Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
On Thu, Oct 27, 2022 at 04:17 PM, rvjansen@... wrote:
I would very, very much have it behave a bit more like modern XEDIT.
What makes XEDIT modern ? VM/SP 3, when REXX and EXTRACT were added ? When (if?) VM/SP 5 at LCM+L ( tty.livingcomputers.org:24 ) is up: which modern features are missing ?
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
Martin said,
No feedback ... are editors much less significant in the (emulated) mainframe community now since most people are doing it the Zowe way ?
I think we are overwhelmed with all the recent activity in this discussion group, and missed your note. I for one would rather see/use the code than look at a screen shot. The difference between Looking at the candy behind the glass, rather than actually having some. :-) ?... Mark S.
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
On Thu, Oct 27, 2022 at 04:45 PM, Bob Bolch wrote:
What I want to see in a full screen editor for VMCE, is the ability to use REXX as a MACRO language
for the editor so we can?implement?new editor commands in REXX and issue 'address XEDIT' type
calls from within these new commands.
Fully agreed.
I need to acquire more knowledge to enter this area. In the 1990s I was a very experienced professional assembler programmer on the Motorola M680x0 family. Used on bare metal including basic setup of the MMU, as a VERSAdos system service or as an interface library for Pascal programmers.
I have to learn all this for System/370. Please tell me where begin to read if there is a faster path than starting at GA22-7000-6? IBM System/370 Principles of Operation .
What do I need for "EE goes XEDIT" ?
- from C call an assembler routine to establish a CMS SUBCOM evironment "XEDIT"
- an assembler routine which receives subcommand calls and calls the C part of the editor
- an assembler routine using EXECCOMM to share REXX variables, EXTRACT needs this
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
±á¾±?¸é±ð²Ô¨¦,? What I want to see in a full screen editor for VMCE, is the ability to use REXX as a MACRO language for the editor so we can?implement?new editor commands in REXX and issue 'address XEDIT' type calls from within these new commands.? There are obstacle in implementing this, including getting EE out of the user area at X'20000' so that more than a small subset of CMS commands can be issued while EE is running. Bob Bolch
?
I think it goes without saying that we would very much like to see Rexx compatibility for EE and a more XEDIT-like working - although it is not at all bad at the moment, just a bit different. Because we are all very busy, it would be good if there would be an easy way to be involved in testing, e.g. by having a way to up/downgrade your EE installation, for example a github repo where we can easily download versions.
I would like to tell you what my personal priorities would be. As I am a late VM user (I had the course in 1987 on (probably) VM/SP R5 and had already my first opportunity to use XEDIT in 2019) (because in 1987 I only ever used VM to IPL MVS in it and then PR/SM came along) - all my Rexx editor macros start with ADDRESS ISREDIT.
Also, as opposed to most people I work with, I use the 3270 emulator keyboard as was intended, with enter mapped to linefeed and ctrl-right to send. So I am using XLIST and I wish it behaved a bit more like FILELIST (we could probably do that in Rexx now), I have X synonymed to EE and I find it almost workable. But I would very, very much have it behave a bit more like modern XEDIT.
best regards,
Ren¨¦ Jansen. On 27 Oct 2022, at 03:34, Martin Scheffler <San-Lorenzo@...> wrote:
No feedback ... are editors much less significant in the (emulated) mainframe community now since most people are doing it the Zowe way ?
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
I think it goes without saying that we would very much like to see Rexx compatibility for EE and a more XEDIT-like working - although it is not at all bad at the moment, just a bit different. Because we are all very busy, it would be good if there would be an easy way to be involved in testing, e.g. by having a way to up/downgrade your EE installation, for example a github repo where we can easily download versions.
I would like to tell you what my personal priorities would be. As I am a late VM user (I had the course in 1987 on (probably) VM/SP R5 and had already my first opportunity to use XEDIT in 2019) (because in 1987 I only ever used VM to IPL MVS in it and then PR/SM came along) - all my Rexx editor macros start with ADDRESS ISREDIT.
Also, as opposed to most people I work with, I use the 3270 emulator keyboard as was intended, with enter mapped to linefeed and ctrl-right to send. So I am using XLIST and I wish it behaved a bit more like FILELIST (we could probably do that in Rexx now), I have X synonymed to EE and I find it almost workable. But I would very, very much have it behave a bit more like modern XEDIT.
best regards,
toggle quoted message
Show quoted text
On 27 Oct 2022, at 03:34, Martin Scheffler <San-Lorenzo@...> wrote:
No feedback ... are editors much less significant in the (emulated) mainframe community now since most people are doing it the Zowe way ?
|
Re: Do I have a looping issue?
Real hardware has no mechanism to report to an operator any
interruption. REAL hardware only reports interruptions to the running
program, an OS in this case. So, no, real hardware will not report
anything.
However, the Hercules "hardware" DOES have the ability to report to an
operator some interruptions. Hence the whole discussion around the
OSTAILOR configuration option.
That is why you did not see it with your real hardware of the past.
Harold Grovesteen
toggle quoted message
Show quoted text
On Thu, 2022-10-27 at 08:38 +0100, Dave Wade wrote: -----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Fish Fish
Sent: 27 October 2022 02:07
To: [email protected]
Subject: Re: [h390-vm] Do I have a looping issue?
Joe Monk wrote:
[...]
But here's the real question ... why is hercules seeing/reporting
it? Huh?! Why WOULDN'T it?! Because the real hardware wouldn't. Pretty sure I never saw anything
like this on any 43xx or 9370 machine I used, and for a great
percentage of the time I was the only user and I was sat at the
system console...
--
"Fish" (David B. Trout)
Software Development Laboratories
mail: fish@... Dave
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
On Thu, Oct 27, 2022 at 10:51 AM, Bertram Moshier wrote:
have you thought about looking at the product KEDIT?
There is a German idiom "Eulen nach Athen tragen (, see also )": KEDIT *is* my principal editor since 1988. I have even written a UTF-8 import/export filter for it to deal with source code containing French and German literals and comments. I'd like to see KEDIT GPLed and opensourced on GitHub when (if?) Mansfield Software (Kevin Kearney) might close down business. Actually I am just implementing KEDIT's??? SET HIGHLIGHT? ? command. This can be done without breaking any XEDIT compatibility. You can expect that I will implement? MORE ,?? LESS? and? TAG as companions to? ALL . ? I am using "z/VM 7.3 XEDIT Command and Macro Reference" and these editors as reference :
- KEDIT for Windows 1.6.1
- XEDIT on z/VM 6.4
- XEDIT on VM/SP Release 5 at
- THE (The Hessling Editor), binaries and source
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
Of course its desirable to have as much of XEDIT as possible under VM/370 CE, except of course it limits the memory available. I tend to use EE to work on CE. ? It would also be nice to see betas etc. ? Dave G4UGM ? ? ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of Martin Scheffler
Sent: 15 October 2022 15:57
To: [email protected]
Subject: [h390-vm] #VMCE #rexx EE goes XEDIT - compiling a wish list? I am using XEDIT and its clones since VM/SP Release 1 up to today ... and counting. EE is great work by Dr. Hans-Walter Latz, but I miss a lot of features that XEDIT and KEDIT have. I have done massive extensions to EE to make it more XEDIT-compatible. A common infrastructure already exists for the SET / QUERY / EXTRACT / MODIFY / TRANSFER subcommands. Actually I am researching how EE could integrate with bREXX.
Having a macro capability powered by REXX is my first wish. The ingredients for the ALL macro will follow (SET SELECT, SET DISPLAY ...).
Do you think it is desirable and feasable to have a decent subset of XEDIT as part of VM/370 CE ?
|
Re: It has been awhile and I forget how to get past a power loss
You are correct in your order.? CKPT, FORCE, and only then COLD
On Wed, Oct 26, 2022, 22:04 Mark A. Stevens via <marXtevens= [email protected]> wrote: On Wed, Oct 26, 2022 at 01:41 PM, Dave Wade wrote:
You need to tell VM what to do as it can¡¯t find warm start data.
So
?
/cold
/force
/ckpt
?
Will do what you want.? I usually use ¡°/force¡± but I think ¡°/cold¡± is safer but it will lose any spool files.
?
Dave
I was taught to try, in the following order:
CKPT
FORCE
COLD
Selected from the GC20-1806-9_VM370_Operators_Guide_Rel_6_PLC_17_Apr81.pdf at
Cold Start
Cold start procedures are usually invoked only on operation of a new
version of the VM/370 system or files. as well as accounting records
are lost or because of a hardware error that prevented valid
continuation.
Failures in system component power, building power, hardware logic,
the CP program, or some (rare) kinds of user programs can be classified
as system operating procedure errors; all can require that you start the
system with a cold start. Some of these malfunctions may require you to
issue the SHUTDOWN command. If the SHUTDOWN command cannot be keyed in,
you can press the LOAD button on the System/370 console. Then after the
appearance of the shutdown message, you should press the LOAD button a
second time to reload the system with VM/370. VM/370 issues a message
to you to perform cold start procedures when it recognizes that
unrecoverable errors have occurred.
Warm Start
Warm start implies that the previous VM/370 session ended with an
orderly shutdown procedure. A warm start should be attempted after
VM/370 has abnormally terminated, and after a successful system dump
operation. Begin warm start procedures if yeu are not alerted by
console messages or visual indications that a checkpoint or cold start
is required. Warm start restores previous spool controls and files;
accounting records remain valid and are to te kept as a base for
continuing operations. In many cases, Vft/310 recovers froll system
failures and restarts itself.
Checkpoint Start
If the system could not perform a warm start because of I/O errors or
invalid data in the warm start area, you can request a checkpoint (CKPT)
start. This option attempts to initialize the system using the
information that has been dynaaically checkpointed during system
operation and stored in the checkpoint area. Spool file chains are
reconstructed; however, the original order of the files is lost.
Accounting and systea message data, recovered under a warm start, is
lost under a checkpoint start~ Ilso, because each spocl buffer aust be
read to reconstruct the record allocation blocks, a checkpoint start
takes longer than a warm start.
Force Start
If the system cannot perform a checkpoint start because of I/O errors or
invalid data in the checkpoint area, you can request a force (PORCE)
start. This option operates like the CKPT start except that any spool
file that contains unreadable or invalid data is dropped from the
system.
When the FORCE option is used, you should not allow any users to log
on to the system. The system should be devoted to reclaiming and
putting out as many spool files as possible. When this has been
accomplished, the system should be re-initialized using the cold (COLD)
start option.
I Hope This Helps.
?... Mark S.
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
Hello,
I would love to see XEDIT on VM/CE.? I know you have a lot invested in EE, but have you thought about looking at the product KEDIT?? While written for Windows 32 bit and most likely in C or C++, I'm wondering if it might help you with the logic?? Maybe not.
Yes, I would love to have XEdit functionality in VM/CE.
Bert / WB8ERT
On Sat, Oct 15, 2022 at 04:57 PM, Martin Scheffler wrote:
The ingredients for the ALL macro will follow (SET SELECT, SET DISPLAY ...).
The first prototype of the ALL command (builtin, not a REXX macro) is here. SET SELECT and SET DISPLAY are operational and almost complete. SET SCOPE is implemented but now there is *much* legwork to do to make all existing commands obey the new situation created by SET SCOPE DISPLAY.?
I will push my enhancements to GitHub after some more testing.
Should I consider publishing binaries as alpha preview releases ?
See screenshots:
look for the numbers in the prefix area
Highlighting of lines with selection level > 0
|
Re: Do I have a looping issue?
On Thu, Oct 27, 2022 at 05:49 AM, Tony Harminc wrote:
On Wed, 26 Oct 2022 at 08:06, Joe Monk <joemonk64@...> wrote:
Dave,
Compare_and_swap is a non-privileged instruction. But he's getting a
protection exception. Now how is that possible?
If we look at his PSW ... 078D2000000383EC
Remember, we are under VM so thats probably not what MVS thinks is the PSW
DAT is controlled by bit 5 in a EC mode PSW. He's got a 7 in the second
nibble, so the DAT bit is on. Thus, DAT is on. So, with DAT on, how is he
getting a protection exception?
As Fish said early in this thread, "the program is running in Problem State
with PSW Key 8 (the "8D" in your PSW), but the storage the 'CS' (Compare
and Swap) instruction is trying to access is Key 0 (the '0' in "K:06")".
That's how he's getting a protection exception.
Again that is what VM has loaded into the PSW. MVS may believe that DAT is off and he is in supervisor state.
Compare and swap is two registers and a memory location. Nothing requiring
a supervisor key...
What is a "supervisor key"? I know of no such thing in the S/370
architecture.
Sorry to ask a novice question, but does supervisor state by-pass storage protection?
Tony H.
Dave
|
Re: Do I have a looping issue?
toggle quoted message
Show quoted text
-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Fish Fish
Sent: 27 October 2022 07:44
To: [email protected]
Subject: Re: [h390-vm] Do I have a looping issue?
Joe Monk wrote:
Fish wrote:
Joe Monk wrote:
[...]
But here's the real question ... why is hercules seeing/reporting
it? Huh?! Why WOULDN'T it?! Because he's running MVS as a guest and the error is supposed to be
trapped at the MVS level. It *IS* being trapped at the MVS level! Isn't it? If it wasn't, I'm sure MVS would
probably crash! (i.e. go into a disabled wait) Only if running native.. We are under VM so its being trapped by VM, otherwise would crash...
I'm confused by your statements, Joe. Are you perhaps forgetting what Hercules
actually IS? Are you forgetting what Hercules DOES? That IT is "the hardware"
and is responsible for PRESENTING the program interrupt to MVS so that it can
be "trapped"? That IT (Hercules) is emulating the HARDWARE, and that the
hardware sees *everything*?
How do you think MVS sees the program interrupt? Magic? :)
Somewhat. It sees it because the smoke and mirrors in VM reflect it to MVS.. ... this isn't the case for all interrupts, some get handled in VM and MVS isn't aware they have occurred
--
"Fish" (David B. Trout)
Software Development Laboratories
mail: fish@...
Dave
|
Re: Do I have a looping issue?
Let me check I have this right¡ ? When running under VM we will always be in ECMODE, DAT ON, and in real problem state. The ¡°virtual PSW¡± which contains what MVS believes is happening is maintained in VMPSW in the VMBLOK When this interrupt occurs we load the appropriate FLIH from REAL page 0. Even when running V=R this will be VMs page zero, not MVSs page 0. VMs first level interrupt handler will then fake the OLD PSW that MVS expects to see which may very be different. It the re-dispatches MVS using the address from the NEW PSW location but once again DAT=ON, ECMODE=ON and SUPERVISOR STATE = OFF ? So its pointless looking at the PSW Hercules reports to see the Machine State that MVS believes it has created. What is reported is the fake VM has created¡ ? Dave ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of Tony Harminc
Sent: 27 October 2022 05:52
To: [email protected]
Subject: Re: [h390-vm] Do I have a looping issue?? points to note:
1. The routine is IEAVEVAL
2. It appears it may be called with DAT both ON and OFF
Sounds pretty quantum to me...
|
Re: Do I have a looping issue?
toggle quoted message
Show quoted text
-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Fish Fish
Sent: 27 October 2022 02:07
To: [email protected]
Subject: Re: [h390-vm] Do I have a looping issue?
Joe Monk wrote:
[...]
But here's the real question ... why is hercules seeing/reporting it? Huh?! Why WOULDN'T it?! Because the real hardware wouldn't. Pretty sure I never saw anything like this on any 43xx or 9370 machine I used, and for a great percentage of the time I was the only user and I was sat at the system console...
--
"Fish" (David B. Trout)
Software Development Laboratories
mail: fish@...
Dave
|
Re: #VMCE #rexx EE goes XEDIT - compiling a wish list
#VMCE
#rexx
No feedback ... are editors much less significant in the (emulated) mainframe community now since most people are doing it the Zowe way ?
|
Re: Do I have a looping issue?
Joe Monk wrote: [...] Thus, DAT is on. So, with DAT on, how is he
getting a protection exception? DAT has absolutely nothing to do with whether a Protection Exception occurs or not. Compare and swap is two registers and a memory location.
Nothing requiring a supervisor key... What are you talking about? Supervisor key? I've never heard of such a thing. Do you mean PSW key 0? -- "Fish" (David B. Trout) Software Development Laboratories mail: fish@...
|
Re: Do I have a looping issue?
Joe Monk wrote: Fish wrote:
Joe Monk wrote:
[...]
But here's the real question ... why is hercules
seeing/reporting it? Huh?! Why WOULDN'T it?! Because he's running MVS as a guest and the error is supposed
to be trapped at the MVS level. It *IS* being trapped at the MVS level! Isn't it? If it wasn't, I'm sure MVS would probably crash! (i.e. go into a disabled wait) I'm confused by your statements, Joe. Are you perhaps forgetting what Hercules actually IS? Are you forgetting what Hercules DOES? That IT is "the hardware" and is responsible for PRESENTING the program interrupt to MVS so that it can be "trapped"? That IT (Hercules) is emulating the HARDWARE, and that the hardware sees *everything*? How do you think MVS sees the program interrupt? Magic? :) -- "Fish" (David B. Trout) Software Development Laboratories mail: fish@...
|
Re: Do I have a looping issue?
Jim Snellen wrote: Fish,
I am NOT using OSTAILOR QUIET.
My configuration file has OSTAILOR VM.
At your suggestion, I've added OSTAILOR OS/390.
When I added the "PGMTRACE -04" statement immediately after
your OSTAILOR statements, I get an error message: <snip> Image not found. Please do not post screen grabs. Post your actual Hercules log file instead. ie. OSTAILOR VM
OSTAILOR +OS/390
PGMTRACE -04 The above worked just fine for me! I then tried:
OSTAILOR VM
OSTAILOR +OS/390 PGMTRACE -04
and still received the same error. The above is invalid. The PGMTRACE statement needs to be on a line by itself, just like you have it in your first example, which *should* have worked. Are you sure you are spelling things correctly? It would help to see your log file. (Not a screen shot! Not an image, but the actual text from your Hercules log file.) Where does the "PGMTRACE -04" go? Anywhere *after* your OSTAILOR statements, just like you have it in your first example above. I have no idea why you are claiming it's not working. It works fine for me! Are you maybe using HercGUI? If so, then yes, HercGUI will complain about the pgmtrace statement. It is not a statement that HercGUI normally expects to see in a configuration file. I will get that fixed in the next release. In the mean time simply click the "Ignore" button whenever that error pops up, or else check [X] the "Ignore any parse errors in this configuration file" option to make it go away permanently for that particular specific configuration file. -- "Fish" (David B. Trout) Software Development Laboratories mail: fish@...
|
Re: Do I have a looping issue?
[...]
points to note:
1. The routine is IEAVEVAL
2. It appears it may be called with DAT both ON and OFF
Sounds pretty quantum to me...
Tony H.
|
Re: Do I have a looping issue?
Dave,
Compare_and_swap is a non-privileged instruction.? But he's getting a protection exception. Now how is that possible?
If we look at his PSW ... 078D2000000383EC?
DAT is controlled by bit 5 in a EC mode PSW. He's got a 7 in the second nibble, so the DAT bit is on. Thus, DAT is on.? So, with DAT on, how is he getting a protection exception?
As Fish said early in this thread, "the program is running in Problem State with PSW Key 8 (the "8D" in your
PSW), but the storage the 'CS' (Compare and Swap) instruction is trying
to access is Key 0 (the '0' in "K:06")". That's how he's getting a protection exception.
Compare and swap is two registers and a memory location. Nothing requiring a supervisor?key...
What is a "supervisor key"? I know of no such thing in the S/370 architecture.
Tony H.?
|
Martin Scheffler
Mark A. Stevens
