How to set up X-Window in OpenVMS IA64

[Manesh Sasankan]

i want to setup the X-window in OpenVMS IA64 machine. This is what i did to do that.

1) My display was not set correctly. When i tried the SHO DIS command, it came up with this.

%DECW-W-OPENIN, error opening DECW$DISPLAY as input
-SYSTEM-W-NOSUCHDEV, no such device available

i set it manually with the SET DISPLAY command

$ SET DIS/CREATE/TRANS=TCPIP/NODE=10.xx.xxx.xxx

then i started a passive connection from my workstation using the XMANAGER software. After doing this, i tried to open up the clock utility using MC DECW$CLOCK and it worked.

But the problem is when i start my server application in X-Window mode, a new window pops up with an error. actually my application trigger this pop up window. but it was working fine a few months back.

So do i have to enable anything other than this to make my application work in X-Window. I just want to know whether this is completely related to me application code, since the clock utility is working fine or do i have to do something else to make this work .

Appreciate if anyone could comment on this.

[EdgarZamora_1]

More information would be helpful, such as the exact error message that "pops up" when you run your application.

[Manesh Sasankan]

this is the exact error thats coming in the pop up window:

"Fatal: Error 406225112 in getting text of message for HLI error 403014050"

please find the attachment for more details.

Thanks,
/Manesh

[Hoff]

What was attached to the message is bitmap image of a Microsoft Windows error dialog box, apparently generated by the X client running on Windows. The text in the dialog is:

Fatal: Error 406225122 in getting text of message for HLI error 403014050

Consider trying your application with a different X Windows server, as a start. One not using eXcursion or whatever X Windows server generated this dialog box. (X11 is included with various operating system distros, including Linux and Mac OS X. I posted up the general X Windows sequence for using X11 on Mac OS X with OpenVMS over at the new HoffmanLabs site about two weeks back.)

Check for ECOs and version upgrades for eXcursion or whatever X Windows server this on this Windows box, do consider reinstalling it, check installation requirements, and I'd then call the HP support center or whomever supports this X Windows server, as digging around inside an X Windows server not going to be easy.

Stephen Hoffman
HoffmanLabs

[Hoff]

Alternative thought: Is that HLI error arising from the application you are seeking to use? If so, check with whomever is supporting the application. SEARCH the application images and files for a fixed substring from that message, as a start, and see if you get any matches. (This would also likely show up if you test with another X Windows server, an outcome which would imply that the problem is on the OpenVMS end, either with OpenVMS or with the application.)

[Manesh Sasankan]

is there any other X-server softwares available
specifically for OpenVMS. Now i'm using X-Manager software. i tried using Hummingbird exceed. but i couldn't start a passive connection from Exceed to openVMS. so this comes up with error to set the display variable properly.

any suggestions for the X-server softwares??

[Anton van Ruitenbeek]

I'm using DECWindows on the OpenVMS site and ReflectionX on the PC site. Everything is working fine.
Possibly you are trying to connect to OpenVMS and start downthere a session wich will create a window on your PC. I think the procedure on the OpenVMS site isn't correct. You need to create an output device every time you login.
I had this problem in the beginning too. It's a long time ago, but try to get it for you.

Simetimes the procedure (PC) runs using SSH or other possibility of remote procedure startups. Make sure these loginprocedures may run. Possible it is not enabled within TCP.

Secondly I have a procedure to run and hereby the procedure
$ define decw$login_norestart true
$ temp = f$element(1, " ", f$getdvi("TT", "TT_ACCPORNAM")
$ set display/create/node='temp'/transport=tcpip/server=0
$ create/terminal/detach !< or whatever you want to start
$ exit

AvR

[Manesh Sasankan]

AVR,

Do you want me to execute all these commands?
======================
Secondly I have a procedure to run and hereby the procedure
$ define decw$login_norestart true
$ temp = f$element(1, " ", f$getdvi("TT", "TT_ACCPORNAM")
$ set display/create/node='temp'/transport=tcpip/server=0
$ create/terminal/detach !< or whatever you want to start
$ exit
======================

could you please explain what the above commands does?

Thanks,
Manesh

[Martin Vorlaender · Accepted solution]

Manesh,

>>>
i tried using Hummingbird exceed. but i couldn't start a passive connection from Exceed to openVMS. so this comes up with error to set the display variable properly.
<<<

eXceed should be fine. Just use the following for the command in the XStart application:

PIPE SET DISPLAY/CREATE/TRANSPORT=TCPIP/NODE=@a && @@SYS$MANAGER:DECW$STARTSM.COM

Or: configure the TCP/IP XDM on the VMS side (see http://h71000.www7.hp.com/doc/83final/6526/6526pro_contents_005.html#toc_chapter_21 ), and use eXceed's XDMCP Broadcast.

HTH,
Martin

[Manesh Sasankan]

Thanks Martin Again.. I was able to start a passive connection from eXceed to VMS. Now the problem is with the source code.

So when i run that image, it's coming with the error.
>>>
"Fatal: Error 406225112 in getting text of message for HLI error 406224986"
<<<
I tried searching both the numbers. its not there in the source code. Even i found the .c file which prompts this error.

But how can i trace the problem.Can i use the CC to debug the problem. Now I'm not to find out the root cause of the problem, the place where it fails.

any help??

[Joseph Huber_1]

VMS a comfortable source code debugger, it is not CC specific.
Compile the source with CC/DEBUG/NOOPTIMIZE,
and link with LINK/DEBUG.
Then when You RUN the image, the debugger window comes up. Use the built in HELP to find out how to trace and set break points.

[Hoff]

Manesh, a little terminology can help avoid some confusion: the X Windows Server is the component that drives the X Windows display, and this server is usually operating on the same system as the display. (While it is possible to run the X Window Server remotely, that's not common.) With your particular configuration, the Hummingbird eXceed package is the X Windows Server. The packages making requests of the server are known as X Windows clients, and these can be local to the server, or remote, or both. In this case, the applications running on OpenVMS are the X Windows clients. Client applications use the display setting (usually the DECW$DISPLAY logical name and its underpinnings, on OpenVMS) to locate and to open up connections to the X Windows Server; to connect to and to display output over on eXceed in this case.

Once you have the source code built and linked with /DEBUG (as described earlier), I'd probably also define the logical name:

$ DEFINE DBG$DECW$DISPLAY " "

Before starting the application. This to bring up the character cell debugger on the OpenVMS box, rather than bringing up the X Windows debugger on eXceed. Having two packages -- yours and the debugger -- both doing remote X Windows can potentially get confusing. Using the character cell debugger means you can navigate your code in the same (and probably DECterm) window where you had invoked the application.

And if you are using X calls in your application, you'll most definitely need the X manuals. These are entirely separate from OpenVMS and its documentation, and also largely separate from the DECwindows documentation. These X manuals can be acquired from OSF or other sources, or can be acquired in the form of books from various sources.

Stephen Hoffman
HoffmanLabs

[Manesh Sasankan]

Thanks all ur replies.

i was able to link the image with /DEBUG parameter and started the image in Debug mode. After performing several steps, this is the error which the debugger show:
>>>>>>>>>>>>

DBG> step
%DEBUG-I-DSFMISMATCH, using I64V82$DKA0:[VMS$COMMON.SYSLIB]LIBRTL.DSF; but it does NOT appear to match with DISK$I64SYS:[VMS$COMMON.SYSLIB]LIBRTL.EXE;1
%DEBUG-I-DYNIMGSET, setting image LIBRTL
%DEBUG-I-DYNMODSET, setting module LIB$FIND_IMAGE
%DEBUG-I-DYNLNGSET, setting language BLISS
%DEBUG-W-UNAOPNSRC, unable to open source file $1$DGA2027:[LIBRTL_5.SRC]LIBFNDIMG.BLI;1
-RMS-F-DEV, error in device name or inappropriate device type for operation
%DEBUG-I-DYNIMGSET, setting image RS1WS
%DEBUG-I-DYNLNGSET, setting language C
stepped to STN\main\%LINE 42447 in THREAD 1
DBG>
<<<<<<<<<<

After this debug error, the pop up window appears with this message

Fatal: Error 406225122 in getting text of message for HLI error 406224986

& the debugger stops responding after clicking the button in the pop up window & displays the following:

%DEBUG-I-EXITSTATUS, is '%SYSTEM-S-NORMAL, normal successful completion'
%DEBUG-I-NOSOURCE, No source for address: 0000000000000000

Appreciate if anyone could comment on this!!!

Thanks
/Manesh

[Christopher Blackburn]

Manesh,

I am also getting the message in the DEBUG package that indicates %DEBUG-I-DSFMISMATCH, using disk[vms$common.syslib]librtl.dsf but it does NOT appear to match librtl.exe.

I get this when I try to set break points within the code. I then find the image being debugged crashes with an access viloation (I have started a new thread for this problem).

Did you ever work out why the DEBUG message comes out.

[The Brit]

I would be less worried about the "DFSMISMATCH" (only an informational message), and be more worried about the

%DEBUG-W-UNAOPNSRC, unable to open source file $1$DGA2027:[LIBRTL_5.SRC]LIBFNDIMG.BLI;1
-RMS-F-DEV, error in device name or inappropriate device type for operation

This message is a DEBUG WARNING, and a VMS(RMS) FAILURE. Is $1$DGA2027 one of your disks, and does [LIBRTL_5.SRC] actually exist??


Dave

[Hoff]

[[[%DEBUG-W-UNAOPNSRC, unable to open source file $1$DGA2027:[LIBRTL_5.SRC]LIBFNDIMG.BLI;1
-RMS-F-DEV, error in device name or inappropriate device type for operation

This message is a DEBUG WARNING, and a VMS(RMS) FAILURE. Is $1$DGA2027 one of your disks, and does [LIBRTL_5.SRC] actually exist??]]]

I'd expect that this is an OpenVMS Engineering disk and an OpenVMS engineering path to the source file for the LIBRTL build involved here.

As for how to resolve this, I'd first look to see if the DSF matches the LIBRTL present; I could envision a case where an ECO delivered a new LIBRTL image and might have omitted a DSF for the file.

In the cases of DSF skewage seen, it's triggered by the omission of the DSF from the ECO kit that updated LIBRTL image. Easiest fix here is to rename the stale DSF out of visibility: RENAME SYS$SHARE:LIBRTL.DSF to LIBRTL.DSF_V82 or some such to get the stale file out of visibility here. (This RENAME is harmless, and serves to suppress the informational. Well, you won't get the symbols from within LIBRTL, but were you really looking to debug that?)

If you want the official RENAME fix or if you want the DSF for the version of LIBRTL that you have here, check with the nice folks at HP. But do check to see if the latest ECO kit includes the DSF, too. This omission was supposed to be fixed.

I'd also suggest and would tend to start up a new thread specific to the topic, rather than what looks to be similar error messages in what are almost certainly entirely dissimilar contexts. You have no "control" over this (old) thread, have no points or other management controls, and the title here tends to get buried. And yes, you could start a thread and link back here as a previous instantiation... (I, for one, was not expecting to find a DSF question underneath a DECwindows configuration title...)

Stephen Hoffman
HoffmanLabs LLC