Problem with fsuipc 3.22 and newer

[fjacobsen]

I have a problem with fsuipc 3.22 and newer versions.

I´m currently running fsuipc 3.129 without any problems.

My problem with fsuipc 3.22 and newer is that if I chose a saved flight with the eurowings pro 2004 ATR or BAE or PMDG´s 737-600/700/800/900 and ends that flight my FS2004 selection screen (Main UI) becomes blank with no text and I have to ctrl+alt+del to close fs2004.

This doesn´t happen if I chose another saved flight with other aircraft, propably because they doesn´t acces fsuipc.

My system:

Amd XP2000 processor

512 Mb DDR RAM

GF2MX200 graphics card

Windows 98SE (Danish)

[Pete Dowson]

My problem with fsuipc 3.22 and newer is that if I chose a saved flight with the eurowings pro 2004 ATR or BAE or PMDG´s 737-600/700/800/900 and ends that flight my FS2004 selection screen (Main UI) becomes blank with no text and I have to ctrl+alt+del to close fs2004.

If you choose a different aircraft first, via the Aircraft menu, does it still occur? This may show whether it is a result of some activity in the advanced aircraft panel code, or actually something which has been corrupted by that code -- in the latter case you'd expect it still to fail even if you changed to another aircraft first.

When you say you have to close FS forcibly (Ctrl-alt-del), is this because FS is solidly hung? i.e. not responsive to anything at all?

If it is okay after changing aircraft, I can check the activity the aircraft has with FSUIPC at the time if you will do the following (with 3.30 installed please):

1. Just before exiting the flight which will cause the hang, go to the FSUIPC Logging page. On the left check the IPC Read and IPC write logging. On the right enter offset 3364 as type U16 and check the "hex" option, then, below, check the normal log option.

2. Ok out of the FSUIPC options, then (with a couple of seconds gap) press ESC and exit the flight.

3. If it hangs as before, close FS, find the FSUIPC.LOG file in the FS Modules folder, ZIP it up and send it to petedowson@btconnect.com with some brief explanation.

Meanwhile, this also sounds rather like a video driver problem. Possibly the only reason it happens with one version and not another is that the memory arrangements will be slightly different, or the timing between one even and another is slightly different. You could probably get different results too by re-ordering the modules in the FS modules folder.

There was one other instance of this reported here someplace -- but this time when using Active Sky 2004, not specific aircraft. It may help if you compared notes with the other user. Possibly you have the same video card and/or driver version? Or maybe there's something else in common you can identify?

Regards,

Pete

[fjacobsen]

Thanks for Ypur quick response Peter.

Well I don´t think that FS2004 is hung because if i press Ctrl+Alt+Del I get the task manager wich shows FS2004 running, but since theres no text I can´t make selections.

I´m currently at work in the other end of my country so I have to wait until I get home on thursday.

What i have read in Your advanced userguide and around different forums I think I might try the rearrange thing with the files in the modules folder, and or the selection of another aircraft before ending the flight with the above mentioned aircraft.

Ofcourse I could stick with version 3.129 but my TCAS on the Eurowings line of aircraft does only work with the newer fsuipc versions (3.30 for instance)

Again thank You very much for Your help

fjacobsen

[Pete Dowson]

Well I don´t think that FS2004 is hung because if i press Ctrl+Alt+Del I get the task manager wich shows FS2004 running, but since theres no text I can´t make selections.

If it simply the display which is wrong, then the video driver is the prime suspect. If Windows doesn't report it as "not responding" then it must still be processing messages and the like. So you may be able to close it more tidily by clicking the close window button/icon (x in top right corner or active window).

Also, see if it the same in both full screen and Windowed modes -- i.e. try the other, the mode you are not using.

I´m currently at work in the other end of my country so I have to wait until I get home on thursday.

What i have read in Your advanced userguide and around different forums I think I might try the rearrange thing with the files in the modules folder, and or the selection of another aircraft before ending the flight with the above mentioned aircraft.

Well, before you work around it, if it looks like some interaction between the aircraft and FSUIPC is responsible, please do send me that Log as I described.

Regards,

Pete

[Pete Dowson]

...but since theres no text I can´t make selections.

There are two other tests you can do here. Both ESCape and Enter should do something in the dialogue even if you don't see any text.

Here, pressing ESCape is the same as clicking the Close (X) Icon top right, and closes FS tidily.

Enter or Return operates the "Fly now" button (which I know you don't see), and effectively re-selects your last selection.

Can you try both of those to see if they operate?

Finally, there was a change in 3.22 to the way FSUIPC polls joystick buttons which did affect some folks adversely (not in any way like your problem though). I did fix this in 3.30, but just in case there's anything going on related to that, please try this:

Edit the FSUIPC.INI file, find the [buttons] section, and add the lines:

PollEpicbuttons=No
PollInterval=66

Do this without FS running, then load and run FS and see if the problem still occurs. Let me know please.

Regards,

Pete

[fjacobsen]

Yep I will try what You mentioned, but I´m not home until thursday, so once I have tried those "fixes" on thursday I will let You know.

Thanks again

fjacobsen

[fjacobsen]

Hallo Peter.

I Finally solved my problem by adding the lines You mentioned above to the [buttons] section of the fsuipc.ini.

But since I´running an unregistered version of Your DLL, the buttons section were not present in my ini file so I simply tipped in the [buttons] section with the lines beneath.

Now everything seems to be Ok.

Thank You very much for Your help.

I think i´m going to register Your "product".

Fjacobsen

[Pete Dowson]

I Finally solved my problem by adding the lines You mentioned above to the [buttons] section of the fsuipc.ini.

But since I´running an unregistered version of Your DLL, the buttons section were not present in my ini file so I simply tipped in the [buttons] section with the lines beneath.

Oh, rightwell done!

Thanks for getting back to me about this.

I wonder what is going on there to cause such a problem? The default button polling cycle is 25 mSecs now, in 3.30 (it was 20 in 3.22, but 60 previously). This should be easily enough time -- but it seems some systems have certain states in which it is taking all that time just to poll buttons on any joysticks connected (USB or game port), so nothing else gets done. The only other instance I had was where the buttons dialogue in FSUIPC was hanging -- this was on a system with a lot of USB connected devices (GoFlight and such).

I will try to figure out a way of keeping the default cycle of 25 mSecs (which is better for faster button detection), yet still prevent this sort of hang. If I think of anything, would you mind testing a version for me if I send it by email attachment? You need to remove those [buttons] lines you added, then try to make the problem recur.

Regards,

Pete

[ronzie]

Little Deja Vu, Pete? See this May 18th thread:

http://forums.simflight.com/viewtopic.phlight=fs9

for details:

Same symptoms. ESC worked to exit in the blank menu screen, any other key crashed.

The good news is that your posted {buttons} entries for this user also RESOLVED my problem on a quick test with a loaded parked aircraft, RC 3.1, and AS2004 all running. Got a clearance from RC, had AS weather refreshed, shut down RC, quit the flight, and the menu text came back and I was able to make further selections!!!

Without this user's details, the only similarity I see is that both of us did not have a {buttons} section in the .ini if I am reading the thread here correctly. If the original poster would kindly jump to that thread (way back about 18 MAY when 3.22 was introduced), maybe he can compare system specs to get some commonality clues.

The problem has carried into 3.30 and is now resolved. (I also let 3.30 build its own .ini file but that did not cure the problem. Does the buttons section only get created when something is assigned in that tab?) Perhaps the buttons section and some default entries are required regardless of EPIC output being applicable or not.

Some PLEASANT news for your recovery period.

I also wonder whether these statements might be required for the extreme possibility that some of your polling services might be looping or going out of bounds due to no default parameters and might be causing stuttering (as suggested on some other forums). Probably not applicable but just a thought.

The old "C Programmer's Rule". Never leave any variable uninitialized. If it is zero'd by the compiler, the routine may be out of bounds as I probably am right now for suggesting this .

Glad to hear your health is doing well with all this eye mess. Hope you enjoyed the Expo with Ray.

[Pete Dowson]

Little Deja Vu, Pete? See this May 18th thread

Yes, I did say a few messages back in this thread "there was one other instance of this reported here someplace", but I thought it was reported as related to AS2004? Sorry, maybe I have mixed things up.

The only reason I suggested these measures was on the off-chance that the hang was in fact a tight loop into Windows APIs which, for some reason unknown to me (still) take a helluva long time on some folks systems in some circumstances.

The only reason I found out that these loops can occur was on another chap's system where he did not report this symptom, but a hang in FSUIPC's own Buttons page. This was with 3.22. After quite a lot of experiments I found the solution, and that was incorporated into 3.30.

At neither time did any relationship with these other symptoms become apparent. This was just a guess on my part after some of the answers I got earlier this thread. This would have happened in your case too if the other (resolved) problem had been reported and investigated beforehand.

the only similarity I see is that both of us did not have a {buttons} section in the .ini

Yes, but that's not really relevant.

Does the buttons section only get created when something is assigned in that tab?)

Yes, it is normally only a list of the assignments made. There are some optional entries, but the defaults are preferable except in exceptional circumstances (like these).

Perhaps the buttons section and some default entries are required regardless of EPIC output being applicable or not.

Actually, I'm pretty sure the Epic part of the addition I suggested is not needed. It's something to do with the polling interval being the same as or even less than the time Windows is taking to check on button status when I poll it. 25 milliseconds is a long time, so this is strange, but that's what it looks like.

Perhaps you can verify that for me? Delete the EPIC line and test again. Thanks.

I also wonder whether these statements might be required for the extreme possibility that some of your polling services might be looping or going out of bounds due to no default parameters and might be causing stuttering (as suggested on some other forums). Probably not applicable but just a thought.

No, not applicable, but thanks. What's this about stuttering? The only real internal cause of stuttering I know in FS2004 is weather setting, and in particular, clouds.

The old "C Programmer's Rule". Never leave any variable uninitialized.

None are, don't worry. there are defaults for everything. That's how FSUIPC can run with no starting INI file.

The difference here is a value of 25 (default now), 20 (default in 3.22) and 60 (default before 3.22). The parameter line to change that has always been available, ever since the button facilities were added. The reason for the decrease was to give better button servicing on rotary switches and the like. These side effects are completely unexpected and rather strange.

Do you want a test version of FSUIPC where I avoid button polling altogether when in FS dialogues? You'd need to delete those lines in [buttons] and try to reproduce the problem. I can send it now if you can test.

Regards,

Pete

[Pete Dowson]

There's one common factor which may be important in both these systems:

W98SE

Do you, by any chance, have a number of USB and/or joystick devices?

I'm wondering if this timing problem is down to Win98's rather half-hearted implementation of USB support. I can't seem to find operating system details for the chap with the Buttons page hang.

Regards,

Pete

[ronzie]

Yes, there are a number of USB devices via a USB 2 powered hub and all are USB 1.1 transport. These are all C&H products without their Control Manager since on my system that was not too swift:

568 COMBAT STICK (no room on my desk for a yoke)

PRO THROTTLE

RUDDER PEDS

I'll run a test without the EPIC line but leave the timing parameter under the BUTTONS section.

If you like send me your suggested version that diables the button polling in the menu. However, would that if it is the problem affect the initial startup menu as well since FS always loads a default flight in the background (verified by your log)? The problem only occurs on a flight exit.

You have my e-mail address and an attachment of that size is not a problem. Are we UP late tonight? You are five hours ahead .

We'll let why I'm on 98SE still go for now. I will say that 2000 refused to get through the hardware detect on my genuine Intel everything! 98SE does seem to have better timing loops and is mandatory (over 98) on over 1 Gb cpus for network related dlls to load (as documented by Intel).

[ronzie]

Perhaps you can verify that for me? Delete the EPIC line and test again. Thanks.

Took out the EPIC line leaving the pollinterval in. It still works OK.

[Pete Dowson]

Yes, there are a number of USB devices via a USB 2 powered hub

Right. That's two out of three so far.

If you like send me your suggested version that diables the button polling in the menu.

It is on its way.

However, would that if it is the problem affect the initial startup menu as well since FS always loads a default flight in the background (verified by your log)? The problem only occurs on a flight exit.

Not sure what the question is here. Sorry.

Are we UP late tonight?

My normal bedtime is around 2 am. I went up soon after sending that last message.

Regards,

Pete

[fjacobsen]

Hello Peter.

Yes offcourse I will try out your "test" version if that can help you solve the problem.

My email adress: fjacobsen@post.tele.dk

By the way I only have one USB device connected to my computer, it´s a Saitek Cyborg 3D joystick for wich I have made a custom FS2004 profile.

fjacobsen

[Pete Dowson]

Yes offcourse I will try out your "test" version if that can help you solve the problem.

Okay, it is on its way. Thanks.

By the way I only have one USB device connected to my computer, it´s a Saitek Cyborg 3D joystick for wich I have made a custom FS2004 profile.

Ah, that spoils one theory then .... hmmm.

Regards

Pete

[fjacobsen]

While waiting for Your mail...

.... I don´t know if it´s to any help, but I think to remember that in my BIOS somthing like USB DOS support is disabled.

Regards

fjacobsen

[fjacobsen]

I have tested your new version and it´s ok

I have also tested the 3.30 version with different numbers, here is my findings:

25 nogo

30 nogo

35 nogo

40 nogo

45 nogo

50 Yes

55 Yes

60 Yes

fjacobsen

By the way I´m just about to go out shopping with my wife so I will return later.

Regards

fjacobsen

[Pete Dowson]

I have tested your new version and it´s ok

Just to be sure, that IS with those [buttons] parameters removed, I assume?

I have also tested the 3.30 version with different numbers, here is my findings:

25 nogo

30 nogo

35 nogo

40 nogo

45 nogo

50 Yes

55 Yes

60 Yes

Phew! Right up to 45!? That is strange.

In that case I can hardly think it's down to timings in the Windows calls I make. It must be something else going on there. I can certainly make the default 50 for Win98/98SE/Me instead of the current 25, but I think I'll try to dig deeper first, see if I can figure out what is really going on here.

Thank you very much for testing. You can continue to use 3.305 without the [buttons] lines for now. The ultimate correction(s), whatever they may be, will be in the next formal release, probably 3.31, but that won't be for several weeks.

Regards,

Pete

[fjacobsen]

Yes the test with your 3.305 version was with the [buttons] section removed.

The test with the different values was done with the 3.30 version with theese lines:

[buttons]

PollInterval=xx

where xx was the numbers tested (25, 30, 35, 40, 45, 50, 55, 60)

Regards

fjacobsen

[Pete Dowson]

Yes the test with your 3.305 version was with the [buttons] section removed.

Great! Thank you!

Pete