Skip to main content

nvUltra (2022.12.23.14.53) hangs on indexing

Completed

Comments

8 comments

  • Fletcher Penney

    If you identify the problematic file, and can share it, I can see what is happening.  The fastest way to find the file is to recursively split the folder in half and see which half fails (or doesn't fail) when you clear the cached index and try to index again.

    0
  • Fletcher Penney

    The next beta will have a troubleshooting option to log each file that is indexed that may help identify the problematic file as well.

    0
  • kallovsky

    I found the problematic file - it was an old "Notes and Settings.txt" file (from nvalt?) containing plist and binary data. The prior version didn't have any problems with that file. Please let me know if I should send you the file (I can't open it anymore and can't tell if it contains private data…).

    And yes, it would be nice to have an option to log each filename when indexing :-)

     

     

    0
  • Fletcher Penney

    If you send me the file attached to a support request I can see where the issue is and should be able to fix it.  All I can do is assure you that I won't look into the contents any further than necessary to find the bug, which involves me running the indexer within Xcode and trying to find infinite loops.  I often don't need to look at anything beyond a single character that is causing the issue.  But I understand if you are concerned about private information.

    0
  • kallovsky

    Just sent you the file!

    Thanks!

    0
  • Fletcher Penney

    Got it. Digging in now.

     

    Thanks!

    0
  • Fletcher Penney

    The issue is that when the indexer hits a byte that is not a valid UTF-8 character, it goes into a loop.

     

    1. The loop is fixed.

    2. It does raise the question as to whether a file that is determined to have binary data should be indexed or not.  This would only happen when the file is labelled as being text (here the `.txt` extension), but turns out to have non-text information inside it.  Files that are not labelled as text are not indexed (with a couple of specific exceptions).  My inclination is to go ahead and allow the indexing to occur.

     

     

    Thanks for sending (and all I had to look at was a single byte of the file....)  :) 

     

    0
  • kallovsky

    Thanks for fixing!

    I don't really have an application where I need those mixed (text and non-text) files to be indexed, the file above just happened to be in that folder. However, indexing them if they have a txt-extension makes sense from my point of view.

    0

Please sign in to leave a comment.

Powered by Zendesk