F$GETDVI(<>,"DEVLOCKNAM") strangeness on V7.3-2

[Robert_Boyd]

It appears that one or more patches cause F$GETDVI to return different types of values.

One one system the value returned is a decent and orderly string with the volume label used for the lock name.

On several other systems with most of the latest patch kits installed, the lexical returns a hexadecimal string with a bunch of "00"s and all the characters when decoded are in reverse order.

What kind of values do you get on your systems?

Which behavior is the "correct" one?

Robert

[Robert Brooks_1 · Accepted solution]

This is a weird one for which I take full credit and/or blame, depending on your point of view!

While doing some $GETDVI work, I noticed that f$getdvi returned a hex string for dvi$_devlocknam. In looking back, I found that for whatever reason, that was the behaviour since its introduction in the early 80's. SYS$GETDVI returns "a decent and orderly string". I "fixed" f$getdvi to return a similar string for V8.2 and for V7.3-2 in one of the SYS/DCL kits (it would take both kits to cause the change).

After the fact, however, we decided that while bizarre and unexpected, we should restore the older behaviour in case there was anyone who had noticed the weird output and had coded around it (there was at least one person in HP who was annoyed that I had "fixed" it). There was also the issue of compatibility within a cluster with VAXes, where the hex string would be returned.

So, a later V7.3-2 SYS/DCL kit will restore the older, weird behaviour. I'm not sure which one, but certainly the latest SYS/DCL kits will return the hex string format.

Note that LIB$GETDVI will follow the F$GETDVI behaviour and not the SYS$GETDVI behaviour, although it's possible that no LIBRTL kits shipped with my "fixed" version.

It's my opinion that my initial change should have been retained, but I was in the minority.

Sorry!

-- Rob

[Robert_Boyd]

Rob,

Thanks for the detailed explanation -- I just today posted a procedure to dcl.OpenVMS.org which prints all F$GETDVI items. I had just fixed up the code to interpret the hexified string and found a couple of systems that didn't work on.

I'm glad there's such an easy explanation :-)
Too bad the people affected couldn't just live with having to detect which format the result was in.

I'll post another version that will be able to display from either form of result.

Robert

[Ian Miller.]

Robert's DCL procedure is at
http://dcl.openvms.org/stories.php?story=06/02/01/2480164