Hello Kris, >The install processes use COPYFILE (OLDDATE, so you should not look for >dates equal to installation. >IMHO, COPYFILE should have OLDDATE as default, so its timestamp can be u sed >as a kind of version identification. > Actually I do not expect the files to have the same date. But as long as the newer files are different in either the date (whatever date) or in sizes the compare of listfile output should detect that.
I agree on the default. Too often OLDD is not used and even equal files h ave different timestamps. > >For an RSU you have no other method to find what has been changed than >comparing LISTFILE information (I've got COMPMDSK for that, part of my >COMPARE package). When applying indivudual PTFs, you can consult the VM SES >PARTCAT for changed files (I wrote SERVICED for that, part of SESTOOLS). This confirms my idea that I can compare the listfile output to find chan ged files. The other way would be to actually compare each file but the compa re on listfile is quicker. > >For saved segments: CMSINST must be rebuilt when one of its execs are >changed. The installation of an RSU wil probaby rebuild it d'office. P ipes >etc, idem ditto. >HELPSEG needs to be rebuilt each time someone accesses 19D in write mode , If segments are changed in the update processes (Indeed I do expect SERVICE/PUT2PROD to perform this) they will be backed up through DCSSBKUP and restored in the target using DCSSRSAV. So changes in CMSINST, PIPES e tc are copied. I already found that the DDR of the 19D is easier during install to prese rve the 1024 blocksize. When either the 19D or the segment is changed we use DDR to copy the disk. Thanks, Berry.
