Re: [evolution-patches] seeking patch review for bug #41846

[Not Zed <notzed ximian com>]

Well i'm sorry, you didn't test it fully then.

It wont handle the user-cancellation stuff properly (Jeff finally just
put that in cvs for smtp sending).

If you cancel sending, it will go to the next message and re-try, rather
than just aborting the whole process.  Only the code has already been
cancelled so each subsequent one will fail too.

But you want to abort immediately then clear the cancellation, and then
try to sync/close the folder.  Otherwise you could end up sending the
message again when you don't need to (if it crashes).


On Mon, 2004-03-15 at 16:03 +0800, Calvin Liu wrote:

> Not Zed, Jeff,
> 
> Just FYI.
> 
> I tested it on my linux box. Seems it works for me, no outstanding
> side-effect.
> 
> Calvin
> 
> On Fri, 2004-03-12 at 11:14, Not Zed wrote:
> > I dunno jeff, it seems to be on the right track - at least for a
> > solution which will work for most of the time.  So long as it checks for
> > the user  cancelled exception, which is the only really important (and
> > configured) one, it could just go onto the next message otherwise.
> > 
> > It probably also needs to do the same as the pop fetching code, if you
> > get a user cancel, clear the cancellation and still do a folder sync (so
> > it doesn't try to re-send messages).
> > 
> > On Thu, 2004-03-11 at 11:26 -0500, Jeffrey Stedfast wrote:
> > 
> > > This isn't the right fix. What if the exception was the the user
> > > cancelled? or that the server was unreachable? or...?
> > > 
> > > I don't think you should bother trying to fix this bug right now. The
> > > proper fix will require a fair amount of work I believe (making sure
> > > all the underlying code sets the proper exceptions, keeping track of
> > > which exceptions we've gotten, etc).
> > > 
> > > I suggest you move on to another bug and leave this one up to me or
> > > Zucchi.
> > > 
> > > Jeff
> > > 
> > > On Thu, 2004-03-11 at 18:45 +0800, Calvin Liu wrote:
> > > 
> > > 
> > > > Hi, there,
> > > > 
> > > > Here's a patch for bug #41846/55290.
> > > > This bugs shows the outbox queue could be blocked when error occurs.
> > > > 
> > > > Patch is simple, just change "break" to "continue" in a for loop.
> > > > Seems work for me.
> > > > 
> > > > Please review it. I'm wondering if there's any side-effect.
> > > > 
> > > > Thanks.
> > > > Calvin
> >