Steema Central

Hello,

I have the latest version of TeeChart (7.0.0.2 of the ocx). I recently upgraded from version 6.

I am having problems with the chart throwing access violations and invalid indexes. I am turning off auto paint, using AddXY to insert values, turning back on auto paint and then doing a refesh.

When it crashes the chart continues to work and my app does not exit. It just shows a message box.

It could be something I am doing wrong, and I would like to get this fixed ASAP because we are releasing a new version of the app that uses TeeChart.

I have a sample VC++ 2003 project that will cause the crash. Where should I send it?

Hello Xenos,

I have the latest version of TeeChart (7.0.0.2 of the ocx). I recently upgraded from version 6.

Please notice that our latest version is TeeChart Pro v7.0.0.3 ActiveX.

I have a sample VC++ 2003 project that will cause the crash. Where should I send it?

Please post it at [url]news://www.steema.net/steema.public.attachments[/url] newsgroup.

narcis wrote:Hello Xenos,

I have the latest version of TeeChart (7.0.0.2 of the ocx). I recently upgraded from version 6.

Please notice that our latest version is TeeChart Pro v7.0.0.3 ActiveX.

I have a sample VC++ 2003 project that will cause the crash. Where should I send it?

Please post it at [url]news://www.steema.net/steema.public.attachments[/url] newsgroup.

Hello Narcis,

I tried version 7.0.0.3 with the same result.

When posting to the newsgroup I keep getting the message that the attachment is too large (it's less than 400K).

Can I just email it to you?

Thanks.

When posting to the newsgroup I keep getting the message that the attachment is too large (it's less than 400K).

Thanks.

I just rechecked, and the post is there now. It did give me an error when posting it though.

Hello,

Have you guys been able to reproduce this issue with the project I sent?

We really need to get this fixed before the next release of our product which will be later this week.

Thanks.

Hello,

This is possibly related to Axis management whilst adding data to the Line Series. We are investigating. As an interim workaround you may find the FastLineSeries better suited to this kind of data volume.

Regards,
Marc Meumann

Marc,

Thank you for the reply. We have changed the series type, and it does seem to help the problem. Although it still happens with a large number of points, it is better than before.

The problem seems to be worse on slower systems.

Any idea on when this problem will be resolved?

Thank you.

Marc wrote:Hello,

This is possibly related to Axis management whilst adding data to the Line Series. We are investigating. As an interim workaround you may find the FastLineSeries better suited to this kind of data volume.

Regards,
Marc Meumann

Hello,

I just tried 7.0.0.4, and the problem is still there. Do you have an idea of when it will be fixed?

We are getting complaints from several customers about this issue and will have to drop TeeChart for another component soon if this problem cannot be resolved.

Thanks.

Hello,

We've spent some time trying to check on the cause of this problem without any clear conclusions, it's proved very difficult to track down. It does not appear to be a load/thread problem as such, as reducing frequency to extremely low values doesn't erradicate the failure though the point quantity per cycle is a factor.

I'm sorry to have kept you waiting, we'll go in once again to see if we can uncover new clues and a solution. I'll post a message up here within a couple of days.

Regards,
Marc Meumann

Hello,

As a followup - as of now we still don't have a concrete TeeChart-internal solution to the problem you've described. TeeChart is Apartment threaded and theoretically as such thread-apartment-protected but seems to be susceptible to the loads in this test. One workaround option to avoid the problem would be to set a critical section around the data load routine. I don't recommend that as being the best final solution, but as a short term measure it will most likely work sufficiently effectively to avoid problems.

We will continue to look for a better generic solution.

Regards,
Marc Meumann

Marc wrote:Hello,

One workaround option to avoid the problem would be to set a critical section around the data load routine. I don't recommend that as being the best final solution, but as a short term measure it will most likely work sufficiently effectively to avoid problems.

Thank you for looking into this. We do appreciate it.

When you talk about a critical section around the "data load routine" do you mean in our code or in the TeeChart internal code? We did play around with it and add a critical section around our data load routine, but it did not help.

Sometimes the crash happens when only calling the data load once.

This testing was done using the same sample that I sent to you.

Again, thank you for looking into this. It is very improtant to us.

Hello,

Yes, the critical section tested was internal to TeeChart. It's not one we'd like to leave there though as it's application would be general by nature and would affect performance.

We'll take another look and see if we can make a clearer recommendation for an application based protected section or include a solution in TeeChart itself.

Regards,
Marc Meumann

Marc wrote:Hello,

Yes, the critical section tested was internal to TeeChart. It's not one we'd like to leave there though as it's application would be general by nature and would affect performance.

We'll take another look and see if we can make a clearer recommendation for an application based protected section or include a solution in TeeChart itself.

Hi Marc,

Would it be possible for us to get a custom "patched" version of TeeChart that has the critical section in it?

That would really help us until an official internal TeeChart solution is released.

Best regards.

Hello Xenos,

We'll run throught the checks. By Tuesday of next week (14th) I should be able to give you an answer about an interim version or alternative solution.

Regards,
Marc Meumann

Marc wrote:Hello Xenos,

We'll run throught the checks. By Tuesday of next week (14th) I should be able to give you an answer about an interim version or alternative solution.

Regards,
Marc Meumann

Hi Marc,

Any update on the status of this?

Thank you.