Developers meeting report

[Abe Kornelis]

All,

Today we've had another developers' meeting for z390.
It was great seeing each other again and discussing
progress of z390 development.

First of all Don gave us an update in the new distribution
he created - V1.7.07 is out. Only the update of the repository
on github still needs to be taken care of.
Also a few of John Ganci's fixes to bat files have been missed out.
John assured us that the issue is only with a borderline case.
For normal operation there should be no issues.
Don will be working with Anthony to get all the V1.7.07
changes promoted into the main repository.

Then we discussed some catalog issues.
Currently z390 supports a catalog that holds its data
in a loadmodule. To support zVSAM V2 we have needed
to define an extended version of this load module.
So we will have V1 catalogs for use with zVSAM V1
and V2 catalogs for use with zVSAM V2.
We're not planning to create support for mixed
environments that need to support both V1 and V2
simultaneously.

With github still being at the level of distribution V1.7.06
the next step will be to commit all V1.7.07 changes into
the repository. Once that is done, all development branches
can be upgraded to the same level.
We browsed around github to have a look at the outstanding
pull requests.

The issue list on github is now fully operational.
We're using it to track all open issues.
The RPI list has been superseded and will no longer be used.

John gave us an update of the status of his ongoing work
with the bat and bash scripts. His changes are due to be
promoted into the repository's main line.
We discussed what needs to be done, and concluded that
first of all V1.7.07 needs to be in the repository, then
John can upgrade his branch to the same level.
Once that is done his changes can be promoted into
the repository's main branch.

Then I gave a short update on zVSAM status.
Some of the programs we need to build our test datasets
will not compile on the back-level assembler that is
present on the current development branch.
The resolution is to upgrade the branch to the V1.7.07
level that Don will have to put into the repo.
Once that has been done I'll resume work on zVSAM V2.

After that we had a discussion on changes needed for open/close.
The current implementation generates an SVC call for each
DCB or ACB to be opened. We'd rather have a parameter list
so that multiple datasets can be opened with a single SVC.
Also, the current implementation will assemble either separate
SVC instructions or a parmlist, depending on the zVSAM option
in effect. We decided that open/close behaviour should not be
influenced by the zVSAM version in effect.

Then we discussed zVSAM V1 and V2 co-existence issues.
As it currently stands the zVSAM version in effect is decided
when the emulator initializes. After initialization only the selected
version will be available. This means that co-existence will be
supported only in the sense that you can run V1 code and/or
V2 code, as long as they 'live'  in separate invocations of the
z390 emulator environment.
This co-existence will be around for quite a while.
At least while zVSAM is being developed, both versions will need
to be present - given the restriction that initialization selects
either the one or the other version. The zVSAM V2 code will need
to be sufficiently tested and stabilized before all V1 can be removed.

Finally, Anthony told us about his efforts to convert extant documentation
to a textual format that is more suitable for git. He added some logic
to make the doco searchable. He showed the web interface into this doco.
And he asked for feedback on the documentation.
@Anthony: Would you please share the web URL ?

Next meeting is planned in three weeks.
I'll keep you posted.

Kind regards & Keep well,
Abe
===