environ.cc: New facility/environment variable MSYS2_ENV_CONV_EXCL

Works very much like MSYS2_ARG_CONV_EXCL. In fact it uses the same
function, arg_heuristic_with_exclusions (). Also refactors parsing
the env. variables to use new function, string_split_delimited ().

The env. that is searched through is the merged (POSIX + Windows)
one. It remains to be seen if this should be made an option or not.

This feature was prompted because the R language (Windows exe) calls
bash to run configure.win, which then calls back into R to read its
config variables (LOCAL_SOFT) and when this happens, msys2-runtime
converts R_ARCH from "/x64" to an absolute Windows path and appends
it to another absolute path, R_HOME, forming an invalid path.

Author: mingwandroid

winsup/cygwin/environ.cc

@@ -1130,6 +1130,10 @@ build_env (const char * const *envp, PWCHAR &envblock, int &envc,
 
   int tl = 0;
   char **pass_dstp;
+#ifdef __MSYS__
+  char *msys2_env_conv_excl_env = NULL;
+  size_t msys2_env_conv_excl_count = 0;
+#endif
   char **pass_env = (char **) alloca (sizeof (char *)
 				      * (n + winnum + SPENVS_SIZE + 1));
   /* Iterate over input list, generating a new environment list and refreshing
@@ -1138,9 +1142,18 @@ build_env (const char * const *envp, PWCHAR &envblock, int &envc,
     {
       bool calc_tl = !no_envblock;
 #ifdef __MSYS__
-      /* Don't pass timezone environment to non-msys applications */
-      if (!keep_posix && ascii_strncasematch(*srcp, "TZ=", 3))
-        goto next1;
+      if (!keep_posix)
+      {
+	/* Don't pass timezone environment to non-msys applications */
+	if (ascii_strncasematch(*srcp, "TZ=", 3))
+	  goto next1;
+	else if (ascii_strncasematch(*srcp, "MSYS2_ENV_CONV_EXCL=", 20))
+	  {
+	    msys2_env_conv_excl_env = (char*)alloca (strlen(&(*srcp)[20])+1);
+	    strcpy (msys2_env_conv_excl_env, &(*srcp)[20]);
+	    msys2_env_conv_excl_count = string_split_delimited (msys2_env_conv_excl_env, ';');
+	  }
+      }
 #endif
       /* Look for entries that require special attention */
       for (unsigned i = 0; i < SPENVS_SIZE; i++)
@@ -1264,7 +1277,8 @@ build_env (const char * const *envp, PWCHAR &envblock, int &envc,
 	    }
 #ifdef __MSYS__
 	  else if (!keep_posix) {
-	    char *win_arg = arg_heuristic(*srcp);
+	    char *win_arg = arg_heuristic_with_exclusions
+		   (*srcp, msys2_env_conv_excl_env, msys2_env_conv_excl_count);
 	    debug_printf("WIN32_PATH is %s", win_arg);
 	    p = cstrdup1(win_arg);
 	    if (win_arg != *srcp)

winsup/cygwin/miscfuncs.cc

@@ -307,6 +307,26 @@ NT_readline::gets ()
     }
 }
 
+/* Searches through string for delimiter replacing each instance with '\0'
+   and returning the number of such delimited substrings. This function
+   Will return 0 for the NULL string and at least 1 otherwise. */
+
+size_t
+string_split_delimited (char * string, char delimiter)
+{
+  if ( string == NULL )
+      return 0;
+  size_t count = 1;
+  string = strchr ( string, delimiter );
+  while (string)
+    {
+      *string = '\0';
+      ++count;
+      string = strchr ( string + 1, delimiter );
+    }
+  return count;
+}
+
 /* Return an address from the import jmp table of main program.  */
 void * __reg1
 __import_address (void *imp)

winsup/cygwin/miscfuncs.h

@@ -82,6 +82,8 @@ void backslashify (const char *, char *, bool);
 void slashify (const char *, char *, bool);
 #define isslash(c) ((c) == '/')
 
+size_t string_split_delimited (char * string, char delimiter);
+
 extern void transform_chars (PWCHAR, PWCHAR);
 extern inline void
 transform_chars (PUNICODE_STRING upath, USHORT start_idx)

winsup/cygwin/path.cc

@@ -3448,7 +3448,6 @@ arg_heuristic_with_exclusions (char const * const arg, char const * exclusions,
     return arg_result;
   }
 
-  debug_printf("Input value: (%s)", arg);
   for (size_t excl = 0; excl < exclusions_count; ++excl)
     {
       /* Since we've got regex linked we should maybe switch to that, but

winsup/cygwin/spawn.cc

@@ -262,8 +262,7 @@ child_info_spawn::worker (const char *prog_arg, const char *const *argv,
   int res = -1;
 
   /* Environment variable MSYS2_ARG_CONV_EXCL contains a list
-     of ';' separated argument prefixes to pass un-modified..
-     It isn't applied to env. variables; only spawn arguments.
+     of ';' separated argument prefixes to pass un-modified.
      A value of * means don't convert any arguments. */
   char* msys2_arg_conv_excl_env = getenv("MSYS2_ARG_CONV_EXCL");
   char* msys2_arg_conv_excl = NULL;
@@ -272,14 +271,7 @@ child_info_spawn::worker (const char *prog_arg, const char *const *argv,
     {
       msys2_arg_conv_excl = (char*)alloca (strlen(msys2_arg_conv_excl_env)+1);
       strcpy (msys2_arg_conv_excl, msys2_arg_conv_excl_env);
-      msys2_arg_conv_excl_count = 1;
-      msys2_arg_conv_excl_env = strchr ( msys2_arg_conv_excl, ';' );
-      while (msys2_arg_conv_excl_env)
-        {
-          *msys2_arg_conv_excl_env = '\0';
-          ++msys2_arg_conv_excl_count;
-          msys2_arg_conv_excl_env = strchr ( msys2_arg_conv_excl_env + 1, ';' );
-        }
+      msys2_arg_conv_excl_count = string_split_delimited (msys2_arg_conv_excl, ';');
     }
 
   /* Check if we have been called from exec{lv}p or spawn{lv}p and mask