From 0748b124725527ddbf71442e054528fcfa03820d Mon Sep 17 00:00:00 2001
From: Axel Kohlmeyer <akohlmey@gmail.com>
Date: Fri, 17 Jul 2020 10:08:46 -0400
Subject: [PATCH] when processing quoted strings, the quotes need to be removed

---
 src/utils.cpp                 | 13 ++++++++-----
 unittest/utils/test_utils.cpp | 27 ++++++++++++++++++++-------
 2 files changed, 28 insertions(+), 12 deletions(-)

diff --git a/src/utils.cpp b/src/utils.cpp
index 745cb72b56..4efa35a8d1 100644
--- a/src/utils.cpp
+++ b/src/utils.cpp
@@ -436,6 +436,7 @@ std::vector<std::string> utils::split_words(const std::string &text)
   const char *buf = text.c_str();
   std::size_t beg = 0;
   std::size_t len = 0;
+  std::size_t add = 0;
   char c = *buf;
 
   while (c) {
@@ -452,8 +453,9 @@ std::vector<std::string> utils::split_words(const std::string &text)
 
     // handle single quote
     if (c == '\'') {
+      ++beg;
+      add = 1;
       c = *++buf;
-      ++len;
       while (((c != '\'') && (c != '\0'))
              || ((c == '\\') && (buf[1] == '\''))) {
         if ((c == '\\') && (buf[1] == '\'')) {
@@ -463,13 +465,14 @@ std::vector<std::string> utils::split_words(const std::string &text)
         c = *++buf;
         ++len;
       }
+      if (c != '\'') ++len;
       c = *++buf;
-      ++len;
 
       // handle double quote
     } else if (c == '"') {
+      ++beg;
+      add = 1;
       c = *++buf;
-      ++len;
       while (((c != '"') && (c != '\0'))
              || ((c == '\\') && (buf[1] == '"'))) {
         if ((c == '\\') && (buf[1] == '"')) {
@@ -479,8 +482,8 @@ std::vector<std::string> utils::split_words(const std::string &text)
         c = *++buf;
         ++len;
       }
+      if (c != '"') ++len;
       c = *++buf;
-      ++len;
     }
 
     // unquoted
@@ -496,7 +499,7 @@ std::vector<std::string> utils::split_words(const std::string &text)
       if ((c == ' ') || (c == '\t') || (c == '\r') || (c == '\n')
           || (c == '\f') || (c == '\0')) {
           list.push_back(text.substr(beg,len));
-          beg += len;
+          beg += len + add;
           break;
       }
       c = *++buf;
diff --git a/unittest/utils/test_utils.cpp b/unittest/utils/test_utils.cpp
index 2a634dd053..22a647f460 100644
--- a/unittest/utils/test_utils.cpp
+++ b/unittest/utils/test_utils.cpp
@@ -21,13 +21,14 @@
 #include <vector>
 
 using namespace LAMMPS_NS;
-using ::testing::Eq;
 using ::testing::EndsWith;
+using ::testing::Eq;
+using ::testing::StrEq;
 
 TEST(Utils, trim_comment)
 {
     auto trimmed = utils::trim_comment("some text # comment");
-    ASSERT_THAT(trimmed, Eq("some text "));
+    ASSERT_THAT(trimmed, StrEq("some text "));
 }
 
 TEST(Utils, count_words)
@@ -59,24 +60,36 @@ TEST(Utils, split_words_simple)
 {
     std::vector<std::string> list = utils::split_words("one two three");
     ASSERT_EQ(list.size(), 3);
+    ASSERT_THAT(list[0], StrEq("one"));
+    ASSERT_THAT(list[1], StrEq("two"));
+    ASSERT_THAT(list[2], StrEq("three"));
 }
 
 TEST(Utils, split_words_quoted)
 {
     std::vector<std::string> list = utils::split_words("one 'two' \"three\"");
     ASSERT_EQ(list.size(), 3);
+    ASSERT_THAT(list[0], StrEq("one"));
+    ASSERT_THAT(list[1], StrEq("two"));
+    ASSERT_THAT(list[2], StrEq("three"));
 }
 
 TEST(Utils, split_words_escaped)
 {
     std::vector<std::string> list = utils::split_words("1\\' '\"two\"' 3\\\"");
     ASSERT_EQ(list.size(), 3);
+    ASSERT_THAT(list[0], StrEq("1\\'"));
+    ASSERT_THAT(list[1], StrEq("\"two\""));
+    ASSERT_THAT(list[2], StrEq("3\\\""));
 }
 
 TEST(Utils, split_words_quote_in_quoted)
 {
     std::vector<std::string> list = utils::split_words("one 't\\'wo' \"th\\\"ree\"");
     ASSERT_EQ(list.size(), 3);
+    ASSERT_THAT(list[0], StrEq("one"));
+    ASSERT_THAT(list[1], StrEq("t\\'wo"));
+    ASSERT_THAT(list[2], StrEq("th\\\"ree"));
 }
 
 TEST(Utils, valid_integer1)
@@ -334,13 +347,13 @@ TEST(Utils, strmatch_whitespace_nonwhitespace)
 TEST(Utils, guesspath)
 {
     char buf[256];
-    FILE *fp = fopen("test_guesspath.txt","w");
+    FILE *fp = fopen("test_guesspath.txt", "w");
 #if defined(__linux__)
-    const char *path = utils::guesspath(buf,sizeof(buf),fp);
-    ASSERT_THAT(path,EndsWith("test_guesspath.txt"));
+    const char *path = utils::guesspath(buf, sizeof(buf), fp);
+    ASSERT_THAT(path, EndsWith("test_guesspath.txt"));
 #else
-    const char *path = utils::guesspath(buf,sizeof(buf),fp);
-    ASSERT_THAT(path,EndsWith("(unknown)"));
+    const char *path = utils::guesspath(buf, sizeof(buf), fp);
+    ASSERT_THAT(path, EndsWith("(unknown)"));
 #endif
     fclose(fp);
 }