

void string_codepoint_iter_end(struct string_codepoint_iter *iter)
{
    /* Do nothing */
    (UNUSED)iter;
}

/* utf8 <- utf16 */
struct string string_from_string16(struct arena *arena, struct string16 str16)
{
    struct string res = {
        .len = 0,
        .text = arena_dry_push(arena, u8)
    };

    u64 pos16 = 0;
    while (pos16 < str16.len) {
        struct string16 str16_remaining = { .len = (str16.len - pos16), .text = str16.text + pos16 };
        struct uni_decode_utf16_result decoded = uni_decode_utf16(str16_remaining);
        struct uni_encode_utf8_result encoded = uni_encode_utf8(decoded.codepoint);

        u8 *dest = arena_push_array(arena, u8, encoded.count8);
        MEMCPY(dest, encoded.chars8, encoded.count8);

        pos16 += decoded.advance16;
        res.len += encoded.count8;
    }

    return res;
}

/* utf8 <- utf32 */
struct string string_from_string32(struct arena *arena, struct string32 str32)
{
    struct string res = {
        .len = 0,
        .text = arena_dry_push(arena, u8)
    };

    u64 pos32 = 0;
    while (pos32 < str32.len) {
        struct string32 str32_remaining = { .len = (str32.len - pos32), .text = str32.text + pos32 };
        struct uni_decode_utf32_result decoded = uni_decode_utf32(str32_remaining);
        struct uni_encode_utf8_result encoded = uni_encode_utf8(decoded.codepoint);

        u8 *dest = arena_push_array(arena, u8, encoded.count8);
        MEMCPY(dest, &encoded.chars8, encoded.count8);

        pos32 += 1;
        res.len += encoded.count8;
    }

    return res;
}

/* utf16 <- utf8 */
struct string16 string16_from_string(struct arena *arena, struct string str8)
{
    struct string16 res = {
        .len = 0,
        .text = arena_dry_push(arena, u16)
    };

    u64 pos8 = 0;
    while (pos8 < str8.len) {
        struct string str8_remaining = { .len = (str8.len - pos8), .text = str8.text + pos8 };
        struct uni_decode_utf8_result decoded = uni_decode_utf8(str8_remaining);
        struct uni_encode_utf16_result encoded = uni_encode_utf16(decoded.codepoint);

        u16 *dest = arena_push_array(arena, u16, encoded.count16);
        MEMCPY(dest, encoded.chars16, (encoded.count16 << 1));

        pos8 += decoded.advance8;
        res.len += encoded.count16;
    }

    return res;
}

/* utf32 <- utf8 */
struct string32 string32_from_string(struct arena *arena, struct string str8)
{
    struct string32 res = {
        .len = 0,
        .text = arena_dry_push(arena, u32)
    };

    u64 pos8 = 0;
    while (pos8 < str8.len) {
        struct string str8_remaining = { .len = (str8.len - pos8), .text = str8.text + pos8 };
        struct uni_decode_utf8_result decoded = uni_decode_utf8(str8_remaining);
        struct uni_encode_utf32_result encoded = uni_encode_utf32(decoded.codepoint);

        u32 *dest = arena_push(arena, u32);
        *dest = encoded.chars32;

        pos8 += decoded.advance8;
        res.len += 1;
    }

    return res;
}

/* ========================== *
 * C narrow strings
 * ========================== */

u64 cstr_len_no_limit(char *cstr)
{
    char *end = cstr;
    if (cstr) {
        while (*end) {
            ++end;
        }
    }
    return end - cstr;
}

u64 cstr_len(char *cstr, u64 limit)
{
    char *end = cstr;
    if (cstr) {
        for (u64 i = 0; i < limit; ++i) {
            if (*end) {
                ++end;
            } else {
                break;
            }
        }
    }
    return end - cstr;
}

char *cstr_from_string(struct arena *arena, struct string src)
{
    u8 *text = arena_push_array(arena, u8, src.len + 1);
    MEMCPY(text, src.text, src.len);
    text[src.len] = 0;
    return (char *)text;
}

char *cstr_buff_from_string(struct string dest_buff, struct string src)
{
    if (dest_buff.len > 0) {
        u64 len = min_u64(src.len, dest_buff.len - 1);
        MEMCPY(dest_buff.text, src.text, len);
        dest_buff.text[len] = 0;
    }
    return (char *)dest_buff.text;
}

struct string string_from_cstr_no_limit(char *cstr)
{
    u64 len = cstr_len_no_limit(cstr);
    return (struct string) {
        .len = len,
        .text = (u8 *)cstr
    };
}

struct string string_from_cstr(char *cstr, u64 limit)
{
    u64 len = cstr_len(cstr, limit);
    return (struct string) {
        .text = (u8 *)cstr,
        .len = len
    };
}

/* ========================== *
 * C wide strings
 * ========================== */

u64 wstr_len_no_limit(wchar_t *wstr)
{
    wchar_t *end = wstr;
    if (end) {
        while (*end) {
            ++end;
        }
    }
    return end - wstr;
}

u64 wstr_len(wchar_t *wstr, u64 limit)
{
    wchar_t *end = wstr;
    if (wstr) {
        for (u64 i = 0; i < limit; ++i) {
            if (*end) {
                ++end;
            } else {
                break;
            }
        }
    }
    return end - wstr;
}

wchar_t *wstr_from_string(struct arena *arena, struct string src)
{
    struct string16 str16 = string16_from_string(arena, src);
    *arena_push(arena, u16) = 0;
    return (wchar_t *)str16.text;
}

wchar_t *wstr_from_string16(struct arena *arena, struct string16 src)
{
    u16 *text = arena_push_array(arena, u16, src.len + 1);
    text[src.len] = 0;
    return (wchar_t *)text;
}

struct string string_from_wstr_no_limit(struct arena *arena, wchar_t *wstr)
{
    struct string16 str16 = string16_from_wstr_no_limit(wstr);
    return string_from_string16(arena, str16);
}

struct string string_from_wstr(struct arena *arena, wchar_t *wstr, u64 limit)
{
    struct string16 str16 = string16_from_wstr(wstr, limit);
    return string_from_string16(arena, str16);
}

struct string16 string16_from_wstr_no_limit(wchar_t *wstr)
{
    u64 len = wstr_len_no_limit(wstr);
    return (struct string16) {
        .len = len,
        .text = (u16 *)wstr
    };
}

struct string16 string16_from_wstr(wchar_t *wstr, u64 limit)
{
    u64 len = wstr_len(wstr, limit);
    return (struct string16)
    {
        .len = len,
        .text = (u16 *)wstr
    };
}