CONTACTS 

SUBJECTS
0702051 F2 Save SDS Records Out of Memory Updating Reference Pointer
Other Files Failing to Archive Preventing Finding Records Based on F

0504 -
0504 -    ..
0505 - Summary/Objective
0506 -
050601 - Follow up ref SDS 2 0000, ref SDS 1 EG8G.
050602 -
050603 - Fixed problem that was preventing searching for records where files
050604 - are listed.  The problem was with the zone parameter for macro 1451
050605 - tht sorts and removes duplicate keys.  The zone was 1 char too short
050606 - causing each line of reference files to seem like a duplicate that was
050607 - then deleted. ref SDS 0 8F9H
050608 -
050609 -
050610 -
050611 -
050612 -
050614 -  ..
0507 -
0508 -
0509 - Problem
0510 -
051001 - Seems like for the past few months that when we do a search for
051002 - records where a particular file was referenced, that the reports are
051003 - failing.  Today, investigation seems to suggest that work on 0702051
051004 - to archive pointers during the save op, is only getting one file from
051005 - "Other Files."  This is a bit of a mess that reduces the ability to
051006 - find things quickly.
051007 -
051009 -  ..
0511 -
0512 -
0513 - Background
0514 -
051401 - On 031128 work was done to process pointers in a separate memory
051402 - segment to reduce the impact of fragmenting that occurs when
051403 - performing a lot of different processes on a large SDS record.
051404 - ref SDS 1 EG8G
051406 -  ..
051407 - A few days later, on 031203 changed the design of the operation to
051408 - archive pointers, ref SDS 2 O54I, using a nested memory segment design
051409 - that processes Reference field data separately. ref SDS 2 3588
051410 -
051411 -    0702053 - not sure what this does just yet, but the explanation
051412 -    sounds suggestive of being the problem area today. ref SDS 2 VP4I
051414 -     ..
051415 -    See it now, the first step saves all of the pointer types in one
051416 -    file, they are processed by 0702053 in a separate memory segment.
051417 -    ref SDS 2 J699
051419 -     ..
051420 -    Looks like problem is with macro 1451 that sorts and eliminates
051421 -    duplicates.  It must be wired to delete all but one of the
051422 -    References, because 0702051 is working correctly to load all the
051423 -    files in References into the weekly pointer file.
051425 -     ..
051426 -    Macro 1451 was upgraded recently on 031217 to avoid having
051427 -    duplicates show up for the record descriptions in the Diary
051428 -    Summary.  The design for macro 1451 in 000008 calls for...
051429 -
051430 -         z 1 27 sets the span of cols to sort for the record ID and the
051431 -         primary line number key, which makes each line unique in SDS
051432 -         records. ref SDS 3 G94J
051434 -     ..
051435 -    This is an error.  z 1 27 misses the 6th position of the line
051436 -    number field which makes each line unique within a record, and so
051437 -    evaluating on col 5 results in deleting most of the references
051438 -    listed for files; so, changed this to z 1 28, reported on 031217.
051439 -    ref SDS 3 G94J
051441 -     ..
051442 -    This change solved the problem.
051443 -
051444 -
051445 -
051446 -
051447 -
051448 -
051449 -
051450 -
051451 -
051452 -
051453 -
0515 -