Blue circles and line become gray and unresponsive

[jlazovskis]

The blue circles and the adjustable line become unresponsive sometimes.

I loaded the file "senate104_RIVET_Codensity20thPercentile.txt" with homology 0, 40 x-bins and 40 y-bins. Then in the "Line Selection Window" when I drag the blue circles on the window boundary, both circles and the line become gray if the left endpoint is at the origin and the right endpoint is dragged right-to-left across the top-left corner of the diagram (I can sometimes reproduce the issue by also dragging the left endpoint left-to-right towards the origin). Then the window looks as in the screenshot below.

Eveything else works - I can click on the bars, hovering over regions shows the dimension, the bottom options are accessible, I just can't move the main line or its endpoints. By clicking on the "Angle" or "Offset" up or down directions I can make the gray circle blue again, but the graph looks jumbled, as in the screenshot below.

As soon as I drag either of the blue circles again, everything goes back to normal.

I cloned Rivet from this repository on the morning of 2018-08-15. My system is as follows:

• Linux Ubuntu 17.10
• Intel Core i7-4510U CPU @ 2.00 GHz
• RIVET 0.4
• CMake version 3.9.1
• QMake version 3.1
• Qt version 5.9.1
• Boost version 1.62

[mlesnick]

Jānis, thank you for this careful bug report. We will have a look. I'm not able to reproduce this on my OS X build, but I will see next week if I can reproduce it on my Linux machine.

On Fri, Aug 17, 2018 at 7:29 PM Jānis Lazovskis notifications@github.com wrote:

The blue circles and the adjustable line become unresponsive sometimes.

I loaded the file "senate104_RIVET_Codensity20thPercentile.txt" with homology 0, 40 x-bins and 40 y-bins. Then in the "Line Selection Window" when I drag the blue circles on the window boundary, both circles and the line become gray if the left endpoint is at the origin and the right endpoint is dragged right-to-left across the top-left corner of the diagram (I can sometimes reproduce the issue by also dragging the left endpoint left-to-right towards the origin). Then the window looks as in the screenshot below.

[image: issue1] https://user-images.githubusercontent.com/14241342/44291882-f1075480-a246-11e8-9fd8-67b3d76acc64.png

Eveything else works - I can click on the bars, hovering over regions shows the dimension, the bottom options are accessible, I just can't move the main line or its endpoints. By clicking on the "Angle" or "Offset" up or down directions I can make the gray circle blue again, but the graph looks jumbled, as in the screenshot below.

[image: issue2] https://user-images.githubusercontent.com/14241342/44292093-66bff000-a248-11e8-9449-66f78445ee4e.png

As soon as I drag either of the blue circles again, everything goes back to normal.

I cloned Rivet from this repository on the morning of 2018-08-15. My system is as follows:

• Linux Ubuntu 17.10
• Intel Core i7-4510U CPU @ 2.00 GHz
• RIVET 0.4
• CMake version 3.9.1
• QMake version 3.1
• Qt version 5.9.1
• Boost version 1.62

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rivetTDA/rivet/issues/133, or mute the thread https://github.com/notifications/unsubscribe-auth/AL-lBO6R2I_QvcRZWGupasmW1oC9vlt3ks5uR1HggaJpZM4WCUgE .

[SimonSegert]

I was able to reproduce this on my Mac. Commit 63e8bb0 seemed to resolve the issue for me.

[mlwright84]

Thanks, Simon, for your work on this. Unfortunately, I'm still experiencing this problem, even with your recent commit. Every time I load the senate104 data set (H_0, 40x40 bins) and move the upper control dot to the left, the line turns grey. I'm using Ubuntu 18.04, with qmake 3.1, Qt 5.9.5, and gcc Ubuntu 7.3.0-16ubuntu3.

[SimonSegert]

Matthew, I’m using Qt Creator 4.5.0, with Qt 5.10.0 and Clang 7.0. Perhaps it has to do with your version of Qt?

On Aug 20, 2018, at 10:52 AM, Matthew Wright notifications@github.com wrote:

Thanks, Simon, for your work on this. Unfortunately, I'm still experiencing this problem, even with your recent commit. Every time I load the senate104 data set (H_0, 40x40 bins) and move the upper contro

[mlwright84]

Nevermind; my mistake. I forgot that we changed the name of the executable. Out of old habit, I was running RIVET instead of rivet_GUI. The issue appears to be fixed!

[mlesnick]

Thank you Simon and Matthew. Jānis, if you find that the fix Simon pushed works on your system, feel free to close the issue.

On Mon, Aug 20, 2018 at 3:28 PM Matthew Wright notifications@github.com wrote:

Nevermind; my mistake. I forgot that we changed the name of the executable. Out of old habit, I was running RIVET instead of rivet_GUI. The issue appears to be fixed!

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rivetTDA/rivet/issues/133#issuecomment-414434438, or mute the thread https://github.com/notifications/unsubscribe-auth/AL-lBP-jW7I8XZmmo8k1FcfBPijI6Owpks5uSw3ugaJpZM4WCUgE .

[jlazovskis]

Using Simon's commit (63e8bb0), everything works! So I suppose this is indeed closed, thanks for your work on this.

However (not sure if I should be posting this here or somewhere else) - using the latest commit (1534e8a), I get a different error. After running ./rivet_GUI in the rivet directory and loading the same file with the same parameters, the progress bar window appears, but as soon as it reaches 100% it crashes, with the last few lines from the terminal as below.

console:  PROGRESS 30489

console:  PROGRESS 30490

console:  BARCODE TEMPLATE COMPUTATION COMPLETE: path traversal and persistence updates took 1270 milliseconds 

console:     building the line arrangement and computing all barcode templates took 2058 milliseconds 

console:  ARRANGEMENT: 

Loading file 
terminate called after throwing an instance of 'std::runtime_error'
  what():  Couldn't open  for reading
Aborted (core dumped)

I tried using ./rivet_console with the same file and parameters, which seemed to run fine, but it didn't write to any file, terminating with simply Wrote arrangement to. Running the same command in Simon's commit the output file is created, terminating with

Wrote arrangement to out.mif
Writing file: out.mif 

[mlwright84]

Thanks, Jānis. It looks like I broke something in my most recent commit. I'll try to fix it now.

On Wed, Aug 22, 2018 at 11:09 AM Jānis Lazovskis notifications@github.com wrote:

Using Simon's commit (63e8bb0 https://github.com/rivetTDA/rivet/commit/63e8bb06e595b484ad3784874f261bbccb7bf656), everything works! So I suppose this is indeed closed, thanks for your work on this.

However (not sure if I should be posting this here or somewhere else) - using the latest commit (1534e8a https://github.com/rivetTDA/rivet/commit/1534e8a4eee6a860b427702bdb5bf3b3b5ba3bf8), I get a different error. After running ./rivet_GUI in the rivet directory and loading the same file with the same parameters, the progress bar window appears, but as soon as it reaches 100% it crashes, with the last few lines from the terminal as below.

console: PROGRESS 30489

console: PROGRESS 30490

console: BARCODE TEMPLATE COMPUTATION COMPLETE: path traversal and persistence updates took 1270 milliseconds

console: building the line arrangement and computing all barcode templates took 2058 milliseconds

console: ARRANGEMENT:

Loading file terminate called after throwing an instance of 'std::runtime_error' what(): Couldn't open for reading Aborted (core dumped)

I tried using ./rivet_console with the same file and parameters, which seemed to run fine, but it didn't write to any file, terminating with simply Wrote arrangement to. Running the same command in Simon's commit the output file is created, terminating with

Wrote arrangement to out.mif Writing file: out.mif

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rivetTDA/rivet/issues/133#issuecomment-415088170, or mute the thread https://github.com/notifications/unsubscribe-auth/AXqxANT9O5dnfSLlCUMgHrvHsYDtFCmRks5uTYJBgaJpZM4WCUgE .

[mlwright84]

Jānis, I think fixed the most recent problem in Commit 4dd01de. Please try again and let us know if you encounter further problems. Thanks!

[jlazovskis]

Everything works fine with the latest commit, no more issues. Thanks again.

[mlwright84]

Great! Thanks for testing RIVET and helping us identify these problems!

On Wed, Aug 22, 2018 at 12:19 PM Jānis Lazovskis notifications@github.com wrote:

Everything works fine with the latest commit, no more issues. Thanks again.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rivetTDA/rivet/issues/133#issuecomment-415109959, or mute the thread https://github.com/notifications/unsubscribe-auth/AXqxAKXculH7eaYItXhWCq-xvJL_S7Mqks5uTZKTgaJpZM4WCUgE .