Merge pull request #2708 from mesonbuild/nirbheek/windows-shebang-parsing

dependencies: Fix parsing of shebangs with spaces

mesonbuild/dependencies/base.py

@@ -560,7 +560,11 @@ class ExternalProgram:
             with open(script) as f:
                 first_line = f.readline().strip()
             if first_line.startswith('#!'):
-                commands = first_line[2:].split('#')[0].strip().split()
+                # In a shebang, everything before the first space is assumed to
+                # be the command to run and everything after the first space is
+                # the single argument to pass to that command. So we must split
+                # exactly once.
+                commands = first_line[2:].split('#')[0].strip().split(maxsplit=1)
                 if mesonlib.is_windows():
                     # Windows does not have UNIX paths so remove them,
                     # but don't remove Windows paths

test cases/common/167 external program shebang parsing/input.txt

@@ -0,0 +1 @@
+some stuff here

test cases/common/167 external program shebang parsing/main.c

@@ -0,0 +1,72 @@
+#include <stdio.h>
+#include <fcntl.h>
+#include <errno.h>
+#include <string.h>
+#include <stdlib.h>
+#include <sys/types.h>
+
+#ifdef _WIN32
+ #include <io.h>
+ #include <windows.h>
+#else
+ #include <unistd.h>
+#endif
+
+/* Who cares about stack sizes in test programs anyway */
+#define LINE_LENGTH 4096
+
+static int
+intrp_copyfile (char * src, char * dest)
+{
+#ifdef _WIN32
+  if (!CopyFile (src, dest, FALSE))
+    return 1;
+  return 0;
+#else
+  return execlp ("cp", "copyfile", src, dest, NULL);
+#endif
+}
+
+static void
+parser_get_line (FILE * f, char line[LINE_LENGTH])
+{
+  if (!fgets (line, LINE_LENGTH, f))
+    fprintf (stderr, "%s\n", strerror (errno));
+}
+
+int
+main (int argc, char * argv[])
+{
+  FILE *f = NULL;
+  char line[LINE_LENGTH];
+
+  if (argc != 4) {
+    fprintf (stderr, "Invalid number of arguments: %i\n", argc);
+    goto err;
+  }
+
+  if ((f = fopen (argv[1], "r")) == NULL) {
+    fprintf (stderr, "%s\n", strerror (errno));
+    goto err;
+  }
+
+  parser_get_line (f, line);
+
+  if (!line || line[0] != '#' || line[1] != '!') {
+    fprintf (stderr, "Invalid script\n");
+    goto err;
+  }
+
+  parser_get_line (f, line);
+
+  if (!line || strncmp (line, "copy", 4) != 0) {
+    fprintf (stderr, "Syntax error: %s\n", line);
+    goto err;
+  }
+
+  return intrp_copyfile (argv[2], argv[3]);
+
+err:
+  fclose (f);
+  return 1;
+}

test cases/common/167 external program shebang parsing/meson.build

@@ -0,0 +1,21 @@
+project('shebang parsing', 'c')
+
+interpreter = executable('aninterp', 'main.c', native : true)
+
+cdata = configuration_data()
+cdata.set('INTRP', interpreter.full_path())
+
+f = configure_file(input : 'script.int.in',
+                   output : 'script.int',
+                   configuration : cdata)
+
+# Test that parsing a shebang with spaces works properly. See `man execve`,
+# specifically the section on "Interpreter scripts" and the one under "NOTES".
+script = find_program(f)
+
+custom_target('interpthis',
+  input : 'input.txt',
+  output : 'output.txt',
+  depends : interpreter,
+  command : [script, '@INPUT@', '@OUTPUT@'],
+  build_by_default : true)

test cases/common/167 external program shebang parsing/script.int.in

@@ -0,0 +1,2 @@
+#!/usr/bin/env @INTRP@
+copy