pike.git/ src/ testsuite.in

2008-06-24

2008-06-24 14:48:45 by Martin Stjernholm <mast@lysator.liu.se>

• f0a7654d554f2124fcab35eedb866b9d3d3f0951 (18 lines) (+17/-1) ^{[ Show | Annotate ]}
  Branch: 7.9

Added a test about a strict_types warning that imho is a bit
overzealous (in any case it's odd).

Also added a note about a more unsettling issue where tests affect
each other.

Rev: src/testsuite.in:1.826

1:   START_MARKER - test_true([["$Id: testsuite.in,v 1.825 2008/06/13 21:14:28 marcus Exp $"]]); + test_true([["$Id: testsuite.in,v 1.826 2008/06/24 14:48:45 mast Exp $"]]);      // This triggered a bug only if run sufficiently early.   test_compile_any([[#pike 7.2]])
664:    function(string,...:int) foo;   ]]);    + test_compile_any([[ +  #pragma strict_types +  void x (void|mapping(string:string) args) +  { +  if (args) +  foreach (indices (args), string a) +  werror ("%O\n", a); +  } + ]]) +  + // For some reason, the test above doesn't generate the warning (about + // the type from indices (args)) if it comes after the one below. + // Also, if it is repeated it doesn't generate the warning the second + // time. Not very reassuring.. /mast +    test_any([[return class Z {   string bonk() { return "oiff"; }