September developer meeting notes

5 views

Skip to first unread message

Abe Kornelis

unread,

Sep 11, 2026, 6:08:04 AMSep 11

to z390dev...@googlegroups.com, z390

All,

we had a good and productive meeting.

Bob, Larry, and Ray joined,
John and Zane had given advance notice
of their absence.

After Ray introduced himself, we went over
the agenda.

1. Language support.
zCobol development has not seen much
progress lately. The status of Anthony's
MVS Cobol port is unknown.
Zane has proposed to create a zRexx
component. Plus, there has been a years-long
effort to create a C compiler for z390.

Since Anthony, Zane, and Tom (the c2z author)
were not in the meeting we could not
elaborate on any of these topics.

2. Test coverage and new instructions.
z390 has lots of regression test cases.
The issue is that many of these never made
it into our automated testing tool.
Abe is working hard to find these 'lost'
testcases, validate them, and integrate them
into the three testing frameworks that
z390 supports: bat scripts, bash scripts,
and groovy scripts.
Progress is steady, a lot has been done,
and a lot more still remains to be done.

For new instructions, z390 supports the
bulk of instructions up to and including z15;
after that we have been unable to keep up
with IBM.

3. 64-bit issues with modal instructions.
When z390 was originally build 64-bit mode
did not yet exist - modal instructions were
implemented as bimodal: they supported
24-bit and 31-bit addressing modes.
We upgraded a few instructions (TR, TRT, etc)
and in the process discovered a lot more
instructions suffer from this same lacking
support for 64-bit mode.
John has started making changes to fix
all of these issues. Current status unclear.
Expected to take a significant amount of time.

4. Java code reformatting.
We have configured a linter that will validate
that our java source code meets the minimum
requirements that we have imposed on
ourselves. The initial boatload of inconsistencies
has been resolved, one program at a time.
The final change is pending a review.
This large effort is now almost done.

5. zVSAM replacement plans
The current zVSAM implementation in z390
has some severe limitations, which led to the
creation of a design for implementing
zVSAM V2. Over time we created two versions
of the new design. The documents have been
converted to markdown, and merged.
Abe is in the process of making decisions
where the old designs conflicted or left
details undecided.
Catalog support in z390 currently does not
support all the attributes we need for
zVSAM V2. The first planned coding effort
will be to upgrade catalog support.

6. Work in progress c.a.
The z390 interactive window that can be
used as an alternative to a cmd window
or a terminal window, had some issues.
To resolve those John has replaced the
engine that z390 used by a better one.
This effort is now complete.

Bob had prepared changes to fix an issue
with error lines on the assembly listing,
that might show retained data from the
previous line on the report. We decided
to turn the rest of the meeting into a
shared working session. Bob shared his
screen and together we went through
the process of getting the changes into
the repository and onto github where
we created a Pull Request.
Abe promised to review the PR.

Next meetings are planned for:
- October 14th; 20:00 GMT
- November 11th; 21:00 GMT
- December 9th, 21:00 GMT
Please mind: these are GMT dates/times
For Australia/New Zealand the meetings
are in the early morning of the next date.