On Thu, 23 Jan 2003, Bill Stoddard wrote: > This patch fixes a segfault I see serving files cached by MMapFile. > First request is okay, subsequent request segfaults in an mmap ring > macro during apr_brigade_destroy because the next pointer is null.. I
Ahhhhhh... right you are. I can explain why if anybody cares. :) In addition to this patch, though, you should be able to remove the apr_mmap_dup(..., cmd->pool), as it's now redundant. Add that, and I'm +1 for this patch, it definitely looks correct. --Cliff
