[Differential] [Commented On] D3287: Make sure we don't break compilation with past broken units

apol (Aleix Pol Gonzalez) noreply at phabricator.kde.org
Mon Nov 7 13:46:15 UTC 2016


apol added a comment.


  In https://phabricator.kde.org/D3287#61135, @dfaure wrote:
  
  > The error message is a bit confusing, it sounds like the right file is ${_tmp_FILE}.
  >  So I would suggest this instead
  >
  >   message(WARNING "${_tmp_FILE}: Broken \"File\" field, make sure it's pointing at the right file")
  
  
  Okay.
  
  > Also, shouldn't the regexp be File=(.*)\n ? Basically, anything in the File field that is not *.kcfg should lead to "Broken File Field", right?
  
  No, the thing is that we were matching `(.*kcfg).*` and this lets `File=something.kcfgc` work despite no something.kcfgc doesn't exist at all.
  
  > So in fact this is a bugfix (not "old broken behaviour") because it gives a better error message in case the file says File=foo : it will say "Broken File Field" instead of "Couldn't read the File field" (which would send people on the wrong track ... the field is here, it can be read, it just has the wrong contents).
  
  It's broken old behaviour because assuming people are going to put the correct filename with things in the end and we'll match the right thing is wong.

REVISION DETAIL
  https://phabricator.kde.org/D3287

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: apol, #frameworks, dfaure
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20161107/d357fc5d/attachment.html>


More information about the Kde-frameworks-devel mailing list