Author |
Message |
Jared Barber (jared_barber)
Member Username: jared_barber
Post Number: 6 Registered: 01-2007
| Posted on Friday, June 08, 2007 - 02:12 am: | |
I have a program (written in C) which calls flexpde5.0.15 with 64-bit addressing. It calls it multiple times. Every 1 out of maybe 4-10 times it comes up with "memory protection fault". When the memory protection fault comes up, the flexpde window stays open and I ask it to try and run the same program again (which it just ran to make an error). Second time it works just fine. Are there any issues that I should be aware of in terms of why this code might have troubles running on Windows Vista or running in a 64-bit capacity? I realize FlexPDE may not yet be ready for Windows Vista, but any help on this you can give would be nice. Thank you, Jared |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 880 Registered: 06-2003
| Posted on Friday, June 08, 2007 - 10:56 pm: | |
We tested FlexPDE briefly on Vista when it was first released, and were convinced that FlexPDE ran correctly. However, we have not done heavy testing on Vista, and never tried the kind of exercise you are doing. We will have to run some more rigorous tests and report later.
|
Michael (madcat)
New member Username: madcat
Post Number: 1 Registered: 11-2007
| Posted on Friday, November 09, 2007 - 04:36 am: | |
Hello at all, I use FlexPDE on a Mac. Since last friday I've the new Mac OS Leopard. FlexPDE 5.0.18 and FlexPDE 5.0.19 produce sometimes a memory protection fault too. For examble: I start the solver and FlexPDE produce a memory protection fault. Than, without I've changed anything in the scriptfile, I run the solver again and FlexPDE works correct. I think thats a problem with the memory management from Vista and Leopard but I'm not sure. Under Mac OS Tiger and Windows XP I haven't such problems. Under Leopard text from the plots (title, scale etc.) are blurred so you can't read it. I'll post a screenshot at the afternoon here to show you what I mean. Is this problem also known? When will these problems to be solved? |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 985 Registered: 06-2003
| Posted on Friday, November 09, 2007 - 02:09 pm: | |
We do not have Leopard installed in-house yet. We do not experience any troubles with Vista. These failures must be related to some specific aspect of your script. Please send us a script that fails, and we will try to track down the problem. We will look into the blurring problem on Leopard. Try turning off the "font smoothing" setting. This was a problem in earlier versions of MacOSX.
|
Michael Köhler (madcat)
New member Username: madcat
Post Number: 2 Registered: 11-2007
| Posted on Friday, November 09, 2007 - 04:07 pm: | |
Here are my scriptfile with some datafiles. Under Windows XP and Mac OS Tiger I haven't any problems with these file. Only sometimes under Mac OS Leopard it doesn't works. How/where can i turn off the "font smoothing" settings?Data-File for Steel 4113 DIN 175 MWG 4113_H.tbl (1.9 k) |
|
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 986 Registered: 06-2003
| Posted on Friday, November 09, 2007 - 08:40 pm: | |
This clearly appears to be a font selection error, and will not be fixed by the font smoothing control. I also don't understand the funny file name. We will have to install Leopard and see what's going on. Font smoothing is in the "Appearance" control panel.
|
Michael Köhler (madcat)
Member Username: madcat
Post Number: 4 Registered: 11-2007
| Posted on Thursday, November 15, 2007 - 05:20 pm: | |
Hello Mr. Nelson, today I installed the new Mac OS update from 10.5.0 to 10.5.1. The font problem is history now.
|
Jared Barber (jared_barber)
Member Username: jared_barber
Post Number: 22 Registered: 01-2007
| Posted on Wednesday, December 26, 2007 - 03:06 pm: | |
Hey, Here's supposedly a script file that fails. Looking at Norm_Flow.dbg shows that a memory protection fault did indeed occur. I have copied this set of files from one file to another and it seems that I can reduplicate this memory protection fault. VERY IMPORTANT: This memory protection fault cannot be duplicated by opening up FlexPDE as one normally would...using a gui. I seem to be able to duplicate the error only by using the command line, in particular using the command line with the "-s" option: [directory of flexpde5]\FlexPDE5.exe -s [directory of FlexPDE Files]\Flex_Files1\Norm_Flow.pde The pde file of interest is the "Norm_Flow" file. By looking at the Norm_Flow.dbg, you can see that it crashed from a memory protection fault (so it has happened and I can get it to happen again easily). I have copied the attached files to different parts of my computer and have been able to reproduce the memory protection fault by using the command line commands above. Hopefully this can be verified on a vista machine as well by running the suggested command on a command line and then investigating (for instance), the dbg file afterwards. I note that if you run this by opening it in the actual window interface (the gui) and using your mouse to click on the "run" button, it seems everything is "reset" and the program will run just fine. In addition, it seems doing so may change the files in the folder so that the "-s" command line call would actually work from then on. Hence, to reproduce the error, do not load it from the FlexPDE window before running the command line call suggested. This has been frustrating because I want to be able to call FlexPDE multiple times overnight but cannot do so because of this incompatibility with Windows Vista. It works just fine on Windows XP (but I do not want to unload windows vista on my fast machine just so FlexPDE5 will work). I have investigated this further trying many different things, but before discussing these investigations, I just wanted to verify that someone might be able to get this memory protection fault (and if that someone is able to get this fault, maybe they may have an idea why it's happening). If you can't get this fault, at least you can see that I have been getting it by looking at the debug file. Thank you for your time and I will talk to whoever may be able to help later. Jared PS I can't get the "Upload Attachment" to work so I will send it via email to "rgnelson@pdesolutions.com" |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 1037 Registered: 06-2003
| Posted on Wednesday, December 26, 2007 - 04:28 pm: | |
The words "SCALAR" and "FINISH", which you use in your script, are deprecated in version 5. If you run the script from the GUI, you will get a warning dialog. This is supressed by the -s switch, so FlexPDE just exits. Normally, each of these only gets reported once, and a tag is written into the flexpde5.ini file, warning FlexPDE to accept the deprecated word. If you move to a new machine or user, so that the flexpde5.ini file is reinitialized, you will get two failures, until the special tags are written to the .ini file. Can this be what you are experiencing? The failure happens during the parsing of the script (ie, immediately), not after it has begun running. After working past the two failures, the script ran to completion on my machine (XP). Try changing SCALAR to GLOBAL and FINISH to CLOSE and see if the failures disappear.
|
Jared Barber (jared_barber)
Member Username: jared_barber
Post Number: 23 Registered: 01-2007
| Posted on Friday, December 28, 2007 - 12:20 am: | |
Hello, I note that I have never encountered this "Memory Allocation" error on my Windows XP machine so it doesn't surprise me that you encountered no errors on your XP machine. I don't think the error can reproduced unless you run it on a Windows Vista machine and unless you call it from the command line using a command like: c:\flexpde5\flexpde5.exe -s Norm_Flow.pde I would appreciate it if someone could try this on a Windows Vista machine to see if they can reproduce this error. If they can't then I can easily give more details of when and how and where this error occurs. If they can, then hopefully we'll have something to start with. I note, the script (and files) that I have given does not fail in all instances, it only fails in one particular instance, when you call it with the "-s" option from the command line. Using the program that I am trying to run (which calls FlexPDE5 multiple times) I can easily come up with scripts that will fail when you call it with the, e.g., "-q" or "-x -r" options as well. I hope someone can try this on Windows Vista. In the meanwhile I will try (though I'm pretty sure this is not the problem) changing my deprecated scalar/global and close/finish commands. Thanks for your time. Jared |
Jared Barber (jared_barber)
Member Username: jared_barber
Post Number: 24 Registered: 01-2007
| Posted on Friday, December 28, 2007 - 11:41 am: | |
Hello, I am able to reproduce the exact same thing when I use "global" and "close" so it doesn't seem like that's the problem. Jared |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 1039 Registered: 06-2003
| Posted on Friday, December 28, 2007 - 05:56 pm: | |
I got it to fail a few times on vista/64, and thought I had a repeatable failure, but now I can't get it to do it again. This is very obscure. I have a suspicion it has to do with file permissions, which Vista has seriously dorked up, but I can't prove it.
|