String functions for trimming whitespace

Author: vector-of-bool

src/common/src/mlib/str.h

@@ -288,6 +288,24 @@ _mstr_adjust_index(mstr_view s, mlib_upsized_integer pos, bool clamp_to_length)
  return pos.bits.as_unsigned;
 }
 
+/**
+ * @brief Obtain the code unit at the given zero-based index, with negative index wrapping.
+ *
+ * This function asserts that the index is in-bounds for the given string.
+ *
+ * @param s The string to be inspected.
+ * @param pos The index to access. Zero is the first code unit, and -1 is the last.
+ * @return char The code unit at position `pos`.
+ */
+static inline char
+mstr_at(mstr_view s, mlib_upsized_integer pos_)
+{
+ size_t pos = _mstr_adjust_index(s, pos_, false);
+ return s.data[pos];
+}
+
+#define mstr_at(S, Pos) (mstr_at)(mstr_view_from(S), mlib_upsize_integer(Pos))
+
 /**
  * @brief Create a new `mstr_view` that views a substring within another string
  *
@@ -441,6 +459,63 @@ mstr_find_first_of(mstr_view hay, mstr_view const needles, mlib_upsized_integer
 #define _mstr_find_first_of_argc_3(Hay, Needle, Pos) _mstr_find_first_of_argc_4(Hay, Needle, Pos, SIZE_MAX)
 #define _mstr_find_first_of_argc_4(Hay, Needle, Pos, Len) mstr_find_first_of(Hay, Needle, mlib_upsize_integer(Pos), Len)
 
+/**
+ * @brief Trim leading latin (ASCII) whitespace from the given string
+ *
+ * @param s The string to be inspected
+ * @return mstr_view A substring view of `s` that excludes any leading whitespace
+ */
+static inline mstr_view
+mstr_trim_left(mstr_view s)
+{
+ while (s.len) {
+ char c = mstr_at(s, 0);
+ if (c == ' ' || c == '\n' || c == '\r' || c == '\t') {
+ s = mstr_substr(s, 1);
+ } else {
+ break;
+ }
+ }
+ return s;
+}
+#define mstr_trim_left(S) (mstr_trim_left)(mstr_view_from(S))
+
+/**
+ * @brief Trim trailing latin (ASCII) whitespace from the given string
+ *
+ * @param s The string to be insepcted
+ * @return mstr_view A substring view of `s` that excludes any trailing whitespace.
+ */
+static inline mstr_view
+mstr_trim_right(mstr_view s)
+{
+ while (s.len) {
+ char c = mstr_at(s, -1);
+ if (c == ' ' || c == '\n' || c == '\r' || c == '\t') {
+ s = mstr_slice(s, 0, -1);
+ } else {
+ break;
+ }
+ }
+ return s;
+}
+#define mstr_trim_right(S) (mstr_trim_right)(mstr_view_from(S))
+
+/**
+ * @brief Trim leading and trailing latin (ASCII) whitespace from the string
+ *
+ * @param s The string to be inspected
+ * @return mstr_view A substring of `s` that excludes leading and trailing whitespace.
+ */
+static inline mstr_view
+mstr_trim(mstr_view s)
+{
+ s = mstr_trim_left(s);
+ s = mstr_trim_right(s);
+ return s;
+}
+#define mstr_trim(S) (mstr_trim)(mstr_view_from(S))
+
 /**
  * @brief Split a single string view into two strings at the given position
  *

src/common/tests/test-mlib.c

@@ -920,6 +920,14 @@ _test_str_view(void)
  // But "Food" > "foo" when case-insensitive:
  mlib_check(mstr_latin_casecmp(mstr_cstring("Food"), >, mstr_cstring("foo")));
  }
+
+ // Trimming
+ {
+ mstr_view s = mstr_cstring(" foo bar \n");
+ mlib_check(mstr_cmp(mstr_trim_left(s), ==, mstr_cstring("foo bar \n")));
+ mlib_check(mstr_cmp(mstr_trim_right(s), ==, mstr_cstring(" foo bar")));
+ mlib_check(mstr_cmp(mstr_trim(s), ==, mstr_cstring("foo bar")));
+ }
 }
 
 static inline void
@@ -1226,7 +1234,7 @@ _test_int_vec(void)
  int_vec_erase(&ints, ints.data + 1, ints.data + 3);
  mlib_check(ints.size, eq, 2, because, "We erased two elements");
 
- mlib_check(mlib_vec_at(ints, -1), eq, -7, because, "Negative index wraps");
+ mlib_check(mlib_vec_at(ints, -1), eq, 1729, because, "Negative index wraps");
 
  int_vec_destroy(&ints);
 }