AnsweredAssumed Answered

vrf multiple copies running

Question asked by VRFuser on Dec 16, 2001

Like • Show 0 Likes0
Comment • 2

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"><HTML><HEAD><META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1"><META NAME="Generator" CONTENT="MS Exchange Server version 5.5.2654.45"><TITLE>RE: vrf Multi-VEE-ing capability</TITLE></HEAD><BODY>Rob,Actually, we found it does release it, but cannot reopen it. Regardless, we can't use it with another app that tries to open & close it. For that reason, we have started an evaluation of an alternate solution (i.e. not use VEE).Paul H. Rubin Staff Engineer Test Engineering Ericsson Amplifier Technologies, Inc. 49 Wireless Blvd. Hauppaugh, New York 11788 Phone: (631) 357-8513 E-Mail: paul.rubin@ericsson.com -----Original Message----- From: ramarquardt@ra.rockwell.com [<A HREF="mailto:ramarquardt@ra.rockwell.com">mailto:ramarquardt@ra.rockwell.com</A>] Sent: Monday, December 17, 2001 9:54 AM To: Paul Rubin (AMT) Subject: RE: vrf Multi-VEE-ing capability I think it's a known VEE bug. VEE doesn't release the resource even if you execute a CLOSE. We don't even try to do this anymore. If pressed, however, we could funnel it through VEE via a TCP/IP socket or use two ports and a remote control switch |at least until this gets fixed>.RobRob Marquardt System Test Engineering Rockwell Automation / Allen-Bradley 1201 So. 2nd Street Milwaukee, Wisconsin, 53204-2496, USA 414-382-3098 Phone 414-382-4820 FAX RAMarquardt@ra.Rockwell.com "Paul Rubin (AMT)" |Paul.Rubin@am1.ericsson.se>@lvld.agilent.com on 12/14/2001 03:53:10 PMSent by: owner-vrf@lvld.agilent.com To: "'Thomas@Vauderwange.de'" |Thomas@Vauderwange.de>, vrf@lvld.agilent.com cc:Subject: RE: vrf Multi-VEE-ing capability Thomas,Apparently you are only claiming that 2 _VEE_ programs can "share" a Com port. Correct? Have you ever succeeded in having VEE share a Com port with any other type application?Paul H. Rubin Staff Engineer Test Engineering Ericsson Amplifier Technologies, Inc. 49 Wireless Blvd. Hauppauge, New York 11788 Phone: (631) 357-8513 E-Mail: paul.rubin@ericsson.com-----Original Message----- From: Thomas Vauderwange [<A HREF="mailto:Thomas@Vauderwange.de">mailto:Thomas@Vauderwange.de</A>] Sent: Friday, December 14, 2001 4:35 PM To: vrf@lvld.agilent.com Subject: Re: vrf Multi-VEE-ing capability Hi Paul,Well, found it hard to believe when I heard it. I was aware that VEE won't let the COM-port loose, once it got hold of it, so the idea to try COM-port-sharing with non-VEE-apps was abandoned.Then, that customer from over there in luvly North Carolina called one day, reporting a 'severe firmware bug' in our machine's microcontroller, as there were the two LEDs that indicate the two possible control modes of that machine blinking! In a very regular manner, the right one - the left one. As something like this was completely unheard of, I started the investigation and found out that my software, which he uses to control/monitor the machine from a PC had been started two times simultaneously on that PC. On the copy in the foreground, that customer had pushed the button that did send (via COM1) the signal to change control mode. The other instance did notice the deviation with its next data exchange and thus ordered the machine back in the other mode. Which led the copy in front of the screen to send another command...Anyway, it worked fine. Unfortunately, you need a real one of these machines to try my software out, which would otherwise be available as a free download from the company's website. But, in case you happen to need a heater-chiller circulator, then that company would be the right address.Thomas"Paul Rubin (AMT)" schrieb:Thomas,I'm quite interested in your comment about RS232 ports. We have not been able to resolve a problem with 2 applications trying to use the same Com port. While we can "Close" the port from the Vee application, there appears to be no way to reopen it after the other application (NOT VEE) releases it. I imagine it would be possible to replicat the other application in VEE, if absolutely necessary, but that would be a significant undertaking as there are a number of different combinations possible.Please advisePaul H. Rubin Staff Engineer Test Engineering Ericsson Amplifier Technologies, Inc. 49 Wireless Blvd. Hauppaugh, New York 11788 Phone: (631) 357-8513 E-Mail: paul.rubin@ericsson.com -----Original Message----- From: Thomas Vauderwange [<A HREF="mailto:Thomas@Vauderwange.de">mailto:Thomas@Vauderwange.de</A>] Sent: Friday, December 14, 2001 4:04 PM To: vrf@lvld.agilent.com Subject: vrf Multi-VEE-ing capabilityHi Shawn,Indeed it is no problem at all to have several different VEE programs running on the same PC the same time. Even tricky stuff, like RS232-comms etc. My customers and the folks down in the production hall do this all the time with my software (Win98, NT4, W2K), and everybody's happy. This happened when I changed from good ol' VEE3-interpreter (where it did not work) to the one that came with VEE5. And: someone found out by coincidence by just mistakenly starting several processes. Even sharing one and the same COM-port (!!) worked without having been specified! All I had to do to make this a real feature is to specify different working directories where different config-files could be placed. And, of course, to allow for an identifier string in the left top of the temperature graph, because otehrwise nobody could tell the processes apart...Regards,ThomasShawn Fessenden schrieb:> Well thank you everybody for your comments! > > There are some real eye-openers there, especially Jens. While I strongly > disagree with VEE's affinity for "structured" code, I suspect that's simply > because we have different ideas of what structured code is but, "that's > probably a discussion more suited to a comp.programming group," (he said, > trying to stay on topic). > > On the other side of the fence, VEE's innate flowchart-like diagramatic / > execution model is definitely a godsend. In fact, whenever customers > explicitly request VEE this model is almost always the reason why. When I > first started using VEE, I just assumed customers wanted VEE because it was > an HP product. > > Anyway Jens, I found this particularly interesting: > > >2 VEE clients and the VEE-controller > >can run on the same PC or on different PCs. > > If I understand this correctly you can run *three* instances of VEE on *one* > computer? If this is the case, then my hat is off to you and I will > henceforth keep my mouth shut! > > Using NT 4.0 I've had terrible problems just keeping *one* instance going on > anything less than a PII 233! At the risk of boring everybody, let me > outline two of these problems. I really need to understand why it is I have > such awful problems and others do not. At the time, we had HP field > application engineers crawling all over what I did and their conclusion was > "It's fine, that's just the way it works". Ever since then I've been saying, > "well then it doesn't work right". > > These two problems are actually just different aspects of the same root > cause, CPU hogging: > > 1. Printing. Customer wants a simplistic graph of readings over time. Using > the Print Screen object, it took over 45 minutes to print one graph. This is > over a 100MB network. Spooler priority was set highest. I abandoned the > spooler and of course that speeded up the printing to a minute or two, but > the UI completely stops responding (of course) and that's unacceptable. I > wound up using Microsoft's DIBAPI to capture the screen myself & print it. > The catch was that this utility had to shift itself to HIGH_PRIORITY_CLASS > to produce a graph in under 5 minutes. This slowed down the UI some, but the > customer accepted the result as a necessary trade-off. > > 2. Browsing. The test scenario is a room full of some 100 UUTs. Test time > varies from 4 hours to three or four days sometimes. Data points consist of > samples of 6 channels every 30 seconds. The DAC box is an Agilent 34970, one > per UUT. Loading the test line is pretty simple. One plug for power, one for > 4 data channels. The other two channels are from sensors placed inside the > UUT by test personnel. A bar code is scanned and the test cycle begins. > > Now, the people who run the test system are union laborers, not engineers. > There are usually three people per test room. It takes all of about an hour > to get a full room under way and then they sit back and wait. It's no fun > sitting there staring at a graph or a summary of all running tests, so they > usually start playing solitare or browsing the web, looking for girly-pix. > Now granted that this isn't the ideal scenario, but realistically it's to be > expected. As soon as VEE is put in the background, problems develop. > > In the first place, after about 40 UUTs VEE starts having a hard time > keeping up with sampling data, updating 34970 displays, checking for card > swaps or entire 34970 swaps, verifying 34970 configuration, analyzing data > to decide on shifts in test phases, testing com ports for proper > communication, watchdog pinging other test computers to make sure they're > ok, etc. This is even with VEE in the bloody FOREGROUND. When it gets put in > the background, look out! > > Now I realize full well that I'm no VEE expert. Like I've said, 30 years of > programming doesn't mean a thing when it comes to VEE so we called in HP > experts for expert advice. After over half a year of tweaking and some > truely extraordinary efforts on the part of various HP engineers, the > customer was still dissatisfied and started threatening to pull out, and for > that I don't blame them. They dumped *huge* amounts of cash on hardware and > they couldn't even run at 50% capacity. > > Even then, memory leaks caused by the infamous error pins would, after only > a few days of continuous operation, cause VEE to reset it's heap (or > something) and all in-progress tests were lost. Granted, there was always a > current status backup stored on the drive, but this was updated every time > data changed on any DAC box and the moment VEE lost it's variables this file > was overwritten with emptyness. The individual data files themselves > remained, but without status information (current test phase, phase test > instructions, current machine state, state machine registers, etc.) the > state machine couldn't possibly continue. > > It was finally decided to move to a compiled language so I locked myself in > a closet for 14 weeks and re-wrote the entire thing in C++. And when I say > re-wrote, I mean exactly that. The C++ version is almost an exact duplicate > of the VEE version. VEE's failures in this instance were definitely not due > to software architecture. This application is currently buzzing happily > away, testing like crazy. There are no issues with printing graphs, playing > solitare or browsing the web. > > The *opinions* I express here are based largely on this one extraordinarily > negative experience. This was definitely the most detailed analysis I've > ever made of VEE on Windows in the wild, though I've had plenty of somewhat > negative experiences with it before. The icing on the cake was HP's own > conclusion that "yes, everything is fine. That's the way it works". Since > then I've recommended to those customers who want to use VEE that they then > should stick with HP-UX. Sometimes that's fine, and sometimes it's not > because Windows is a requirement. In that case, we use "Windows specific" > tools like Visual Basic, Visual C++, Delphi or C++ Builder. > > Now I challenge any of you to put yourself in my shoes: Put your life on > hold for 14 weeks of consecutive 16-20 hour days WITHOUT PAY, bear the whip > from management AND the customer, watch HP go over everything you've done > with a fine-tooth comb and shrug their shoulders, watch five years of your > life helping to build a company from a two-man operation to a small global > enterprise go down the drain as law suits and bankruptcy ensue and then have > anything positive at all to say about VEE. Here it is, on public record: > > IN SPITE OF ALL OF THIS I STILL LIKE VEE!!! > > So please, if anybody has anything to say to me personally about VEE bashing > then lets keep it off the list as it probably won't be productive at all and > I guarantee I'll respond to all flames with a blow-torch. > -SHAWN- > > _________________________________________________________________ > Join the world's largest e-mail service with MSN Hotmail. > <A HREF="http://www.hotmail.com" TARGET="_blank">http://www.hotmail.com</A> > > --------------------------------------------------------------------- > This is the "vrf" maillist, managed by Majordomo. To send messages to > this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and > unsubscriptions are done through the address "vrf-request@lvld.agilent.com". > If you need details, just send a message containing the text "help" > to "vrf-request@lvld.agilent.com". > ------------------------------------------------------------------------------------------------------------------------------------------ This is the "vrf" maillist, managed by Majordomo. To send messages to this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and unsubscriptions are done through the address "vrf-request@lvld.agilent.com". If you need details, just send a message containing the text "help" to "vrf-request@lvld.agilent.com". --------------------------------------------------------------------- </BODY></HTML>

No one else has this question

Outcomes

2 Replies

bwalden Dec 16, 2001 12:49 AM
Mark Correct
Correct Answer
Hi Folks,

Can anybody tell me how a particular Vee application can tell if another
copy of itself is already running when it starts? I'm using Windows
2000.

Thanks in advance.

Barrie
whoi

---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
Like • Show 0 Likes0
Actions
VRFuser Dec 16, 2001 9:40 AM
Mark Correct
Correct Answer
A runtime-only Linux version of VEE as a first step faces at least one very
serious obstacle: programs created on one platform are almost guaranteed to
fail in new and mysterious ways when run on another platform. Without a
development environment on Linux we will have little hope of figuring out
what went wrong when our programs die.

...david

David Sherrill
Ultrasound Software Engineering
Philips Medical Systems
---------------------------------------------------------------------
This is the "vrf" maillist, managed by Majordomo. To send messages to
this maillist, just email to "vrf@lvld.agilent.com". Subscriptions and
unsubscriptions are done through the address "vrf-request@lvld.agilent.com".
If you need details, just send a message containing the text "help"
to "vrf-request@lvld.agilent.com".
---------------------------------------------------------------------
Like • Show 0 Likes0
Actions

vrf multiple copies running

Outcomes

Related Content

Recommended Content