Errors with more than 7 arguments given to REXX EXECs, bREXX does not even start

[SanLorenzo59]

I have run into a bug that a CMS EXEC in REXX cannot have many blank-delimited arguments,
bREXX does not even start. The string "CvG" is arbitrary.

You can use a dummy file (only "/*" inside) to get this bug.
See the examples below with 7, 8 and 9 arguments.

[marXtevens]

So far I am mucking about within BREXX and GCC code, in and around abort processing due to memory allocation. Nothing definite so far.

If someone else has knowledge to share ... I'm listening.

"Oh, what a tangled web we weave, when first we practice to code C" ( apologies to Sir Walter Scott, 1808)

[BobBolch]

Vm/370 space for an extended plist might be limited. After buildin the EXEC with the large number of arguments, write a second EXEC to call it, then call the second EXEC. Outer exec /* */ Call Inner with large number of arguments. /* Just as a test */ Bob Bolch Bob

…

On Sun, Sep 18, 2022, 7:06 PM marXtevens ***@***.***> wrote: So far I am mucking about within BREXX and GCC code, in and around abort processing due to memory allocation. Nothing definite so far. If someone else has knowledge to share ... I'm listening. "Oh, what a tangled web we weave, when first we practice to code C" ( apologies to Sir Walter Scott, 1808) — Reply to this email directly, view it on GitHub <#83 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARUFCBJEK2CXJB3AQ3GGZPLV66N7VANCNFSM6AAAAAAQPFAKRI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[SanLorenzo59]

Comparing VM/370 CE and LCM+L's VM/SP Release 5 :

an outer REXX exec calls an inner EXEC1 exec and an inner REXX exec - VM/370 CE fails, VM/SP runs.

[BobBolch]

Not the experiment I proposed. The inner exec I proposed testing is the rexx exec that accepts the long argument list. I want to see whether rexx called by rex can accept a long parameter list, as oppsed to rexx called by CMS. Bob Bolch

…

On Mon, Sep 19, 2022, 5:33 AM SanLorenzo59 ***@***.***> wrote: Comparing VM/370 CE and LCM+L's VM/SP Release 5 : an outer REXX exec calls an inner EXEC1 exec and an inner REXX exec - VM/370 CE fails, VM/SP runs. [image: Issue #83_0002d] <https://user-images.githubusercontent.com/91172642/190986622-98b2ca12-c62d-478c-bcba-9239e3ac7049.png> [image: Issue #83_0002a] <https://user-images.githubusercontent.com/91172642/190986676-8bf5d0cd-90dd-44e2-81f6-bd0afde58022.png> [image: Issue #83_0002c] <https://user-images.githubusercontent.com/91172642/190986706-684b0cbd-cc03-4d3d-92be-1ed2f4005eab.png> — Reply to this email directly, view it on GitHub <#83 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARUFCBN6ZBMNWT2CFZ273DLV7AXPJANCNFSM6AAAAAAQPFAKRI> . You are receiving this because you commented.Message ID: ***@***.***>

[SanLorenzo59]

The problem that needs to be fixed is with CMS calling the REXX interpreter:
EXEC processing initiated ? DONE
INNER EXEC A found ? YES
Beginning of REXX comment found in first line ? YES
REXX interpreter called and handed over fileid and arguments ? FAILED, [filename] field mangled

[SanLorenzo59]

******************** Problem Solved ***********************

VM/370 Community Edition Version 1 Release 1.2 has an outdated saved segment GCCLIB .
Running GCCSEG EXEC Y2 as MAINT solves this problem.

[BobBolch]

Excellent news! I have been in the hospital for a surgery, and I was just going to
start working on this today. You have saved me a lot of work.
Bob Bolch

[marXtevens]

I'm glad we have a solution. Curiosity is going to drive me to figure out what changed, or nuts, or both.

[SanLorenzo59]

I noticed "by accident" that this bug did not occur in a 16M virtual machine. I made this screenshot on a freshly installed "untouched" VM/370 Community Edition Version 1 Release 1.2 as MAINT.

[BobBolch]

When the VMSIZE is 16M, then the GCCLIB runtime code is loaded from the GCCLIB TEXT file (which has the correct code). The code in the distributed segment is back level, because I missed the step of resaving the segment. Resaving the segment resolves the issue, because the segment is used when theVMSIZE is small enough to load the segment. Bob Bolch

…

On Sun, Oct 16, 2022, 5:45 AM SanLorenzo59 ***@***.***> wrote: I noticed "by accident" that this bug did not occur in a 16M virtual machine. I made this screenshot on a freshly installed "untouched" VM/370 Community Edition Version 1 Release 1.2 as MAINT. [image: image] <https://user-images.githubusercontent.com/91172642/196028585-27f220f4-ba78-49ad-bf1c-49df7188bd40.png> — Reply to this email directly, view it on GitHub <#83 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARUFCBPMKPK3IRDKCDVOLNLWDPFETANCNFSM6AAAAAAQPFAKRI> . You are receiving this because you commented.Message ID: ***@***.***>

[SanLorenzo59]

When/how will this issue be closed ? The cause is known and instructions to fix an installed system are here. This can be fixed right after install. New users of http://vm370.org/VM/V1R1.2 however have to follow the link to https://github.com/s390guy/vm370/issues to learn of this issue and that a DIY fix is available.

Refer to

https://github.com/s390guy/vm370/issues

to see issues opened for the 1.1.1 release. Those marked "fixed in next release" are included in this 1.1.2 release.

This could be misunderstood ... have I to look elsewhere to see issues opened for the 1.1.2 release ?

[BobBolch]

When the next release (1.1.3) is published, the issues marked 'fixed in next release', on the issues page, will be included. Users will have to go to the issues page to see issues still open after the 1.1.2 release. To see issues fixed for the 1.1.1 release, and included in 1.1.2, go to https://github.com/s390guy/vm370/issues and click on the word 'closed' on the line that says 'xxx Open yyy Closed' Clicking on 'open' shows issues for the current (1.1.2) release. Bob Bolch PS These logistics questions should be posted on the https://groups.io/g/h390-vm forum, and not in an individual issue.

…

On Sun, Oct 16, 2022 at 8:07 AM SanLorenzo59 ***@***.***> wrote: When/how will this issue be closed ? The cause is known and instructions to fix an installed system are here. This can be fixed right after install. New users of http://vm370.org/VM/V1R1.2 however *have* to follow the link to https://github.com/s390guy/vm370/issues to learn of this issue and that a DIY fix is available. Refer to https://github.com/s390guy/vm370/issues to see issues opened for the 1.1.1 release. Those marked "fixed in next release" are included in this 1.1.2 release. This could be misunderstood ... have I to look elsewhere to see issues opened for the 1.1.2 release ? — Reply to this email directly, view it on GitHub <#83 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARUFCBI2QAETA6JEONEI2LDWDPVYFANCNFSM6AAAAAAQPFAKRI> . You are receiving this because you commented.Message ID: ***@***.***>