RE: Loop code

JT,

Just pointing out something here....  I hear programmers getting confused
about this, and (IMO) usually it's because they never paid very close
attention to what the indicators are really representing.  Take a look at
CHAIN in the good ol' ILE RPG Reference, and you see the indicators columns
are labeled NR (no record), ER (error), and --.  READ, and all it's cousins,
labels them as --, ER, and EOF.  SETLL is NR, ER, and EQ.  

In practice, I think most RPG programmers just wait for SEU to tell them
they need an indicator wherever SEU placed the cursor, and never really
develop a feel for what the indicators are really telling them.  Its just
simply "there" (indicator off) or "not there" (indicator on).  

As I recall, some of these "rules" about the compiler setting off the %eof
indicator after a chain was PTFd recently, because too many programmers
missed out on how to use the new BIFs.

Eric DeLong
Sally Beauty Company
MIS-Project Manager (BSG)
940-898-7863 or ext. 1863



-----Original Message-----
From: jt@xxxxxx [mailto:jt@xxxxxx]
Sent: Monday, March 29, 2004 5:57 PM
To: RPG programming on the AS400 / iSeries
Subject: RE: Loop code


Thx Paul,

But:  For the same reason that a CHAIN, or a READ, or a READE, or a READC
sets the indicator(s) on or off.  Iow, every single one of these opcodes
either sets an indicator on or off.  (And, not coincidently, it works the
same way if you're using the < indicator for errors on these opcodes, and in
fact, every single opcode I can think of off the top of my head.)  To me,
intuitively, it would make as much sense to leave %EOF *unchanged* after
code is executed, as it would to leave the indicator specified on a LOOKUP
*unchanged*, unless there was a match and ONLY THEN you set the indicator
on.  Otherwise it just stays the same it was.  Now THAT makes no sense to me
(and I would suspect most-all everybody who has actually used RPG, although
there will be exceptions 'course).

So I'm afraid I don't see a close relationship at all between how a FILE
POINTER is used, but a great deal of relationship (like near-identical)
between indicators and the %EOF built-in.  Except that's not how the %EOF
works.  So, as far as making "perfect sense", very obviously this is not the
case.  And I'm not even sure it makes even a LITTLE sense, from the
discussion I've seen so far.

(Btw, I never found out if you can set these built-ins, like you can
indicators, in DBG??)

| -----Original Message-----
| [mailto:rpg400-l-bounces@xxxxxxxxxxxx]On Behalf Of Paul Tuohy
| Sent: Monday, March 29, 2004 5:07 PM

| Sorry jt, but I think this makes perfect sense.
|
| If a CHAIN is successful, it means you got a record, so you are
| not at EOF,
| so %EOF(filename) is set off.
|
| If a CHAIN is unsuccessful, the file pointer is left where it was, so
| %EOF(filename) is unchanged.
|
| An unsuccessful chain does not position the file pointer at EOF - so why
| would it set on %EOF?
|
| Paul Tuohy
|
| ----- Original Message -----
| From: "jt" <jt@xxxxxx>
| Sent: Friday, March 26, 2004 7:35 PM

| > This part seems ENTIRELY UNintuitive.  Is there some benefit or
| Reason WHY
| > it would NOT change??  What logic went into this decision, if I may ask?
| > (Obviously, I may ask, but it might never get posted nor answered, of
| > course...;-)
| >
| > | -----Original Message-----
| > | [mailto:rpg400-l-bounces@xxxxxxxxxxxx]On Behalf Of Paul Tuohy
| > | Sent: Friday, March 26, 2004 12:37 PM
| >
| > | "The following operations, if successful, set %EOF(fielname)
| off. If the
| > | operation is not successful, %EOF(filename) is not changed."



_______________________________________________
This is the RPG programming on the AS400 / iSeries (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/rpg400-l.