Dale Johannesen wrote: > This is wrong, because 4 bytes starting at 73 goes outside the original > object and can > cause a page fault.
FYI You can write a testcase for this by using mmap to allocate a page of memory, putting a copy of the structure 76 bytes from the end of the mmapped region, and then trying to access the field b. Having a testcase will help ensure that the bug is fixed and stays fixed. There is a existing similar testcase in gcc.dg/20030711-1.c that you can use as a model. I haven't tried looking at the bug yet. -- Jim Wilson, GNU Tools Support, http://www.specifix.com
