Desktop API crashes

[JimBecker17]

Ok - I haven't been on here for over a month, so here is the rant you have all been missing...

The eSignal crashing on use has been pointed out numerous times by myself and Chris James, as well as several others. It is not just realtime quotes, it is an asyncronous issue overall.

There are periods of time everything runs ok, but there is some BUG which causes crashes as well.

It's not based on inefficiency of design, it's a BUG in the software. It seems to be a NULL Pointer bug.

I've provided details and written a test application which show this problem. The test application will crash eSignal 100% of the time, doing the same thing over and over with no increase in load - just something to scare up the bug.

If the problem were merely poor design, then the net result would be things slow down but continue without crashing.

THIS IS A BUG WHICH IS PROBABLY QUITE SIMPLE TO FIX IF eSignal WOULD TAKE A LOOK AT IT. IT IS AN ASYNCHRONOUS ERROR. IT SEEMS OBVIOUS THEY ARE NOT USING THEIR API IN ANY REAL APPLICATIONS AS IT WOULD TURN UP AND THEY WOULD RESOLVE IT QUICKLY.

I am SCREAMING here because this has been pointed out since Summer and we continue to be told to totally switch APIs or other major work. Regardless of if there is EVER a replacement API, I for one am not going to spend hundreds more hours doing analysis on why something from eSignal does not work and working around it.

I continue to contend with some focus on looking into this, rather than finger pointing or promising something else, the problem would be resolved. I bet it could be resolved in a couple hours with my provided test program and a little coffee.

Last night got a dozen or so crashes of the same thing, and this morning probably over thirty. Same bug. There is basically NO LOAD on eSignal during a lot of these - it's just a BUG. (i'm sitting here typing this and it crashes a few times, not much going on except getting five minute quotes here and there.)

If those from eSignal will note, I have had to e-scream and e-yell to get you to recognize that RequestHistory() and the T&S data were in error, after being told it was either not a problem or my coding error. Eventually they were fixed, but months were spent convincing someone to investigate on your end.

This is the same deal. Please fix.


-Jim

[JimBecker17]

For what it's worth - here is a test application which will crash the latest eSignal in my world. Project is VC++ 6.0.

It does the same thing over and over, should crash in four to eight minutes. Counts on having 1000 symbols, but can be changed on line 526 of eSignalTestDlg.cpp.

It packages eSignal into a class derived from CButton, which makes dropping their API into an application pretty easy.

Select individual buttons to toggle on/off, or turn all on/off to get a lot happening in one thread but pseudo asynchronously.

The logic which restarts eSignal is not included.

-Jim

[wscully]

Desktop API crashes

I have a simple application that correlates information about stocks and stock options obtained from S&P, IBD, Valueline, and eSignal (on stocks in my portfolio and stocks that I am considering). Compared to a commercial application, my workload is trivial. I have numerous abends of eSignal during the day - and always with the same error. My annual subscription to eSignal is up in January and I have given up. I am looking for an alternative data supplier for realtime stock and stock options information. I saw one advertised today on CNBC (Trading Desk?) that claims to be highly rated by Barrons. Please let me know if you have any suggestions.

The error:
"DBC NL2 Message Sink: WinSig.Exe Application Error

The instruction at 0x5ad71531 referenced memory at 0x00000014 that could not be 'read'."

[JimBecker17]

Yes, same error. Search for postings by me and you will find many details about said error going way back - by myself and enough other people to play rugby against an eSignal team...

Sometimes it might show up differently, but this is the one they continue to ignore - driving away customers.

The totally sad thing is this has got to be something very easy to fix once someone will sit down to try and find it. But it doesn't happen. My sample app was provided in early October after months of trying to get them to look into this crash.

BTW the test application I provided can be used to test all kinds of calls, just change the logic in the OnTimer() routine to what one wants to press. Not that any errors found will be addressed.... but certainly they can be demonstrated that way.

-Jim

[DionLoy]

Hi folks,

Sorry for the lack of attention regarding this problem. eSignal will crash if you attempt to Unregister your listener while in the middle of a quote or bar callback.

This problem has been fixed in eSignal 7.8 due out in a week or two. Please email me though at dloy [at] mail.esignal.com and I will forward you the URL to the latest 7.8 beta for verification.

I've been running the program you had attached with 7.8 for awhile now with no ill effects (as you mentioned it definitely does crash 7.7 and earlier).

Thanks!

[Guest]

As Dion mentioned, we did correct an issue with the API. I am currently working with JimBecker, wscully and Dierk Droth to verify that their problems are in fact corrected.

I encourage you all to post your results, as there are some resources currently available to correct some of these hard to find issues.

[JimBecker17]

Ok, I take it all back! Well, actually wanted to report that eSignal read yesterdays diatribe/rant and directly contacted me.

Robi called and I emailed with DionLoy where he provided me with 7.8 beta candidate. If I read between the lines correctly on what Dion sez it is, sounds like the rug was pulled out under an object in it's handler. I've done this plenty of times myself, and it's no fun chasing it down.

The great news is this beta seems to have cleared up the 0x014 crash which I've been whining about for months.

For me this is a big thing - as it's been lots of crashes per day.

Thank you for listening and fixing the problems, guys! Really appreciate it.

-Jim

[Guest]

Great, now you can spread positive news about us.

Glad to hear your problem was corrected. Now, on to the rest!

[Enlighten]

http://www.esignal.com/hubs/matrix.asp

Is this link currently up to date ?

Thanks,

[Guest]

Our XML product has morphed a bit, but yes it is up to date.

[HeiDie]

I installed eSignal 7.8 today, recompiled a simple program and
got a crash of eSignal while exiting eSignal.

BTW the crash reporter did not work either.

The program in question only called the entitlement and worked on eSignal 7.6 reliably.

Anybody else seeing problems?

I sent the incident by e-mail to [email protected]

[Guest]

Yes, I have seen that problem myself. I was not able to add the correction before the 7.8 release, but this problem has been corrected and will be addressed in the 7.9 Beta version. If it poses a serious problem, you should roll back to 7.7, as there were not too many modifications to the API in the 7.8 release of eSignal.

[JimBecker17]

There is still a crash problem with exiting. Apparently eSignal has done some fixes for this in upcoming release - not sure 7.8 or 7.9.

The 0x014 crash error still pops up occasionally, but much less than before. I think running another app on the same machine using CPU cycles might tickle it a bit.

The CoreReports for 7.8 apparently go to a new file store, which normal users might not have upload access. I've alerted them to this situation.

On a different matter, I still haven't been able to consistently get eSignal to stop prompting me to do the "Switch Pages Without Saving". Keeps popping response dialogs up when loading layouts... The checkbox setting in Preferences seems to uncheck itself! Perhaps a bug that could be fixed?

I'm not complaining here - just some details and noting a nit.


-Jim

[DionLoy]

Hi Jim, does your sample app (the one with all the buttons in a grid) produce that 0x014 crash errror consistently? I've been leaving it running for awhile unable to crash it.

I'll check on that switch pages w/o saving problem.

Thanks.

[JimBecker17]

Hi Dion,

No, the sample app just barely caused the initial crash compared to my "real app". The "real app" caused the intial crash all the time. It also causes these subsequent crashes, with today being the prime day for crashes.. Don't know why but there are Monday blues with this crash.

I've probably gotten about twenty crashes this morning, that is much better than before the last fix. Normally other days of the week don't see that many crashes - but today for some reason there were a lot this morning. This is the highest have seen since getting 7.8 beta, generally it's been quite good.


-Jim