RJLRef: $CASE/gen/genmerge/paleeFdbk.981212 (recovered from hung email)

Hi pat. Did you do 'man cvs' and read the command details? (extensive).
THere are two possibilities beside read access to group you belong to:
-------------------
jupiter.cs.uml.edu(56)> groups palee
grad 98f523 98s522
-------------------
The other two  possiblities:
(1) any side-branch checkout needs an explicit branch 'tag' to be checked out.
This is the likely cause of your trouble.
(2) you must give the path name of your checkout subtree  relative to $CVSROOT
(being above or below the point you shojld be results in a failed co).

E.g., here are two CVS ROOT symbols: (grep my ~/.login):
---------------
jupiter.cs.uml.edu(59)> echo $BDEROOT
/usr/proj3/case/95s523/95sbde/base/Master
jupiter.cs.uml.edu(60)> echo $GENROOT
/usr/proj3/case/gen/base/Master
-------------
E.g., below $BDEROOT is bde/{src, include, pr_util etc.}
To get src/bde.cc, you must do 'cvs co bde/src/bde.cc'
and be in the parent node where bde will be created.
Normally you do cvs co bde there and get the full tree,

Below $GENROOT is chgen with src, doc, etc. subdirectories:
-------------
jupiter.cs.uml.edu(91)> ls -d $GENROOT/chgen/{src,doc}
/usr/proj3/case/gen/base/Master/chgen/doc
/usr/proj3/case/gen/base/Master/chgen/src 
----------------
so to check out the chgen tree' main branch, latest rev., 
BELOW $PWD=$CWD, do 'cvs co chgen' there.
To check out a side branch, a branch tag is needed
(similar to checking out an  earlier rev.: 
e.g., 'cvs co -r1.1 path/file')(see man cvs for naming conventions)

At this point the link to jupiter hung and I lost lots more advice re: 
branch tags and the genmerge *.hlp file.  Please refer to it.  
I fixed a typo in the symlink to there.

I take this line:
...	 [genv10_branch] chgen =ver_10dflt=....
to mean that the tag genv10_branch or 10dflt is needed to check out this
version. (see man cvs or man 5 cvs)

There  were two gendflt projects - see $CASE/gen/ver_10_extensions.hlp:
genDfltRows added standardized dflt values into table TA
and (I believe) made pr_create initialize to these.
genDfltSuper made pr_create add a chain of superclass ancestors, if any, 
and properly install their pkeys into fkeys of subclass instances. 
Both of these avoid error-prone calls to pr_set for fields that 
do not require non-default values.  GenDfltSuper also saves ancestor creation 
and pkey to fkey copying via pr_set_key(pr_get_key()).

I added a number of *.dif files in gen/genmerge comparing various src paths:
They show that genv10dflt matches gen/ver_10 more than the others.
They do NOT show which dflt version resides here. I need to inspect src code
for this. I also checked the report genDFltVals.doc and converted it to 
Word7. It states that it is based on gen/ver_10 but not ver_10log.
I think this version should be merged into gen/ver_10 FIRST before
ver_10log or genDfltSuper.
---------------
jupiter.cs.uml.edu(321)> pwd
/usr/proj3/case/gen/genmerge/asabnis/ver_10dflt/src
jupiter.cs.uml.edu(322)> ls *.dif
cfiles.wc.dif          genv10src.dif          genv9.1src.dif
gen_pr_add.93to94.dif  genv10srcFull.dif      genv9src.dif
genv10logsrc.dif       genv8src.dif
jupiter.cs.uml.edu(323)>              
-------------------

> 

> Hi Professor Lechner:
> 
> I looked at /usr/proj3/case/gen directory and I found ver_10dfltRows and
> ver_10log directories.  But these two directories aren't under cvs version
> control, so why when I did a "cvs history -a" command, it gave me the
> following information:
> 
> % cvs history -a
> O 05/14 01:44 +0000 asabnis chgen     =ver_10=
> /usr/proj3/case/97s523/genme*
> O 05/19 02:48 +0000 asabnis [genv10_branch] chgen     =ver_10dflt=
> /usr/proj3/c*
> O 05/17 04:32 +0000 asabnis chgen     =ver_10log=
> /usr/proj3/case/97s523/genme*
> O 05/10 19:56 +0000 asabnis chgen     =ver_8=
> /usr/proj3/case/97s523/genme*
> O 05/12 04:39 +0000 asabnis chgen     =ver_9=
> /usr/proj3/case/97s523/genme*
> O 04/23 06:20 +0000 lechner CVSROOT   =CVSROOT=
> /nfs/jupiter/proj3/case/gen/*
> O 08/04 05:35 +0000 lechner chgen/doc =chgen/doc=
> /usr/proj3/case/gen/ver_10/cc
> O 09/12 01:57 +0000 lechner chgen/src =chgen/src=
> ~/genv10Updates/chgen/src
> 
> Also, when I attempted to do "cvs checkout ver_10dfltRows", I got this
> error message:
> cvs checkout: cannot find module `ver_10dfltRows' - ignored
> 
> Am I doing somthing wrong?  I thought that ver_10dfltRows and ver_10log
> are supposed to be under cvs revision control, and my job is to checkout
> both modules and then merge them, test them, and finally checkin the
> merged version back in cvs.  Am I right?
> 
> Regards,
> Patrick
>