Hi,

to reproduce the error you can do:

1. Create a new Pipeline job
2. In the job configuration, choose "Pipeline Script from SCM"
3. Enter a existing git repository URL (and credentials if required) containing a Jenkinsfile
4. Select "trigger from SCM and enter a time"
5. Wait until job gets triggered  (I let the job fail on purpose by providing a incorrect Jenkinsfile, but for this issue I think this is irrelevant)
6. After job got triggered go back to configuration page and change the repository URL (containing a different Jenkinsfile)
7. Change the trigger and wait until job get triggered
8. Now have a look at the query protocol. My protocol indicates that jenkins is still using the first repository:  

    > git.exe ls-remote -h git@oldRepository.git # timeout=10
   Found 1 remote heads on git@oldRepository.git

   The same seems to account for the other values entered through the web interface (Name and RefSpec).