Re: File attributes PF38, LF38, PF38-DTA, PF38-SRC

[Neil Palmer <neilp@xxxxxxxxxxx>]

John,

If they were normal simple logicals built over physical files with a
single member, then yes.  The problem is many of the logicals are join
logicals.  The physical files have multiple members.  Different logicals
are built over specific members in certain physical files, which need to
be specified on the CRTLF command, and some of the logicals have multiple
members, so it's a bit of work to dig all this up 15 years after it was
created to recreate all the files correctly.  ;-)

...Neil





"John Earl" <john.earl@powertechgroup.com>



        To:     <midrange-l@midrange.com>
        cc:
        Subject:        Re: File attributes PF38, LF38, PF38-DTA, PF38-SRC


Neil,

I'm just curious, but what happens if you take your LF38 source (that does
not have a PFILE library qualification) and recompile it into a temporary
library as an LF?

Will it use the existing access path?  I'd be surprised if it did not.

In which case it would be pretty fast and painless to compile all of the
LF38's as LF's (assuming proper library lists and PF to LF dependencies),
then delete the LF38's (or move them to a temporary holding library) and
move the LF's into the production library.  Sounds like it should take
about
2 minutes per LF, no?

This should be easy and painless to test.  And you'll know in minutes if
you
got it right.

(John's law:  Everything is easy if someone else has to do it :)

Does this sound reasonable to you?

jte


--
John Earl