From: Bob Lechner <lechner@cs.uml.edu>
Subject: gen/ver_10 actions TBD - notes to me 
To: lechner@saturn.cs.uml.edu (Bob Lechner)
Date: Fri, 17 Apr 1998 22:28:23 -0400 (EDT)
RJLRef: $CASE/gen/ver_10_extensions.hlp


I compared several versions of genv10 including 

A. $CASE/95f523/bdesrc/maldred/chgen 
Mark Aldred's changed gen/ver_10 for alpha machines and Makefiles.  
(grep maldred <anygenpath>/src/*.c)

On 98/4/17, I merged maldred's Makefile.alpha into 
	$CASE/gen/ver_10/chgen/src/Makefile,
under the condition that $MACHINE = 'alpha', on 98/4/17.
(I have not tested this or checked it into CVS.)

---------------------------------------------

B. $CASE/gen/ver_10log (from $CASE/96s523/gen/mmurphy)


------------------------------------------------

C. 96s523/bdesym/rcostell/chgen10/chgen10 
(the bdesym version of chgen is identified as version 10.2.

(Bdesym began with gen v9.5, later upgraded to genv10.2.)
Besides fixing minor bugs in genv10
bdesym adds code under the condition 'if (cli_log == TRUE)'.
This suggests that gen/ver_10log was the final 
target version of gen which they modified for bdesym.

Explanation of bdesym changes and use of gen/ver_10:
(genv10 is in $CASE/gen/ver_10log but not checked into CVS):

$CASE/96s523/bdesym/rcostell/temp/* includes 4 help files
on bdesym's changes to pr_set/get_str and how tokstr and symref
representations are made transparent to the programmer.
This happens if is_key is 'x' or 'y' in the schema.

Tables TT and TA are navigated at runtime to decide which 
representation will be use for the string.
Therefore, gen/ver_10 is used and TT-->TA is loaded at runtime.
(gen/v9.5 onward produces TT-->TA metadata.

---------------------------------------------------

D. $CASE/ver_10dfltSuper - 
(moved from $CASE/96s523/gen/smoore) 

Recurses up the superclass hierarchy creating
superclass instances and inserting parent pointers
and fkeys on the way back, 
whenever a subclass instance is created.
 
-----------------------------------------------

E. $CASE/ver_10dfltRows - 

(moved from $CASE/96su523/gendfltrows, which 
followed 96s523/genlog (now in ver_10log).
THis makes altName field of TA into dfltVal field,
a metaschema change with little impact since
the altName field was unused before.
TT,TA metadata is available to check default value
type constraints but this is not implemented yet.

-----------------------------------------------

F. $CASE/gen/97sgenmerge - (an aareadme file 
moved here from the $CASE/97s523/genmerge project)

This project placed gen/ver_8 under cvs control 
and checked in other ver_10 extensions 
as tagged branches but not on the main sequence
of revisions. 

Merging these revisions remains to be done.
I did not check the up-to-date-ness of the versions I moved
here vs. the versions checked in by genmerge.