OSWeb version 1.3.1

Comments

• sebastiaan

  May 2019

  Hi Saamah,

  > I made the gaze cuing experiment in the basic tutorial and converted it to JATOS and it seems to work.

  Good!

  > Does anyone know why this happens?

  There's an error message that you don't see. It depends on the browser how you can see that though. In Chrome, you can right click on the window, select Inspect and then open the Console tab, as shown below. What does it say there?

  

  > Also, how do you make it so that JATOS records the actual key response, rather than just recording whether it was correct or not? In the normal offline version of Open Sesame that is recorded as "response_keyboard_response".

  The response will be stored as the variable response.

  Cheers!

  Sebastiaan

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!
  
• saamah

  May 2019 edited May 2019

  Hi Sebastiaan,

  Thanks for the quick reply. On the response data, the variable response does not exist in the results file JATOS exported. On the running error, I did indeed find this error info. It says the following:

  OSWeb has stopped running due to a fatal error. osweb.1.3.1.bundle.js:1:102646
      addError http://127.0.0.1:9000/publix/5/5/js/osweb.1.3.1.bundle.js:1
      _callee$ http://127.0.0.1:9000/publix/5/5/js/osweb.1.3.1.bundle.js:1
      tryCatch http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:22
      invoke http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:22
      t http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:22
      asyncGeneratorStep http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:1
      _throw http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:1
  Error reading osexp: Error: Could not read remote osexp, TypeError: e.substr is not a function osweb.1.3.1.bundle.js:1:102711
      addError http://127.0.0.1:9000/publix/5/5/js/osweb.1.3.1.bundle.js:1
      _callee$ http://127.0.0.1:9000/publix/5/5/js/osweb.1.3.1.bundle.js:1
      tryCatch http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:22
      invoke http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:22
      t http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:22
      asyncGeneratorStep http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:1
      _throw http://127.0.0.1:9000/publix/5/5/js/vendors~osweb.1.3.1.bundle.js:1
  

  
  

  Many thanks

  Saamah
• sebastiaan

  May 2019

  Hi Saamah,

  Thanks! Could you attach the experiment here? And which browser (+ browser version) are you using?

  Cheers!

  Sebastiaan

  Ping @Daniel

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!
  
• saamah

  May 2019

  
  

  Practice2 - Tutorialgazecuing.osexp.zip

  Morning Sebastiaan!

  Thank you again. Here is the zip file produced when I exported the experiment using OSweb 1.3.1. I've been using Firefox 66.0.3. But I also tried it with Google Chrome and have the same problem.

  Thanks

  Saamah
• Daniel

  May 2019

  This must be a regression error. I will look into it soon!

  
• saamah

  May 2019

  Hi @Daniel - did you find a solution to this problem?

  Thanks

  Saamah
• Daniel

  May 2019

  Hi Saamah,

  I'm actually looking into it now and it just works for me, so the error must occur elsewhere.

  The `Error reading osexp: Error: Could not read remote osexp, TypeError: e.substr is not a function osweb.1.3.1.bundle.js:1:102711
  

  Give me the impression that this has to do with the loading of the file, and that there is an error in the error handler for that part (I know that sounds weird). I'm going to try and fix this, so we can get the real error message in your case of what is going wrong.

  I'll be back shortly.

  
• saamah

  May 2019

  Ok -thanks!
• Daniel

  May 2019

  Hi Samaah,

  I can't seem to reproduce the error, on my Jatos installation the experiment just works fine, so it's hard for me to pinpoint what is going wrong.

  There are only 2 places where the substr function is used, which is mentioned in your error message. Another element of the error message gives me the impression that the error occurs during loading of the filepool items, as if the name of a file is empty and thus not handled correctly...

  What version of Jatos are you using?

  @sebastiaan shall I create a temporary jatos.cogsci.nl account for saamah to see if it works for her there? If so, then this has to do with her Jatos installation, but I don't like it that the error message is not more informative, so do want to build in some extra checks.

  One thing that has changed between 1.3.0 and 1.3.1 is the unicode filename handling, or specifically

  currentFile.name.replace(/^.*[\\/]/, '').replace(
              /U\+([0-9A-F]{4})/g, (whole, group1) => {
                // Parse encoded characters back to their unicode counterparts
                return String.fromCharCode(parseInt(group1, 16))
              }
   )
  

  so it might have something to do with that too, although I don't understand why it just works for me then...

  
• sebastiaan

  May 2019

  > shall I create a temporary jatos.cogsci.nl account for saamah to see if it works for her there? If so, then this has to do with her Jatos installation, but I don't like it that the error message is not more informative, so do want to build in some extra checks.

  Yes, let's do that. It's a very strange error.

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!
  
• saamah

  May 2019

  Hi, thanks for taking this so seriously.

  I tried it on the JATOS test server (Cortex). It didn't work there either.

  Saamah
• Daniel

  May 2019

  @saamah, can you tell me your e-mail address and name in a private message so I can create an account for it?

  
• saamah

  May 2019

  Thanks @Daniel for finding the solution! For anyone else reading, it turns out that my problem was to do with the Adblocker. Needs to be deactivated when running experiments complied with OSWeb 1.3.1. Thanks Open Sesame team!
• Daniel

  May 2019 edited May 2019

  @sebastiaan detecting adblockers can be quite daunting, but shouldn't we warn people that experiments might not work correctly if adblockers are enabled? The flipside of the coin is that people may become wary that osweb may feed them ads if we issue such warnings, so I'm not sure what the right approach is here...

  
• sebastiaan

  May 2019

  @Daniel Yes, good point. I'll add this to the documentation.

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!