progwolff added a comment.

  In https://phabricator.kde.org/D7671#142826, @aacid wrote:
  
  > I don't know who told you this is the correct behaviour of kdirwatch, but i 
kind of disagree.
  >
  > Since we're not watching the file but *also* the folder.
  
  
  From the KDirWatch class reference:
  
  > void KDirWatch::dirty       (       const QString &         path    )       
signal
  >  Emitted when a watched object is changed.
  > 
  > For a directory this signal is emitted when files therein are created or 
deleted. [...]
  > 
  > When you watch a directory, changes in the size or attributes of contained 
files may or may not trigger this signal to be emitted depending on which 
backend is used by KDirWatch.
  
  So, let's assume we have a file `path/file` loaded in Okular and we watch the 
directory `path` for changes. We remove `file` and create a new file named 
`file`. The directory `path` will get dirty, as a new file is created.
  This means, a signal `dirty("path")` is emitted. However, Okular does not 
reload the document when the signal `dirty("path")` is received, but waits for 
a signal `dirty("path/file")`.
  
  The behaviour of KDirWatch might not be the most convenient, but from my 
perspective it is consistent with the documentation.
  Maybe the docs could be a little clearer at this point, like Henrik and 
Oliver proposed.
  
  ____________
  
  In https://phabricator.kde.org/D7671#142844, @rkflx wrote:
  
  > (Julian: Sorry to intrude, feel free to pick up from here.)
  
  
  Always happy when someone contributes good ideas to a discussion.

REPOSITORY
  R223 Okular

REVISION DETAIL
  https://phabricator.kde.org/D7671

To: progwolff, aacid
Cc: sander, rkflx, #okular, aacid

Reply via email to