#1260 Stopgap fix for br916

closed
nobody
None
5
2013-02-04
2013-02-02
Martin Whitaker
No

As discussed in the bug report, as a stopgap measure, we need to reject any attempt to pass a value stored in thread memory to $strobe or $monitor (with a "sorry" message), as the thread memory may get reused or deallocated before the $strobe or $monitor actually executes.

I think this patch should catch all the cases that cause problems (and hopefully doesn't generate any false positives).

1 Attachments

Discussion

  • Here are a couple of regression tests to add to the test suite. Note that the gold files contain the expected results for when the bug is fixed properly - with the stopgap fix, both tests are expected to fail (gracefully) at run time.

     
    Attachments
  • Cary R.
    Cary R.
    2013-02-04

    • status: open --> closed
     
  • Cary R.
    Cary R.
    2013-02-04

    I have applied these patches and then made a second patch with minor changes.

    Specifically there was a space before a tab problem on one line, I reformatted the NOTE message, fixed a word problem "round" instead of "around" and set the global br916 variable so that the note is only displayed once since I believe that was the intended functionality.

    I also updated the test suite and results.