4.1 release candidate 3

> Damn, I thought this had gotten fixed.  This has been an annoying bug
> for us for a very long time; it happens all the time in 3.9, but I had

I saw this for the first time at all and even not on the first update
run. And never read about any complains, that this "happens all the
time in 3.9". Is there some  bug report or e-mail thread about this? I
wasn't always following the mailing list in the past.

I feel it is tricky to update the SystemProgressMorph using MC while
it is actually in use during updating itself. I got bitten by this a
few times and then was reluctant to do any more changes because of the
difficulties that could arise.
The last cosmetic change was to remove the borderlines of the bars.
Now this just takes an effect after the SystemProgressMorph gets
reset again. Such code was missing at first in the postscript
and I added it later on.
Now I think the empty progress morph just stems from that attempt to
initialize and reset it during a postscript while it is actually in use.
So my guess is, that this will only be seen once (or at max every time
SystemProgressMorph will be reset while in use).

> not seen in in 4.1, so I assumed someone had fixed it, but I guess it
> didn't get fixed afterall!  :(

>From my view there is and was nothing to be fixed.

Alex

On 4/13/2010 11:41 PM, Alexander Lazarević wrote:
> I feel it is tricky to update the SystemProgressMorph using MC while
> it is actually in use during updating itself. I got bitten by this a
> few times and then was reluctant to do any more changes because of the
> difficulties that could arise.

It's not tricky at all. The next time you need to reset it from MC,
simply use "SystemProgressMorph future reset". This will work nicely
(try it with the snippet I sent earlier).

Cheers,
   - Andreas

2010/4/14 Andreas Raab <[hidden email]>:
> On 4/13/2010 11:41 PM, Alexander Lazarević wrote:
> It's not tricky at all. The next time you need to reset it from MC, simply
> use "SystemProgressMorph future reset". This will work nicely (try it with
> the snippet I sent earlier).

Thanks Andreas. What I still don't grok is, when the message send will
happen. Ok, some time in the future, but who says it won't be during
the SPM is still in use? I'm afraid Future/Promises/Traits (and a lot
more) are still white spots on my personal Squeak map!

Updating a SPM still feels tricky for me because I ran into some MNU
(as I recall) when I updated a method that contained a new message
send to a new method that was about to be loaded after the updated one
(there is no defined order how methods get loaded, is there?). Because
the updated method was used before the new method was loaded I got a
MNU.
So I had to split this up into two commits and make a mcm for the first one.

Alex