RE: "Reference #"

[qsrvbas@xxxxxxxxxxxx (Tom Liotta)]

Buck:

Yep, and that's an added reason why I wouldn't use it as a unique identifier in 
the first place as Scott and one or two others also mentioned. I'd almost 
certainly go with a UUID or similar to begin with. And if I really _needed_ a 
sequential number, I'd likely create a server-type function, perhaps receiving 
requests and serving responses via keyed data queues, or implement (quite 
possibly service program) procedures with maybe a core procedure like:

pgm     ( +
          &pqSpcNam   +
          &pIncVal    +
        )

   dcl    &pqSpcNam       *char   20
   dcl    &pIncVal        *dec (  15 0 )


   dcl    &SpcNam         *char   10
   dcl    &SpcLib         *char   10


   chgvar    &SpcNam      %sst( &pqSpcNam  1 10 )
   chgvar    &SpcLib      %sst( &pqSpcNam 11 10 )

   alcobj    obj( ( &SpcLib/&SpcNam  +
               *usrspc *excl ) )

   chgvar    &pIncVal   ( &pIncVal + 1 )

   dlcobj    obj( ( &SpcLib/&SpcNam  +
               *usrspc *excl ) )

   return

endpgm

...and a few related procedures to create the space and initialize when it 
doesn't exist, to retrieve a pointer to it and to call this procedure passing 
the pointer in place of &pIncVal in the parm list, and you have a fairly small, 
reasonably easy to maintain sequential number generator that's also pretty 
quick. Obviously, add proper error handling and choose how to handle lock 
timeouts.

A 15-digit packed number will go up to about a quadrillion before needing to be 
rolled over, so it could be used just about anywhere in a system. But if that 
ain't enough, then bump the pointer by 8 and create a new number in the same 
space. I suppose you could have a description field followed by a number 
followed by a description field followed by... etc., if you need to get fancy, 
and it should still be pretty fast.

ILE CL is handy for some procedures. In this case, the ALCOBJ and DLCOBJ are 
what is really used. Of course, since the lock(), unlock() and similar APIs are 
available, there's no real use for ILE CL in this except for concept.

Tom Liotta

rpg400-l-request@xxxxxxxxxxxx wrote:

>   4. RE: "Reference #" (Buck Calabro)
>
>>this does make me think that even when time-change 
>>does apply, it might be helped by simply appending the UTC 
>>offset system value to the string. Somebody might make use of it.
>
>When people set the system value.  ICT poll shows that 3 of 15 people have
>it set to +00:00.  Doesn't invalidate the idea though, just limits the
>implementation.

-- 
-- 
Tom Liotta
The PowerTech Group, Inc.
19426 68th Avenue South
Kent, WA 98032
Phone  253-872-7788 x313
Fax    253-872-7904
http://www.powertechgroup.com


__________________________________________________________________
Try AOL and get 1045 hours FREE for 45 days!
http://free.aol.com/tryaolfree/index.adp?375380

Get AOL Instant Messenger 5.1 for FREE! Download Now!
http://aim.aol.com/aimnew/Aim/register.adp?promos=380455