[comp.mail.elm] mailbox lock on 4.3 BSD-tahoe

paul@uxc.cso.uiuc.edu (12/07/89)

Platform: Elm 2.2 PL 14 on VAX 3500 running 4.3 BSD-tahoe

Problem: I don't wish to run ELM either setuid or setgid as it's not needed
for mail box locking.  4.3 BSD-tahoe uses only flock().  Creation of .lock
files should be #ifdef'ed the same as LOCK_WITH_FLOCK.  A patch file will
be sent to the contact address once I finish the install and test for a
while.

         Paul Pomes

UUCP:     {att,iuvax,uunet}!uiucuxc!paul     ICBM: 40 06 47 N / 88 13 35 W
Internet, BITNET: paul@uxc.cso.uiuc.edu      Phone: 217 333 6262
US Mail:  UofIllinois, CSO, 1304 W Springfield Ave, Urbana, IL  61801-2987

rob@PacBell.COM (Rob Bernardo) (12/08/89)

In article <200600002@uxc.cso.uiuc.edu> paul@uxc.cso.uiuc.edu writes:
+Problem: I don't wish to run ELM either setuid or setgid as it's not needed
+for mail box locking.  4.3 BSD-tahoe uses only flock().  Creation of .lock
+files should be #ifdef'ed the same as LOCK_WITH_FLOCK.  A patch file will
+be sent to the contact address once I finish the install and test for a
+while.

The problem is that Configure cannot accurately tell if you system actually
*relies* on locking by flock(), only that your system *might use* flock().
Consequently, if ELM uses flock() it can't be sure that your system
doesn't need locking with .lock files.
-- 
Rob Bernardo      ...![backbone]!pacbell!pbhyf!rob -or- rob@pbhyf.PacBell.COM
  Product engineer, UNIX/C Reusable Code Library        Editor, "Go `C' UNIX"
  Office: (415) 823-2417                Pacific * Bell, San Ramon, California
  Residence: (415) 827-4301                     R BAR JB, Concord, California

jgd@csd4.csd.uwm.edu (John G Dobnick) (12/10/89)

From article <6542@pbhyf.PacBell.COM>, by rob@PacBell.COM (Rob Bernardo):
> In article <200600002@uxc.cso.uiuc.edu> paul@uxc.cso.uiuc.edu writes:
> +Problem: I don't wish to run ELM either setuid or setgid as it's not needed
> +for mail box locking.  4.3 BSD-tahoe uses only flock().  Creation of .lock
> +files should be #ifdef'ed the same as LOCK_WITH_FLOCK.  A patch file will
> +be sent to the contact address once I finish the install and test for a
> +while.
> 
> The problem is that Configure cannot accurately tell if you system actually
> *relies* on locking by flock(), only that your system *might use* flock().
> Consequently, if ELM uses flock() it can't be sure that your system
> doesn't need locking with .lock files.

Which is why, once again, I suggest that the installer/configurator
be given the *option* of selecting either one or the other or both
methods of locking.  Even is Configure can't, with certainty, figure 
out whether flock() is relied upon, some human at the site may know.
Allowance should be made for these cases.

Thank you,
-- 
John G Dobnick
Computing Services Division @ University of Wisconsin - Milwaukee
INTERNET: jgd@csd4.csd.uwm.edu
UUCP: uunet!uwm!csd4.csd.uwm.edu!jgd

"Knowing how things work is the basis for appreciation,
and is thus a source of civilized delight."  -- William Safire