Re: diagnosing test failures

[Terry Lambert <terry@whistle.com>]

Rick Jones wrote:
> I've managed to get a system call trace of slapd that shows the source
> of the EBADF error coming from the berkeley db-2.7.7 when test003 is
> run:
> 
> write(2, "  ) \n", 3) .................................... = 3
> open("./test-db/dn2id.dbb", O_RDWR|O_CREAT|O_EXCL|O_LARGEFILE, 0600)
> ERR#17 EEXIST

Got an error because of O_CREAT, and the file exists, and
the tests do not clean up after themselves, and you didn't
clean up the files before running test 3.


> fcntl(-1, F_SETFD, 1) .................................... ERR#9 EBADF

Tried to keep using the "fd" returned; someone is apparently
not checking the open return code for a -1, and barging on
ahead as if everything is fine, when it isn't.

> under HP-UX at least if O_CREAT and O_EXCL are set on
> the open, and the file already exists, the open will
> fail with an EEXIST. i'm not sure if that is the
> behaviour on other platforms.

Only on POSIX compliant ones...


-- Terry Lambert
-- Whistle Communications, Inc., an I.B.M. Company
-- terry@whistle.com
-------------------------------------------------------------------
This is formal notice under California Assembly Bill 1629, enacted
9/26/98 that any UCE sent to my email address will be billed $50
per incident to the legally allowed maximum of $25,000.