u 0.5.0 → 1.0.0

data/ext/u/rb_u_string_to_inum.h

@@ -0,0 +1 @@
+VALUE rb_u_string_to_inum(VALUE str, int base, bool verify);

data/ext/u/rb_u_string_to_str.c

@@ -0,0 +1,17 @@
+#include "rb_includes.h"
+
+/* @return The String representation of the receiver, inheriting any taint and
+ *   untrust, encoded as UTF-8 */
+VALUE
+rb_u_string_to_str(VALUE self)
+{
+        const struct rb_u_string *string = RVAL2USTRING(self);
+
+        VALUE result = NIL_P(string->rb) ?
+                rb_u_str_new(USTRING_STR(string), USTRING_LENGTH(string)) :
+                string->rb;
+
+        OBJ_INFECT(result, self);
+
+        return result;
+}

data/ext/u/rb_u_string_to_sym.c

@@ -0,0 +1,12 @@
+#include "rb_includes.h"
+
+/* @raise [EncodingError] If the receiver contains an invalid UTF-8 sequence
+ * @raise [RuntimeError] If there’s no more room for a new Symbol in Ruby’s
+ *   Symbol table
+ * @return [Symbol] The Symbol representation of the receiver */
+VALUE
+rb_u_string_to_sym(VALUE self)
+{
+        /* NOTE: Lazy, but MRI makes it hard to implement this method. */
+        return rb_str_intern(StringValue(self));
+}

data/ext/u/rb_u_string_tr.c

@@ -0,0 +1,290 @@
+#include "rb_includes.h"
+#include "rb_u_string_internal_tr.h"
+
+#include "output.h"
+
+struct tr_range
+{
+        uint32_t begin;
+        uint32_t end;
+};
+
+static int
+tr_ranges_setup(struct tr *tr, struct tr_range *ranges)
+{
+        int n = 0;
+        bool was_inside_range = false;
+        while (tr_next(tr) != TR_FINISHED) {
+                if (tr->inside_range) {
+                        if (!was_inside_range) {
+                                ranges[n].begin = tr->now;
+                                was_inside_range = true;
+                        }
+                } else {
+                        if (was_inside_range)
+                                ranges[n].end = tr->now;
+                        else
+                                ranges[n].begin = ranges[n].end = tr->now;
+                        n++;
+                        was_inside_range = false;
+                }
+        }
+
+        return n;
+}
+
+
+struct tr_trans_closure
+{
+        struct tr_range *from;
+        int n_from;
+        struct tr_range *to;
+        int n_to;
+};
+
+static uint32_t
+tr_trans_replace_exclude(UNUSED(uint32_t c), void *closure)
+{
+        return *((uint32_t *)closure);
+}
+
+static int
+tr_trans_replace_include_offset_of(struct tr_range *ranges, int range, uint32_t c)
+{
+        int offset = 0;
+
+        for (int i = 0; i < range; i++)
+                offset += ranges[i].end - ranges[i].begin + 1;
+        offset += c - ranges[range].begin;
+
+        return offset;
+}
+
+static int
+tr_trans_replace_include_find_from_range(struct tr_trans_closure *closure, uint32_t c)
+{
+        for (int i = closure->n_from - 1; i >= 0; i--)
+                if (closure->from[i].begin <= c && c <= closure->from[i].end)
+                        return i;
+
+        return -1;
+}
+
+static uint32_t
+tr_trans_replace_include_find_to_u_char(struct tr_trans_closure *closure, int offset)
+{
+        for (int i = 0, seen = 0; i < closure->n_to; i++) {
+                int size = closure->to[i].end - closure->to[i].begin + 1;
+                if (seen + size > offset)
+                        return closure->to[i].begin + (offset - seen);
+                seen += size;
+        }
+
+        return closure->to[closure->n_to - 1].end;
+}
+
+static uint32_t
+tr_trans_replace_include(uint32_t c, void *v_closure)
+{
+        struct tr_trans_closure *closure = (struct tr_trans_closure *)v_closure;
+
+        int from = tr_trans_replace_include_find_from_range(closure, c);
+        if (from == -1)
+                return closure->to[closure->n_to - 1].end;
+
+        int offset = tr_trans_replace_include_offset_of(closure->from, from, c);
+
+        return tr_trans_replace_include_find_to_u_char(closure, offset);
+}
+
+static void
+tr_trans_real_squeeze(const char *str, const char *end,
+                      struct tr_table *translation,
+                      uint32_t replace(uint32_t, void *), void *closure,
+                      struct output *output, bool *modified)
+{
+        size_t n = output->n;
+        const char *p = str;
+        uint32_t prev_c = U_N_CODEPOINTS;
+        while (p < end) {
+                const char *prev = p;
+                uint32_t c0 = u_decode(&p, p, end);
+                if (tr_table_lookup(translation, c0)) {
+                        uint32_t c = replace(c0, closure);
+                        if (prev_c == c)
+                                continue;
+                        prev_c = c;
+                        output_char(output, c);
+                        if (c != c0)
+                                *modified = true;
+                } else {
+                        output_string(output, prev, p - prev);
+                        prev_c = U_N_CODEPOINTS;
+                }
+        }
+
+        if ((size_t)(end - str) > (output->n - n))
+                *modified = true;
+}
+
+static void
+tr_trans_real_standard(const char *str, const char *end,
+                       struct tr_table *translation,
+                       uint32_t replace(uint32_t, void *), void *closure,
+                       struct output *output, bool *modified)
+{
+        const char *p = str;
+
+        while (p < end) {
+                const char *prev = p;
+                uint32_t c = u_decode(&p, p, end);
+                if (tr_table_lookup(translation, c)) {
+                        uint32_t replacement = replace(c, closure);
+                        output_char(output, replacement);
+                        if (replacement != c)
+                                *modified = true;
+                } else
+                        output_string(output, prev, p - prev);
+        }
+}
+
+static void
+tr_trans_real(const char *str, const char *end,
+              struct tr_table *translation,
+              uint32_t replace(uint32_t, void *), void *closure, bool squeeze,
+              struct output *output, bool *modified)
+{
+        if (squeeze)
+                tr_trans_real_squeeze(str, end,
+                                      translation,
+                                      replace, closure,
+                                      output, modified);
+        else
+                tr_trans_real_standard(str, end,
+                                       translation,
+                                       replace, closure,
+                                       output, modified);
+}
+
+static VALUE
+tr_trans_do(VALUE self, struct tr_table *translation,
+            uint32_t (*replace)(uint32_t, void *), void *closure, bool squeeze)
+{
+        const struct rb_u_string *string = RVAL2USTRING(self);
+
+        const char *begin = USTRING_STR(string);
+        const char *end = USTRING_END(string);
+        bool modified = false;
+        struct output output = OUTPUT_INIT(NULL, 0);
+        tr_trans_real(begin, end,
+                      translation,
+                      replace, closure, squeeze,
+                      &output, &modified);
+        if (!modified)
+                return self;
+        output.result = ALLOC_N(char, output.n + 1);
+        output.m = output.n + 1;
+        output.n = 0;
+        tr_trans_real(begin, end,
+                      translation,
+                      replace, closure, squeeze,
+                      &output, &modified);
+        output_finalize(&output);
+
+        return rb_u_string_new_c_own(self, output.result, output.n);
+}
+
+static VALUE
+tr_trans(VALUE self, VALUE rbfrom, VALUE rbto, bool squeeze)
+{
+        const struct rb_u_string *string = RVAL2USTRING(self);
+        const struct rb_u_string *from = RVAL2USTRING_ANY(rbfrom);
+        const struct rb_u_string *to = RVAL2USTRING_ANY(rbto);
+
+        if (USTRING_STR(string) == NULL || USTRING_LENGTH(string) == 0)
+                return self;
+
+        if (USTRING_LENGTH(to) == 0)
+                return rb_u_string_delete(1, &rbfrom, self);
+
+        struct tr tr_from;
+        tr_init(&tr_from, USTRING_STR(from), USTRING_END(from));
+
+        struct tr tr_to;
+        tr_init(&tr_to, USTRING_STR(to), USTRING_END(to));
+
+        struct tr_table translation;
+        tr_table_initialize(&translation, rbfrom);
+
+        if (tr_should_exclude(&tr_from)) {
+                /* This case is easy.  Just include everything by default and
+                 * exclude the rest as always.  Replace characters found by the
+                 * last character found in tr_to. */
+                while (tr_next(&tr_to) != TR_FINISHED)
+                       ; /* We just need the last replacement character. */
+                return tr_trans_do(self, &translation, tr_trans_replace_exclude,
+                                   &tr_to.now, squeeze);
+        }
+
+        /* This case is hard.  We need a full-fledged lookup of what character
+         * to translate to, not simply a check whether to include it or not. */
+        struct tr_trans_closure trans_closure;
+
+        struct tr_range from_ranges[u_n_chars_n(USTRING_STR(from), USTRING_LENGTH(from))];
+        trans_closure.from = from_ranges;
+        trans_closure.n_from = tr_ranges_setup(&tr_from, from_ranges);
+
+        struct tr_range to_ranges[u_n_chars_n(USTRING_STR(to), USTRING_LENGTH(to))];
+        trans_closure.to = to_ranges;
+        trans_closure.n_to = tr_ranges_setup(&tr_to, to_ranges);
+
+        return tr_trans_do(self, &translation, tr_trans_replace_include,
+                           &trans_closure, squeeze);
+}
+
+/* @overload tr(from, to)
+ *
+ *   Returns the receiver, translating characters in FROM to their equivalent
+ *   character, by index, in TO, inheriting any taint and untrust.  If
+ *   TO{#length} < FROM{#length}, TO[-1] will be used for any index _i_ >
+ *   TO{#length}.
+ *
+ *   The complement of all Unicode characters and a given set of characters may
+ *   be specified by prefixing a non-empty set with ‘`^`’ (U+005E CIRCUMFLEX
+ *   ACCENT).
+ *
+ *   Any sequence of characters _a_-_b_ inside a set will expand to also
+ *   include all characters whose code points lay between those of _a_ and _b_.
+ *
+ *   @param [#to_str] from
+ *   @param [#to_str] to
+ *   @return [U::String] */
+VALUE
+rb_u_string_tr(VALUE self, VALUE from, VALUE to)
+{
+        return tr_trans(self, from, to, false);
+}
+
+/* @overload tr_s(from, to)
+ *
+ *   Returns the receiver, translating characters in FROM to their equivalent
+ *   character, by index, in TO and then squeezing any substrings of
+ *   {#length} > 1 consisting of the same character _c_ with _c_, inheriting
+ *   any taint and untrust.  If TO{#length} < FROM{#length}, TO[-1] will be
+ *   used for any index _i_ > TO{#length}.
+ *
+ *   The complement of all Unicode characters and a given set of characters may
+ *   be specified by prefixing a non-empty set with ‘`^`’ (U+005E CIRCUMFLEX
+ *   ACCENT).
+ *
+ *   Any sequence of characters _a_-_b_ inside a set will expand to also
+ *   include all characters whose code points lay between those of _a_ and _b_.
+ *
+ *   @param [#to_str] from
+ *   @param [#to_str] to
+ *   @return [U::String] */
+VALUE
+rb_u_string_tr_s(VALUE self, VALUE from, VALUE to)
+{
+        return tr_trans(self, from, to, true);
+}

data/ext/u/rb_u_string_upcase.c

@@ -0,0 +1,12 @@
+#include "rb_includes.h"
+
+/* @overload upcase(locale = ENV['LC_CTYPE'])
+ *   @param [#to_str] locale
+ *   @return [U::String] The upcasing of the receiver according to the rules of
+ *     of the language of LOCALE, which may be empty to specifically use the
+ *     default, language-independent, rules, inheriting any taint and untrust */
+VALUE
+rb_u_string_upcase(int argc, VALUE *argv, VALUE self)
+{
+        return _rb_u_string_convert_locale(argc, argv, self, u_upcase, NULL);
+}

data/ext/u/rb_u_string_upper.c

@@ -0,0 +1,13 @@
+#include "rb_includes.h"
+
+/* @overload upper?(locale = ENV[LC_CTYPE])
+ *   @param [#to_str] locale
+ *   @return [Boolean] True if the receiver has been upcased according to the
+ *     rules of the language of LOCALE, which may be empty to specifically use
+ *     the default, language-independent, rules, that is, if _a_ =
+ *     _a_{#upcase}(LOCALE), where _a_ = {#normalize}(`:nfd`) */
+VALUE
+rb_u_string_upper(int argc, VALUE *argv, VALUE self)
+{
+        return _rb_u_string_test_locale(argc, argv, self, u_upcase);
+}

data/ext/u/rb_u_string_valid.c

@@ -0,0 +1,10 @@
+#include "rb_includes.h"
+
+/* @overload valid?
+ *   @return [Boolean] True if the receiver contains only valid Unicode
+ *     characters */
+VALUE
+rb_u_string_valid(VALUE self)
+{
+        return _rb_u_character_test(self, u_char_isvalid);
+}

data/ext/u/rb_u_string_valid_encoding.c

@@ -0,0 +1,12 @@
+#include "rb_includes.h"
+
+/* @overload valid_encoding?
+ *   @return [Boolean] True if the receiver contains only valid UTF-8
+ *     sequences */
+VALUE
+rb_u_string_valid_encoding(VALUE self)
+{
+        const struct rb_u_string *string = RVAL2USTRING(self);
+
+        return u_valid(USTRING_STR(string), USTRING_LENGTH(string), NULL) ? Qtrue : Qfalse;
+}

data/ext/u/rb_u_string_wide.c

@@ -0,0 +1,21 @@
+#include "rb_includes.h"
+
+/* @overload wide?
+ *
+ *   Returns true if the receiver contains only “wide” characters.  Wide
+ *   character are those that have their East_Asian_Width property set to Wide
+ *   or Fullwidth.
+ *
+ *   This is mostly useful for determining how many “cells” a character will
+ *   take up on a terminal or similar cell-based display.
+ *
+ *   @return [Boolean]
+ *   @see http://www.unicode.org/reports/tr11/
+ *     Unicode Standard Annex #11: East Asian Width
+ *   @see #wide_cjk?
+ *   @see #width */
+VALUE
+rb_u_string_wide(VALUE self)
+{
+        return _rb_u_character_test(self, u_char_iswide);
+}

data/ext/u/rb_u_string_wide_cjk.c

@@ -0,0 +1,21 @@
+#include "rb_includes.h"
+
+/* @overload wide_cjk?
+ *
+ *   Returns true if the receiver contains only “wide” and “ambiguously wide”
+ *   characters.  Wide and ambiguously wide character are those that have their
+ *   East_Asian_Width property set to Ambiguous, Wide or Fullwidth.
+ *
+ *   This is mostly useful for determining how many “cells” a character will
+ *   take up on a terminal or similar cell-based display.
+ *
+ *   @return [Boolean]
+ *   @see http://www.unicode.org/reports/tr11/
+ *     Unicode Standard Annex #11: East Asian Width
+ *   @see #wide?
+ *   @see #width */
+VALUE
+rb_u_string_wide_cjk(VALUE self)
+{
+        return _rb_u_character_test(self, u_char_iswide_cjk);
+}

data/ext/u/rb_u_string_width.c

@@ -0,0 +1,19 @@
+#include "rb_includes.h"
+
+/* Returns the width of the receiver.  The width is defined as the sum of the
+ * number of “cells” on a terminal or similar cell-based display that the
+ * characters in the string will require.
+ *
+ * Characters that are {#wide?} have a width of 2.  Characters that are
+ * {#zero_width?} have a width of 0.  Other characters have a width of 1.
+ *
+ * @return [Integer]
+ * @see http://www.unicode.org/reports/tr11/
+ *   Unicode Standard Annex #11: East Asian Width */
+VALUE
+rb_u_string_width(VALUE self)
+{
+        const struct rb_u_string *string = RVAL2USTRING(self);
+
+        return UINT2NUM(u_width_n(USTRING_STR(string), USTRING_LENGTH(string)));
+}

data/ext/u/rb_u_string_word_break.c

@@ -0,0 +1,63 @@
+#include "rb_includes.h"
+
+#define BREAK2ID(value, symbol) \
+        case U_WORD_BREAK_##value: { \
+                static ID id_##symbol; \
+                if (id_##symbol == 0) \
+                        id_##symbol = rb_intern(#symbol); \
+                return ID2SYM(id_##symbol); \
+        }
+
+static VALUE
+break_to_symbol(enum u_word_break value)
+{
+        switch (value) {
+	BREAK2ID(ALETTER, aletter);
+	BREAK2ID(CR, cr);
+	BREAK2ID(EXTEND, extend);
+	BREAK2ID(EXTENDNUMLET, extendnumlet);
+	BREAK2ID(FORMAT, format);
+	BREAK2ID(KATAKANA, katakana);
+	BREAK2ID(LF, lf);
+	BREAK2ID(MIDLETTER, midletter);
+	BREAK2ID(MIDNUM, midnum);
+	BREAK2ID(MIDNUMLET, midnumlet);
+	BREAK2ID(NEWLINE, newline);
+	BREAK2ID(NUMERIC, numeric);
+	BREAK2ID(OTHER, other);
+	BREAK2ID(REGIONAL_INDICATOR, regional_indicator);
+        default:
+                rb_u_raise(rb_eNotImpError, "unknown word break value: %d", value);
+        }
+}
+
+/* Returns the word break property value of the characters of the receiver.
+ *
+ * The possible word break values are
+ *
+ * * :aletter
+ * * :cr
+ * * :extend
+ * * :extendnumlet
+ * * :format
+ * * :katakana
+ * * :lf
+ * * :midletter
+ * * :midnum
+ * * :midnumlet
+ * * :newline
+ * * :numeric
+ * * :other
+ * * :regional_indicator
+ *
+ * @raise [ArgumentError] If the string consists of more than one break type
+ * @return [Symbol]
+ * @see http://www.unicode.org/reports/tr29/
+ *   Unicode Standard Annex #29: Unicode Text Segmentation */
+VALUE
+rb_u_string_word_break(VALUE self)
+{
+        return _rb_u_string_property(self, "word break", U_WORD_BREAK_OTHER,
+                                     (int (*)(uint32_t))u_char_word_break,
+                                     (VALUE (*)(int))break_to_symbol);
+}