[prev in list] [next in list] [prev in thread] [next in thread] 

List:       apr-dev
Subject:    [BUG?] apr_proc_other_child_refresh passes wrong status flavor to maintenance function
From:       "Jeff Trawick" <trawick () gmail ! com>
Date:       2007-01-24 14:44:22
Message-ID: cc67648e0701240644l76e26b6fm1ee906f05e15e7c1 () mail ! gmail ! com
[Download RAW message or body]

In this code, a Unix-style signum|exitcode|bits is passed to the
maintenance function from within apr_proc_other_child_refresh().

    waitret = waitpid(ocr->proc->pid, &status, WNOHANG);
    if (waitret == ocr->proc->pid) {
        ocr->proc = NULL;
        (*ocr->maintenance) (APR_OC_REASON_DEATH, ocr->data, status);
    }

But for other calls to the maintenance function with
APR_OC_REASON_DEATH we expect that a signum_or_exitcode is passed to
the maintenance function (API breakage mentioned separately via
http://svn.apache.org/viewvc?view=rev&revision=499417).  It doesn't
make sense to pass a Unix-style status here.

Thoughts?
[prev in list] [next in list] [prev in thread] [next in thread]