Topic: Graph 4.5 Beta - Display bugs

I have encountered two display bugs in that beta build 545, using Windows 7 64-bit and up to date, Python 3.2 32-bit.

Points won't appear when I load a file, no matter the size of the marker. Markers are shown in the legend, but not in the graph itself, see attached screenshot. I have tried:
-changing the marker size to 0 and then back to 2 and 5. Still no markers.
-adding a new point series with markers. Still no markers.
-Attempting the Graph.Redraw() function in the console. Still no markers.
-Closing and reopening the file. Still no markers.

On some files, upon loading, I see two sets of axes, see attached screenshot. I can work around that bug by changing the zoom and getting back to the original zoom, even through Undo. Some files never get that bug. Another file showed it on 8 out of 10 tries at opening-closing the file.

I hope this helps.

Post's attachments

Attachment icon Axis display bug.png 177.21 kb, 465 downloads since 2013-01-30 

Attachment icon Points display bug.png 170.66 kb, 441 downloads since 2013-01-30 

Re: Graph 4.5 Beta - Display bugs

I don't seem to be able to reproduce the problem with two axes. Could you please attach the grf file where you can reproduce it often?

I have however fixed the problem with point series not shown. You can find it in the new beta version:
http://padowan.dk/graph/DownloadBeta.php

I have also made some memory optimizations for large point series, so if you use the trick with setting Max undo steps to 0, I think Graph can handle a point series with a million points.

Re: Graph 4.5 Beta - Display bugs

Here's the file, I just checked that it still does the display bug. I hope it can be reproduced on another machine.

I'll download the new Beta and try big point series again.

Thanks!

Post's attachments

Attachment icon Convectrons.grf 56.91 kb, 773 downloads since 2013-02-01 

4 Reply by GraphUser3546 (edited by GraphUser3546 2013-02-02 00:03:55)

Re: Graph 4.5 Beta - Display bugs

I have tried "Set Undo steps to 0" with the new build 546 with large amounts of data. Working with a single instance of Graph opened at a given time, I have done the following list of operations (operations are described below):

a-c-d -> EEFFACE error
a-d -> EEFFACE error
a-b-d -> EEFFACE error
a-b-e-d -> OK
a-c-b-e-d -> OK

a- loading 1.3 million points, from a tabulation separated floats .txt file (39Mb)
b- saving the file
c- altering the data through a Python script (it changes the x-coordinate of all points, clears the old series, adds the modified series, changes axes)
d- Changing the name of the series or any other attribute, or one coordinate of a single point in the series
e- Closing and reloading the file.

It seems like closing and reopening the file has a beneficial effect on getting rid of the EEFFACE error after a memory-extensive operation. EDIT: I forgot to mention, I can alter the axes anytime without problems within these tests.

I hope this helps.

Re: Graph 4.5 Beta - Display bugs

GraphUser3546 wrote:

Here's the file, I just checked that it still does the display bug. I hope it can be reproduced on another machine.

Thanks. I can reproduce the problem now. I will fix it as soon as possible. For now you can press F5 to update the view when you see the problem.

GraphUser3546 wrote:

a- loading 1.3 million points, from a tabulation separated floats .txt file (39Mb)

I think 1.3 million points are too much. At least I couldn't make it work with 1.2 million either, however I could make it work with 1.0 million points.

Re: Graph 4.5 Beta - Display bugs

Agreed, that's a lot of points. But still, with closing and re-opening the file, I can work around the error. Sincerely, it is very impressive considering Microsoft Excel can only graph 65k points.

Thanks

Re: Graph 4.5 Beta - Display bugs

I have now fixed the problem with a second set of axes showing. You can find it in the beta version at: http://padowan.dk/graph/DownloadBeta.php

Thank you for the help in solving this.

Re: Graph 4.5 Beta - Display bugs

I can confirm, the two axes sets problem seems fixed.

Thanks!

Ivan Johansen wrote:

Thank you for the help in solving this.

and you're welcome.