D24962: [DeleteJob] Use a separate worker thread to run actual IO operation

[Méven Car]

meven added a comment.


  In D24962#554961 <https://phabricator.kde.org/D24962#554961>, @dfaure wrote:
  
  > Any reason why you didn't implement my suggestion of
  >
  >    DeleteJobIOWorker *ioworker() {
  >        if (!m_ioworker) {
  >              ...
  >        }
  >        return m_ioworker;
  >    }
  >   [...] 
  >   QMetaObject::invokeMethod(ioworker(), "rmfile", [...]);
  >   
  >
  > ?
  >  A call to an initSomething() method can easily be forgotten, while an on-demand getter ensure that the worker is created when it's needed (for the first time).
  >  Sorry for the nitpicking :-)
  
  
  Since m_ioworker is accessible where worker() would be needed, nothing keeps the user to use m_ioworker instead of worker() which is in the end is equivalent to forget to call initSomething.
  It was my own habit to use a init or ensureInit function in such cases, and is the main reason I was using one.
  But it is more explicit to have an accessor and for code coherence I have changed the code to have a *worker() function.

INLINE COMMENTS

> dfaure wrote in deletejob.cpp:412
> marked as done but I still see removeFirst, I'm confused.

I missed this line, I did it line 430.

REPOSITORY
  R241 KIO

REVISION DETAIL
  https://phabricator.kde.org/D24962

To: meven, dfaure
Cc: kde-frameworks-devel, LeGast00n, GB_2, michaelh, ngraham, bruns
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20191028/f903bdd3/attachment-0001.html>