[Virtualacorn-list] Why doesn't VRPC register itself

[Jeremy Nicoll - ml virtualacorn]

Mik Towse <mik.towse at xemik.com> wrote:

> In article <51acff1bf2dave at triffid.co.uk> Dave Symes  wrote:
> > In article <mpro.lhconl007cwl301rw at wingsandbeaks.org.uk.invalid>,
> >    Jeremy Nicoll - ml virtualacorn <jn.ml.vac.83 at wingsandbeaks.org.uk>
> > wrote:

What do R-COmp say?

> > Dunno! Not asked, they'll probably say it's a VRPC thing, nothing to do
> > with UniServer.

> And they may be right. 

But then again they might not be.  UniServer has both VA and Windows
components and Dave is seeing a problem when it's used.  R-Comp would be my
first port-of-call in diagnosing this.

I'm not a Uniserver user, but run VA and Firefox a great deal, and have
never seen any interference between them.  That suggests to me that the
problem is somewhere in the way that one Windows app (VA) talks to another
(Uniserver).  It shouldn't be in the way that Uniserver talks to Firefox
because I would expect that to be using completely (Windows-)standard APIs.


Having said that, I do recall some previous discussion - was it on this
topic? - that hinted at problems when VRPC is running full-screen (something
I also never do).  Does the Uniserver/Firefox problem happen with VA in
full-screen?  And if so, does the problem stop if VA is running in a window?




> I have Win XP Firefox (WF) running alongside VRPC pretty much all the
> time.
> 
> I have a permanent 're-direct' for all VRPC email links, HTML files etc to
> be sent to WF[1]. However, the first time I click on a link (after VRPC is
> first run) will cause ROS to report back the standard message that it
> doesn't know of an app'n to load the file. Clicking again, will load the
> link into WF /and/ open C:\. WF usually also contains a blank tab for the
> failed load.
> 
> This isn't actually a WF error either. All PDF files are sent to Windows,
> too and if the first file to be launched to WXP is PDF, I will get the
> same error.

When you get the 'no known application' message, presumably the RO system
variable that describes what to do with a PDF file is not set?  When it then
works, has the var become set?  That'd be var

  Alias$@RunType_xxx      where xxx is the hexadecimal filetype of a PDF

(or possibly Alias$@LoadType_xxx)  I'm not sure about mailto etc, but they
may also be controlled inthe first place by the content of a system
variable.



 
> But it gets stranger... My VRPC is stored on a removable USB drive. So
> when I am 'on the road', I plug it into my wife's netbook[2]. In this
> case, the first click always launches correctly.
> 
> I've discussed the issue with Andrew, but since the VRPC installation has
> merely been moved from one system to the other, where do you start?

First you remember that even if RO boots the same every time (it didn't use
to on my real RPC, because minor timing differences in eg hardware
initialising or whether or not there was a printer attached to the parallel
port) affected it) the two underlying Windows systems are not identical.
That may affect how VA boots too, apart from how the Windows end of
Uniserver works.
 

You start with a diagnostic trace from the RO end of Uniserver reporting
what it requested and what worked or didn't in each case, probably in
conjunction with something similar from the Windows end.  If the problem is
as repeatable as you say, tracking down why it happens should be easy.  (Not
for you, but for R-Comp who should provide you with some way to collect the
info they need to diagnose it.  I'd be amazed if both ends of the process
can't create detailed logs because I'd expect that to have been needed by
the programmer as features were developed.  If Andrew wasn't the programmer,
I'd want to be /sure/ that he'd discussed the issue with the programmer.)  

-- 
Jeremy C B Nicoll - my opinions are my own.