common,w32: Refine the command line parsing for \ in quotes.

* common/t-w32-cmdline.c (test_all): Add new test cases.
* common/w32-misc.c (strip_one_arg): Add arg endquote.
(parse_cmdstring): Take care of backslashes in quotes.
--

I found some new test vectors from Microsoft.

2 files changed, 33 insertions, 6 deletions

diff --git a/common/t-w32-cmdline.c b/common/t-w32-cmdline.c
index 8686a376a..172489c70 100644
--- a/common/t-w32-cmdline.c
+++ b/common/t-w32-cmdline.c
@@ -59,8 +59,18 @@ test_all (void)
     { "a\\\\\\b d\"e f\"g h", 3, { "a\\\\\\b", "de fg", "h" }},
     { "a\\\\\\\"b c d",       3, { "a\\\"b",   "c",     "d" }},
     { "a\\\\\\\\\"b c\" d e", 3, { "a\\\\b c", "d",     "e" }},
+    /* Examples from "Parsing C Command-Line Arguments" dated 11/09/2020.
+     * https://docs.microsoft.com/en-us/cpp/c-language/\
+     * parsing-c-command-line-arguments?view=msvc-160
+     */
+    { "\"a b c\" d e",          3, { "a b c",    "d",     "e" }},
+    { "\"ab\\\"c\" \"\\\\\" d", 3, { "ab\"c",    "\\",    "d" }},
+    { "a\\\\\\b d\"e f\"g h",   3, { "a\\\\\\b", "de fg", "h" }},
+    { "a\\\\\\\"b c d",         3, { "a\\\"b",   "c",     "d" }},
+    { "a\\\\\\\\\"b c\" d e",   3, { "a\\\\b c", "d",     "e" }},
+    { "a\"b\"\" c d",           1, { "ab\" c d"               }},
     /* Some arbitrary tests created using mingw.
-     * But I am nire sure whether their parser is fully correct.
+     * But I am not sure whether their parser is fully correct.
      */
     { "e:a  a b\"c\" ",       3, { "e:a", "a", "bc" }},
     /* { "e:a  a b\"c\"\" d\"\"e \" ", */
diff --git a/common/w32-misc.c b/common/w32-misc.c
index 8aef12e4a..ae194facb 100644
--- a/common/w32-misc.c
+++ b/common/w32-misc.c
@@ -46,7 +46,7 @@ count_backslashes (const char *s)
 
 
 static void
-strip_one_arg (char *string)
+strip_one_arg (char *string, int endquote)
 {
   char *s, *d;
   unsigned int n, i;
@@ -62,6 +62,12 @@ strip_one_arg (char *string)
             if ((n&1)) /* Odd number of backslashes.  */
               *d++ = '"';  /* Print the quote.  */
           }
+        else if (!s[n] && endquote)
+          {
+            for (i=0; i < n/2; i++)
+              *d++ = '\\';
+            s--;
+          }
         else /* Print all backslashes.  */
           {
             for (i=0; i < n; i++)
@@ -94,10 +100,21 @@ parse_cmdstring (char *string, char **argv)
         {
           if (*p == '\\' && p[1] == '"')
             p++;
+          else if (*p == '\\' && p[1] == '\\')
+            p++;
           else if (*p == '"')
             {
-              if (argv && (p[1] == ' ' || p[1] == '\t' || !p[1]))
-                *p = 0;
+              if (p[1] == ' ' || p[1] == '\t' || !p[1])
+                {
+                  if (argv)
+                    {
+                      *p = 0;
+                      strip_one_arg (p0, 1);
+                      argv[argc] = p0;
+                    }
+                  argc++;
+                  p0 = NULL;
+                }
               inquote = 0;
             }
         }
@@ -126,7 +143,7 @@ parse_cmdstring (char *string, char **argv)
               if (argv)
                 {
                   *p = 0;
-                  strip_one_arg (p0);
+                  strip_one_arg (p0, inquote);
                   argv[argc] = p0;
                 }
               argc++;
@@ -144,7 +161,7 @@ parse_cmdstring (char *string, char **argv)
       if (argv)
         {
           *p = 0;
-          strip_one_arg (p0);
+          strip_one_arg (p0, inquote);
           argv[argc] = p0;
         }
       argc++;