🚀 go-pugleaf

In article <1993Jan8.182534.14955@lds.loral.com>, eastwood@lds.loral.com (Jeff Eastwood 5304) writes:
|> 
|> I'm using DEC's PHIGS rel. 2.3C under VMS/POSIX on a VAXstation 4000. I can't
|> start a second instance of a trivial application (one that just opens up phigs,
|> writes a line of text, posts it, and closes.)  The first invocation works
|> fine. The second (started while the first is still running) fails in the
|> popenws() call. The workstation we are using is the PPWSTYPE_MOTIF_PEX_WIDGET
|> type.
|> 
Jeff: 

Sorry I can't solve your current problem.  But a question poped into my mind.
Please don't take offense:  

   Is this an example of why many developers do their best to buy software 
   from software developers rather than from hardware manufacturers?

Don't forget my bias:  TGS sells FIGARO+, the leading hardware independent
implementation of the ISO/ANSI PHIGS standard.  And FIGARO+ doesn't have
any problem with simultaneous applications....

Regards,

Joe

-- 
---------------------------------------------------------------------
 Joseph Stewart
 Senior Applications Engineer (FIGARO+, FIGt, FIGraph)
 Template Graphics Software (TGS)
 959 Concord Street 
 Framingham MA 01701-4613 (USA)
 508-872-8700 x340
 joes@tgs.com -- tgs!joes
---------------------------------------------------------------------

In article <1993Jan11.143505.5065@lpi.liant.com> joes@tgs.com writes:
>In article <1993Jan8.182534.14955@lds.loral.com>, eastwood@lds.loral.com (Jeff Eastwood 5304) writes:
>|> 
>|> I'm using DEC's PHIGS rel. 2.3C under VMS/POSIX on a VAXstation 4000. I can't
>|> start a second instance of a trivial application (one that just opens up phigs,
>|> writes a line of text, posts it, and closes.)  The first invocation works
>|> fine. The second (started while the first is still running) fails in the
>|> popenws() call. The workstation we are using is the PPWSTYPE_MOTIF_PEX_WIDGET
>|> type.
>|> 
>Jeff: 


	Any error messages (PHIGS, X, or otherwise)? Or does it just drop out?
I know I'm developing rather large molecular dynamics code with a Motif/PEX 
workstation on DECstation 5000's using their PHIGS (2.3a I think). The
resulting executable is LARGE to say the least due to the lack of shared
libraries I suppose. Stripped, the code is 6.5 meg. X may be free, but it
sure is expensive! Anyway, if I try to run two distinct copies of the
same program, the machine runs out of swap space/memory/whatever else.
However, I AM able to copy simulations from within the code itself,
and open several workstations without a problem. Might your problem be
memory related (ie, a lack thereof when you try to load more than one
executable into user space)??? 


				keith

In article <1993Jan8.182534.14955@lds.loral.com> eastwood@lds.loral.com (Jeff Eastwood 5304) writes:
>
>I'm using DEC's PHIGS rel. 2.3C under VMS/POSIX on a VAXstation 4000. I can't
>start a second instance of a trivial application (one that just opens up phigs,
>writes a line of text, posts it, and closes.)  The first invocation works
>fine. The second (started while the first is still running) fails in the
>popenws() call. The workstation we are using is the PPWSTYPE_MOTIF_PEX_WIDGET
>type.
>
>The guess is that too many colormaps are being allocated, so that the 2nd invocation
>can't alloc it's share. I've tried setting the various logicals, such as
>phigs$num_red_for_true, ...blue_for_true, ...green_for_true, and phigs$num_pseudo_colors,
>to small values (even 2). This does not help.
>
>Thanks for any ideas you might provide,
>
>Jeff Eastwood
>(813)-371-0811 x5304
>

And the answer is: don't try to run > 1 instance, when the executable is 
linked with /DEBUG. Debug executable size is ~10M; non-debug version
is ~1.5M.

Kudos to Keith Baccki for suggesting this solution.