April meeting minutes

7 views

Skip to first unread message

Abe Kornelis

unread,

Apr 9, 2026, 3:19:49 AMApr 9

to z390dev...@googlegroups.com, z390

Hello all,

Yesterday's April meeting for z390 developers
was attended by Don Higgins, John Ganci, Larry Simmons,
and Abe Kornelis.

We discussed zCobol. There are changes in the works to
improve support for Search and String/Unstring verbs.
The changes are not yet complete.

Abe reported completing the last updates to fix javadoc
comments. Two Pull Requests are pending for a review.

Increasing test coverage is currently stuck on a zCobol
issue related to multiplying packed decimal fields.

We also discussed the implementation of Floating Point
Registers. The data reside not in a single register array,
but in a series of arrays. The current data for each register
is in a different array, depending on the format of the
data created by the last operation that changed the value
of the register. There is a separate array that indicates
where the current value of each FPR can be found.
If we're going to implement the new vector facility,
we'll run into the issue that the FPR registers in z/Architecture
overlap with some vector registers. With the current
logic, that's going to be overly complex to implement.
We've discussed the options to revamp the current
implementation of the floating point registers to get
the data into a single array.

Abe has been working to revive the zCics regression tests.
He has converted the documentation to markdown
for inclusion in the repository. He has also been able to
run the tests, but some are broken. Abe indicates he has
no time to deep-dive into the issues to resolve them.
There is a Pull Request to cover the work done so far.
The PR is pending for a review.

Abe asked Don to see if there are any more documents
in his archive that might not be present in the current
repository. Abe offered to deduplicate and convert any
such documents.

Abe and Zane have been working to convert and prepare
zVSAM design documentation. Essentially there are two
versions that need to be reconciled and updated before
we can start work to prepare for actually implementing
any of the zVSAM logic. It's going to be a very large project,
and we're very keen on cutting it into really small chunks
to keep the work manageable.

There's a Pull Request pending for improving error
detection and reporting for some instructions that
have restrictions on register operands. The PR
needs to be reviewed. Abe to take action to find
a reviewer.

John brought up there are some issues with the
interactive test support. One can add multiple
breakpoints for execution addresses, but only a single
one for register data monitoring, and another single
one for storage alteration monitoring.
Also, we lack a command to list defined breakpoints.
On top of that, after disabling all breakpoints,
re-establishing one or more of the disabled breakpoints
results in illogical behaviour.
We discussed options to change the defined behaviour
and its implementation to achieve more consistency.

Larry mentioned he's still looking into the issue of
the TRT instruction not clearing the hi word when
executed in amode 64.

Next meeting is planned on May 13th, 21:00 UTC.
Depending on your time zone this may be a
Wednesday afternoon, a Wednesday evening,
or a Thursday morning.