On Monday 29 July 2013 19:35:22 Thomas Lübking wrote:
> +
> + AddressbookJob *job = m_completionRequests.take(toEdit);
> +
> + if (job) {
> + disconnect(job, 0, 0, 0);
> + job->stop();
> + job->deleteLater();
> + }
>
> The completion race (you type faster than the backend responds
> and never get any response) is still open. What about having
> a job queue[2], only replace the recent job and move the
> recent job to oldjob once oldjob returns?
> This way really slow backends (ie. remote ones) would for sure
> send data from time to time, yet we don't spam them with long
> request chains.
> Now I'm using two m_completion(First/Second)Requests. And when new job is preparing it replace last started job. > > +void EnvelopeView::checkAddressesWork() > "checkAddressesKnown()"? > Changed. > Also, why not launch the jobs for all addresses in parallel? > (Ideally condense addresses before or in ::addToSet(), there > could be multiple entries) > For emails with lot of recipients this could lead to application crash... For each email address trojita is starting job. > + } > + > + checkAddressesWorkDone(); > "finishAdressesKnownCheck()" / "finishCheckAdressesKnown()"? - > "checkAddressesWorkDone()" sounds like a signal > Changed to finishCheckAdressesKnown. > + QTimer::singleShot(0, this, SLOT(checkAddressesWork())); > *cough* > What to use? > > > +void ImapPage::reloadPasswords() > this reloads only one password, doesn't? > Yes, changed. > > +void OutgoingPage::reloadPasswords() > singular as well? > Changed. -- Pali Rohár [email protected]
signature.asc
Description: This is a digitally signed message part.
