https://bugs.kde.org/show_bug.cgi?id=452729
--- Comment #14 from Méven Car <meven....@kdemail.net> ---
(In reply to Harald Sitter from comment #12)
> (In reply to Méven Car from comment #11)
> > The bug might be there in fact.
> > Missing
> > if (cmd == CMD_NONE) {.
> > break;
> > }
> > ?
>
> That would resume the get() wouldn't it? What we actually want is for it to
> not resume.
In fact this is already handled by
if (!isSubCommand(cmd)) {
break;
}
That breaks for CMD_NONE
It is worth inspecting more the solution you found.
@@ -245,9 +245,7 @@ void KIO::MimeTypeFinderJobPrivate::scanFileWithGet()
m_suggestedFileName =
job->queryMetaData(QStringLiteral("content-disposition-filename"));
}
- if (!m_url.isLocalFile()) { // #434455
- job->putOnHold();
- }
+ job->kill();
q->emitResult();
});
}
Interestingly `void SchedulerPrivate::putSlaveOnHold(KIO::SimpleJob *job, const
QUrl &url)` which is called by `putOnHold` would kill the slave if it is called
twice for the same ioworker. So it seems to validate your patch, making this
more hastly.
Conceptually this makes prefect sense to stop the worker once we received the
mimetype.
--
You are receiving this mail because:
You are watching all bug changes.
