Author: Chris Guffey
Date: 12/5/2006 - RJL Rev. 061205,061211
RJLRef: $CASE/06f522/asgnt3_report_cg_rl061211.txt
[Was ./cguffey/asgnt3/assignment3_report.txt]

Work accomplished:
1) Spent lots of time in group discussion and exchanging emails trying to clarify the functionality and scope of the project
2) Implemented code for keyValueListExists and associated functions (see code history below)
3) Worked with Alison to resolve compilation and runtime errors


Revision History:
file: $case/06f522/cguffey/asgnt3/assignment3.c
changes:
Initial attempt at coding keyValueListExists. Won't compile and only half implemented.


file: $case/06f522/cguffey/asgnt3/assignment3_2.c
changes:Introduced functions ExtractValues(...) and DetermineColType(...), 


file: $case/06f522/cguffey/asgnt3/assignment3_3.c
changes:
Introduced functions doesDEKeyExist(...), GetValueFromDECurr(...) and InitializeTableAbbr(...). This was to modularize the code, separate concerns and make it more readable to the human eye. Removed function DetermineColType(...).


file: $case/06f522/cguffey/asgnt3/assignment3_4.c
changes:
- Created a linked list of structs that hold fname and fvalue
characteristics of our CK
- rewrote ExtractValue to use hcg_parse
- pValueList is now only parsed once, not during every new DE row
- reduced the number of nested loops in DoesDEKeyExist(...), because we extract the CK info before hand now we no longer needed the inner 2 loops, they were replaced with one less costly loop


file: $case/06f522/cguffey/asgnt3/assignment3_5.c
changes:
- Fixed assorted compilation errors


[RJL> Next change (below) may be a candidate for GVars extension: 
The CA relation relates TA ftype with CK fname; CQ  holds fvalue in ASCII. 
CA like TA is metadata (has CK not CQ fkey) so cannot hold 
fvalue (encoded OR ASCII).  CQ COULD hold encoded as well as 
ASCII value for validation use during  search loop.
<RJL]

file: $case/06f522/cguffey/asgnt3/assignment3_6.c
changes:
- stricter error checking
- speed inprovement by removing reliance on string compares for non string field types, there is an increase in disk space cost directly proprotional to the number of CAs associated with the CK, this should be negligible especially considering we're only dealing with one CK at a time. This included reintroducing the ftype check from version 2
- figured out the malloc issue from last night and fixed that, I
merely needed to type cast the malloc result
- merged in RJLs modifications
- fixed compilation error found by Alison


file: $case/06f522/cguffey/asgnt3/assignment3_7.c
changes:
- Removed the TTabbrev input arg from keyValueListExists since it was redundant once we find the TTcurr. Other functions modifed as appropriate to reflect this.


file: $case/06f522/cguffey/asgnt3/assignment3_8.c
changes:
- Removed table_loop_all(TT) in keyValueListExists, TTcurr can simply be assigned by TTcurr = CKcurr->TTid_pp;

file: $case/06f522/cguffey/asgnt3/assignment3_9.c
changes:
- Alison found a bug where pFieldList was not being properly initialized. A simple modification in the function ExtractValues fixes this.