In <[EMAIL PROTECTED]>, on 02/16/2007
at 08:03 AM, Paul Gilmartin <[EMAIL PROTECTED]> said:
>This "hole" has existed in the Classic file system since OS/360.
>(Shmuel might correct my conjecture about history). Likewise there's
>the venerable pitfall of allocating a PDS and failing to specify a
>member, thus overwriting the directory.
We lost a proclib that way in SVS. That prompted be to write an
IEBGENER alternative that would not allow a PDS name without a member
on SYSUT2. There was also the problem of an explicit DCB parameter
changing the attributes of an existing PDS, causing use of existing
members to fail.
>Rightly, both holes should be closed once and for all (perhaps
>leaving a back door to allow reading a PDS directory sequentially).
I wouldn't call that leaving a hole; OPEN should and does treat input
and output differently.
>But the fix should be done not haphazardly by individual
>applications, each applying its developers' idiosyncratic concept of
>what the rule should be, but uniformly, in OPEN code.
That must be done carefully, because it is legitimate to omit the
member name when the application program will supply it.
>For the PDS/PS fix, Rexx is an individual application that gets it
>right: It fails the attempt to open a PDS with no member specified,
>but allows it if the programmer allocates with overriding DSORG(PS)
>(the back door).
That sounds like it's still too easy to inadvertently destroy a
directory.
--
Shmuel (Seymour J.) Metz, SysProg and JOAT
ISO position; see <http://patriot.net/~shmuel/resume/brief.html>
We don't care. We don't have to care, we're Congress.
(S877: The Shut up and Eat Your spam act of 2003)
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
