home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!olivea!isc-br!tau-ceti!dejavu!salnick
- From: salnick@dejavu.spk.wa.us (There is no substitute for displacement...)
- Newsgroups: comp.databases.ingres
- Subject: Re: Help - ABF/4GL questions.
- Message-ID: <salnick.07zw@dejavu.spk.wa.us>
- Date: 7 Jan 93 13:52:58 GMT
- References: <00966285.BBAC44C0.3043@hal.hahnemann.edu> <1993Jan7.044714.10382@csc.canberra.edu.au>
- Organization: Not Organized
- Lines: 50
-
- In article <1993Jan7.044714.10382@csc.canberra.edu.au> mmr@asgserver.canberra.edu.au (Micheal Rainford-Watson) writes:
- >In article <00966285.BBAC44C0.3043@hal.hahnemann.edu> "Anthony J. Rzepela" <rzepela@hal.hahnemann.edu> writes:
- >>In response to a messy locking situation where the user
- >>GOLD@ILP.MIT.EDU (Mark D. Gold) wanted to avoid deadlock on
- >>a shared resource (a temporary table), it was pointed out
- >>by Mark Jaeger <cs_mj@gsbvax.uchicago.edu> that
- >>
- >>1> ABF has a neat lock detection and deletion feature.
- >>
- >>2> User-interrupts are not easily handled in the 4GL.
- >>
-
- It seems that several people have missed a key issue here: These
- temporary tables will be crerated in each of the users's contexts. As a
- consequence, they will be PRIVATE tables. No locking is required, since
- the tables are visible only to their creators, even if they all have the
- same name.
-
- If they are created in the context of the DBA, then I have to ask "Why is
- it being done this way???"
-
- bob
-
- >>In relation to the first matter, I have had the experience where
- >>an interrupted (^Y, ^C) session of ABF still exists as a lock,
- >>as the locks from these "dead" sessions show up until explicitly
- >>deleted by the new ABF user. (I am at 6.3/03 on VMS 5.4-2).
- >>
- >>What would be really slick (and I don't know about feasibility)
- >>would be to include in the application a user-maintined lock system, in which
- >>the application desiring a resource could check not only for one of these
- >>pseudo-locks, but also, if found, the status of the process that created it.
- >>(Is it still running, in a PFW, etc.?)
- >
- >... keep going ..
- >
- >And if the resource is being held by a user (for too long), they are
- >informed that someone is waiting for that resource and they should
- >'get on their bike!'.
- >
- >--
- >Micheal Rainford-Watson _--_|\ E-Mail: mmr@asgserver.canberra.edu.au
- >Administrative Systems Group / \ Phone : +61 6 201 2397
- >University of Canberra \_.--.*/ Mail : PO Box 1 BELCONNEN ACT 2616
- >^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^v^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-
- --
- RW Salnick, Spokane,WA | Home: salnick@dejavu.spk.wa.us
- Amiga 1000, WB 1.3 | Work: mfgsys!salnick@kaiser.spk.wa.us
- WA9BVE |
-