z390 August developer meeting
Abe Kornelis
All,
Our next developer
meeting (the August edition) will be on August 6th/7th
(depending on time zone)
Here's the meeting link: https://teams.live.com/meet/936640015321?p=jqWPw73ipLka3NPAKo
| Location | Local Time | Time Zone | UTC Offset |
|---|---|---|---|
| Brisbane (Australia – Queensland) | Thursday, 7 August 2025 at 6:00:00 am | AEST | UTC+10 hours |
| Dallas (USA – Texas) | Wednesday, 6 August 2025 at 3:00:00 pm | CDT | UTC-5 hours |
| Amsterdam (Netherlands) | Wednesday, 6 August 2025 at 10:00:00 pm | CEST | UTC+2 hours |
| Corresponding UTC (GMT) | Wednesday, 6 August 2025 at 20:00:00 |
Some topics we might
want to discuss:
- zCobol support, test
suite, MVS Cobol port
- Optable support, opcode tables, and the new z17
- Scripts in bat / bash / groovy files
- Unable to compile z390 due to java 8 not accepting --release parameter
- Missing Javadoc
comments
- Anything else
Kind regards to you all
& hope to see you in two weeks!
Abe
===
Abe Kornelis
All,
we had another productive z390 developer meeting.
We discussed whether or not to synchronise the
functionality of the groovy, bash, and bat scripts
and decided that it's not worth the effort,
although we'll do try to keep the basic
user-facing scripts in sync. That is: the scripts that
most of us use to run assemblies/compiles,
link-edits and test runs of our programs.
We also discussed support for non-default
extensions. Support for non-default extensions
has been with z390 for a very long time, but it
was never documented. We agreed to improve
documentation and to fix the code where the
default is enforced, rather than being applied
only when not specified explicitly.
This will only hold for input files, not for
files created by z390 during assembly or
link-edit.
We noticed some issues with the z390 build
process: the java compiler options used in the
bat and bash scripts are not identical.
The use of -xLint option suppresses some warnings
that we'd rather seek to resolve.
We'll try to address these issues and also look
at the compatibility level of the compiler,
which is still at Java 8.
Finally, we discussed the availability of z390
documentation when working off-line.
Anthony indicated he'd make an overview
of alternatives of how this can be achieved.
Kind regards & hope to see you again next month.
Abe
===
Abe Kornelis
All,
John just pointed out to me that I'd forgotten
to mention the discussion we had on floating point
arithmetic. Bob and John have both spent
considerable time and effort in trying to understand
the internal logic that z390 uses to implement the
floating point instructions.
Bob discovered (some time ago) a bug that could
be fixed relatively easily, but the underlying problem
would not be resolved. We feel it's a bad plan to
swipe the issue under the rug.
The complexity seems lo lie largely with the fact
that the floating point registers can hold data in
different formats (base-2, base-16, or base-10).
Only the base-2 version is supported by standard
Java math libraries, other formats require
conversions back and forth. We have a suspicion
that the issue that Bob detected is ultimately
caused by these conversions. This means that any
other floating point instructions might be
affected as well.
Kind regards & happy programming,
Abe
===
René Jansen
--
You received this message because you are subscribed to the Google Groups "z390" group.
To unsubscribe from this group and stop receiving emails from it, send an email to z390+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/z390/b0a06983-041e-4868-a070-84fdff711daa%40bixoft.nl.