topic Re: _XDispatchInput return 0 ??????? in Operating System - OpenVMS

_XDispatchInput return 0 ???????

Jansen_8 — Mon, 30 Jan 2006 05:57:28 GMT

I am trying to port a big opensource package( gtk2 ). When runnig a simple test program I get different results depending on de X-display I try to use:

Linux-system with Xrender-extension : correct

OpenVMS system (8-bits or 24 bits /Pseudo-color or True color (without Xrender) :
first error is :_XDispatchInput return 0
after that the program either crashes or gives me :
[_XDispatchInput - entered in reply state]

I cannot find any info on this message. What does it tell me?

Re: _XDispatchInput return 0 ???????

Wim Van den Wyngaert — Mon, 30 Jan 2006 06:09:10 GMT

Seems rather old to me but any way.

http://h30266.www3.hp.com/MasterIndex/cover_letter/cover_letter_0023ff3c.txt

In DECwindows Motif V1.2-3, remote logins often fail when the server and client are running different versions of OpenVMS,
and if the DECW$LOGIN.LOG file is enabled the error message "[_XDispatchEvent - entered in reply state]" is written to the
log file. For example, the remote login will fail if the client system is an Alpha system running DECwindows Motif
V1.2-3 and OpenVMS Alpha V6.1 and the X11 display server is a VAXstation running OpenVMS VAX V6.1, but it would succeed if
the client system were running OpenVMS Alpha V1.5. This problem is corrected in the remedial kit.

Wim

Re: _XDispatchInput return 0 ???????

Jansen_8 — Mon, 30 Jan 2006 07:09:39 GMT

Forgot to tell you. The OpenVMS system I useis a OpenVMS8.2 alpha system with DecWindows 1.5. The application giving the error is just an X-application competely compiled from source and runned by an unpriviledged account.
The source code does not contain a call to _XDispatchInput. I think it is called seomewhere in the X11-libraries.

Re: _XDispatchInput return 0 ???????

Wim Van den Wyngaert — Mon, 30 Jan 2006 07:42:19 GMT

Your problem doesn't seem to be in the release notes but did you have eco 1 installed ?

http://www4.itrc.hp.com/service/patch/patchDetail.do?BC=patch.breadcrumb.main|patch.breadcrumb.search|&patchid=DEC-AXPVMS-DWMOTIF_ECO01-V0105--4&context=openvms:alpha:8.2

Wim

Re: _XDispatchInput return 0 ???????

Jansen_8 — Mon, 30 Jan 2006 08:16:59 GMT

All patches/eco's are applyed to the systems.
I used the latest releases of the compilers to create the application.

Re: _XDispatchInput return 0 ???????

Jan van den Ende — Mon, 30 Jan 2006 09:09:12 GMT

Hallo,

I know little about X, but

"is just an X-application competely compiled from source"
in combination with
"first error is :_XDispatchInput return 0"
always triggers this Pavlow-reactie in me:

"Is this been ported from one-or-the-other Unixian, and did they forget to translate the error passing?"

in U*X, 0 usually means "all fine", whereas in VMS the ODD numeric values are success statusses, and any even status is some error, with 0 being an unspecified Warning.

Could that be the case here also?

... just my EUR 0,02.

Proost.

Have one on me.

jpe

Re: _XDispatchInput return 0 ???????

Jansen_8 — Mon, 30 Jan 2006 10:04:39 GMT

Yes, it is a port from Unix, but... It the code itself never calls XDispatchInput. That is done somewhere in the DECW-Libraries. Also the error message is generated somewhere in side the DECW-libraries

Re: _XDispatchInput return 0 ???????

Edward Miller_1 — Fri, 15 Sep 2006 18:29:34 GMT

We recently converted from Alpha VMS 7.3-2 (DECW v1.3) to VMS 8.2 (DECW V1.5)
and encountered the "_XdispatchInput return 0" error message. In our case it appeared multiple
times during the initialization of our X application (running on Alpha VMS), but
not after initialization. Apart from the messages, there was no indication of any
problem with the application.

I inquired about this with HP support and received the following response:

"_XDispatchInput return 0" is a debug message that accidentally got left out
in DECwindows V1.5 ECO kit, DWMOTIF_ECO01 V1.5. This message is now removed
and the fix will be available in next release of DECwindows. This debug message
does not indicate any error so customer does not need to handle it.

I have no further information as to which release does/will have this problem
corrected.