Add new C-style string splitting function to string_utils.

The motivation for this is that we need to split strings into
C-style `**char`, to later pass them on to system calls that
are based on that interface.

Co-authored-by: Fotis Koutoulakis <[email protected]>
Co-authored-by: Thomas Given-Wilson <[email protected]>

Author: NlightNFotis

src/util/string_utils.cpp

@@ -1,3 +1,5 @@
+
+
 /*******************************************************************\
 
 Module:
@@ -12,6 +14,7 @@ Author: Daniel Poetzl
 
 #include <algorithm>
 #include <cctype>
+#include <cstring>
 #include <iomanip>
 
 /// Remove all whitespace characters from either end of a string. Whitespace
@@ -252,3 +255,23 @@ std::string wrap_line(
 
   return result;
 }
+
+char **string_to_cstr_array(const std::string &s, const char *delims)
+{
+  char **res = NULL;
+  int n_spaces = 0;
+  char *p = strtok(strdup(s.c_str()), delims);
+
+  while(p)
+  {
+    res = reinterpret_cast<char **>(realloc(res, sizeof(char *) * ++n_spaces));
+    INVARIANT(res != nullptr, "Memory allocation failed");
+    res[n_spaces - 1] = p;
+    p = strtok(NULL, delims);
+  }
+
+  res =
+    reinterpret_cast<char **>(realloc(res, sizeof(char *) * (n_spaces + 1)));
+  res[n_spaces] = 0;
+  return res;
+}

src/util/string_utils.h

@@ -146,4 +146,14 @@ std::string wrap_line(
   const std::size_t left_margin = 0,
   const std::size_t width = 80);
 
+/// Split a string into an array of char*.
+///
+/// This was motivated by string handling for passing arguments to low
+/// level system calls, that have C style signatures.
+/// \param s: The string to split
+/// \param delims: The delimiter characters to split on
+/// \return a NULL terminated char** that has each of the space delimited
+///  substrings of the input.
+char **string_to_cstr_array(const std::string &s, const char *delims);
+
 #endif

unit/Makefile

@@ -142,6 +142,7 @@ SRC += analyses/ai/ai.cpp \
        util/string_utils/join_string.cpp \
        util/string_utils/split_string.cpp \
        util/string_utils/strip_string.cpp \
+       util/string_utils/string_to_cstyle_array.cpp \
        util/string_utils/wrap_line.cpp \
        util/symbol_table.cpp \
        util/symbol.cpp \

unit/util/string_utils/string_to_cstyle_array.cpp

@@ -0,0 +1,54 @@
+/// \file string_to_cstr_array Unit Tests
+/// \author Diffblue Ltd.
+
+#include <testing-utils/use_catch.h>
+#include <util/string_utils.h>
+
+TEST_CASE(
+  "Split a simple string into three parts",
+  "[core][utils][string_utils][string_to_c_style_array]")
+{
+  const std::string test_string = "The Jabberwocky bites";
+  char **result = string_to_cstr_array(test_string, " ");
+
+  REQUIRE(strcmp(result[0], "The") == 0);
+  REQUIRE(strcmp(result[1], "Jabberwocky") == 0);
+  REQUIRE(strcmp(result[2], "bites") == 0);
+  REQUIRE(result[3] == 0);
+}
+
+TEST_CASE(
+  "Split an empty string",
+  "[core][utils][string_utils][string_to_c_style_array]")
+{
+  const std::string test_string = "";
+  char **result = string_to_cstr_array(test_string, " ");
+
+  REQUIRE(result[0] == 0);
+}
+
+TEST_CASE(
+  "Split a simple string into three parts with more spaces",
+  "[core][utils][string_utils][string_to_c_style_array]")
+{
+  const std::string test_string = "The   Jabberwocky bites";
+  char **result = string_to_cstr_array(test_string, " ");
+
+  REQUIRE(strcmp(result[0], "The") == 0);
+  REQUIRE(strcmp(result[1], "Jabberwocky") == 0);
+  REQUIRE(strcmp(result[2], "bites") == 0);
+  REQUIRE(result[3] == 0);
+}
+
+TEST_CASE(
+  "Split a simple string into three parts with more spaces and tabs",
+  "[core][utils][string_utils][string_to_c_style_array]")
+{
+  const std::string test_string = "The \tJabberwocky \tbites";
+  char **result = string_to_cstr_array(test_string, " \t");
+
+  REQUIRE(strcmp(result[0], "The") == 0);
+  REQUIRE(strcmp(result[1], "Jabberwocky") == 0);
+  REQUIRE(strcmp(result[2], "bites") == 0);
+  REQUIRE(result[3] == 0);
+}