May meeting minutes

[Abe Kornelis]

All,

During our May developers meeting we discussed
the z390 status quo with Don Higgins, John Ganci,
and me (Abe Kornelis).

For zCobol, Abe mentioned that all testcase programs
have now been cleared of issues and add to the 
z390 test suite. For the NIST test suite we still have 
tons of issues, but that is of later concern.
The MVS Cobol port that Anthony has been working on
has an unknown status.

The javadoc update effort has finally been completed.
All javadoc now generates without warnings.
And, what's more: Anthony has made an update to the
github.io main page for z390 to have a direct link
to the javadoc pages!
Main z390 page: https://z390development.github.io/z390/
On the contributing sub-page, under the menu icon,
there is an entry point to the javadoc at
https://z390development.github.io/z390/javadoc/package-summary.html

Next we discussed the options for applying a single style
for the java code. As it stands various authors have applied
there own preferences. We decided that John will work on
the lz390 program to come up with a concrete proposal.

Another ongoing project is improving test coverage.
Various pre-existing test programs never made it into
the groovy scripts that we use to ensure we do not 
cause regression during maintenance.
The work has steadily progressed, and now that the
zCobol test cases have been done, I'm planning to
continue with other z390 components.

For zCics, Don has found a collection of PDF files that
Melvyn Maltz created. Abe is in the process of converting
these documents to markdown for inclusion in the z390
repository, where they can be merged with existing zCics
documentation. John said he'd have a look at the change
that covers the first of these converted documents.
With zCics regression tests we ran into an issue: they contain
at least one test that is currently broken. I've let that issue rest.
John mentioned he'd have a look at the issue, time permitting.

The zVSAM V2 effort is on hold for a bit. I am working on 
too many things at the same time. I'm planning to resume
merging the two versions of the design documents
after completing the zCics documents have all been converted.

Abe mentioned that the current implementation of Floating
Point is not compatible with the vector register setup in
current zArchitecture. IBM has made the FP registers coincide
with selected sections of the Vector Registers.
This overlap is hard to mimic unless we re-implement
the FP registers and the Vector Registers as a single array,
with the same overlap that the zArchitecture describes.
For now, we decided that Abe will write an issue to cover
this. (see issue #804 on github)

Next we discussed the that we're still using Java V8.
Which is still supported, although quite old.
There has been an attempt - many years ago -
to upgrade to a newer version of Java, but we could
not bring back memories of what caused that effort
to fail. For now we decided that Abe will make an
inventory of all issues that need to be looked into.

Then we discussed creating a new distribution.
We decided to create v1.8.4.3 first to release all the
zCobol improvements already dine. We'd like to complete
the pending enhancements for SEARCH/STRING/UNSTRING
verbs before creating v1.8.5

Next meeting: June, tenth, 20:00 UTC.
For New Zealand / Australia: that's June 11th, early morning.

Kind regards & happy programming!

Abe
===