RC2 out

Hi

RC2 is out with:
        - fix squeakmap
        - fix toggle
        - pragmas fixes
        - duplicate halos

I hope to get soon some fix for compiledMethod class and we can go  
final.

Stef

On Oct 13, 2006, at 15:01, stephane ducasse wrote:

I just loaded RC2 and when I run SqueakMap I get an error.  It's in  
SMSqueakMap>>categories, instance var objects is undefined.

Anyone else this is happening to?

------------------------
Frank Caggiano
frankcag at crystal-objects dot com
http://www.crystal-objects.com

The best education for the best is the best education for all.
                     Robert Maynard Hutchins

Frank Caggiano a écrit :
No for me evrything is ok.

Thanks a lote Marcus and Steph :)

        Math

Stef,

I just grabbed 7063 and the current win32 VM, and opened the image.
Something is not right with the system window stripe coloring.  The
active state is somehow not following the keyboard focus.  Click in the
various workspaces and you should see what I mean.  If it is a corrupt
instance, I would start looking at the lower right workspace, but that
is simply a hunch.  One could argue that the lower right workspace
appears not to deactivate???

As a related aside, I still recommend flipping the #isActive tests (e.g.
appending #not) in #setStripeColorsFrom:.  Try it (see below) for a few
minutes and I suspect you will see what I mean - suddenly the active
window is "stronger" than the inactive ones.

Bill



setStripeColorsFrom: paneColor
        "Set the stripe color based on the given paneColor"

        labelArea ifNotNil: [labelArea color: Color transparent].
        self updateBoxesColor: (self isActive not
                                ifTrue: [paneColor]
                                ifFalse: [paneColor muchDarker]).
        stripes ifNil: [^self].
        self isActive not
                ifTrue: [self fillStyle: (self gradientWithColor:
paneColor lighter lighter lighter)]
                ifFalse: ["This could be much faster"
                                self fillStyle: (self gradientWithColor:
paneColor duller)].





Wilhelm K. Schwab, Ph.D.
University of Florida
Department of Anesthesiology
PO Box 100254
Gainesville, FL 32610-0254

Email: [hidden email]
Tel: (352) 846-1285
FAX: (352) 392-7029

On 14 oct. 06, at 17:13, Bill Schwab wrote:

I do not understand where it comes from. On mac I have the correct  
behavior
or I think I see it. ;)

> As a related aside, I still recommend flipping the #isActive tests  
> (e.g.
> appending #not) in #setStripeColorsFrom:.  Try it (see below) for a  
> few
> minutes and I suspect you will see what I mean - suddenly the active
> window is "stronger" than the inactive ones.

This is strange because I get no difference.
The topmost is the brightest.

Stef

Stef,

===============================
> I just grabbed 7063 and the current win32 VM, and opened the image.
> Something is not right with the system window stripe coloring. The
> active state is somehow not following the keyboard focus. Click in
> the
> various workspaces and you should see what I mean. If it is a corrupt
> instance, I would start looking at the lower right workspace, but that
> is simply a hunch. One could argue that the lower right workspace
> appears not to deactivate???

I do not understand where it comes from. On mac I have the correct
behavior
or I think I see it. ;)
===============================

With the downloaded image, just alternately click between the bottom two
workspaces.  I am seeing no change in their stripe coloring, suggesting
that something is not respecting and/or representing its active/inactive
status.


===============================
> As a related aside, I still recommend flipping the #isActive tests
> (e.g.
> appending #not) in #setStripeColorsFrom:. Try it (see below) for a
> few
> minutes and I suspect you will see what I mean - suddenly the active
> window is "stronger" than the inactive ones.

This is strange because I get no difference.
The topmost is the brightest.
===============================

Note that you will need to activate each system window in turn to see
the effects.  The inactive views will then have lighter stripes, but I
maintain that it is much more intuitive that the darker/stronger valued
stripe belongs to the active window.

Bill





Wilhelm K. Schwab, Ph.D.
University of Florida
Department of Anesthesiology
PO Box 100254
Gainesville, FL 32610-0254

Email: [hidden email]
Tel: (352) 846-1285
FAX: (352) 392-7029

Bill Schwab wrote:
> I just grabbed 7063 and the current win32 VM, and opened the image.
> Something is not right with the system window stripe coloring.  The
> active state is somehow not following the keyboard focus.  Click in the
> various workspaces and you should see what I mean.  If it is a corrupt
> instance, I would start looking at the lower right workspace, but that
> is simply a hunch.  One could argue that the lower right workspace
> appears not to deactivate???

The image was saved in a funny state. The lower right window wasn't
correctly deactivated. No idea how this happened but to fix it, click on
the window title instead of the text in it. Once you've done it,
everything should be fine.

On a related note: I noticed that activating the top-window takes about
a second or so which is due to the absence of the true-type caching fix
that I wrote about a year ago
(http://impara.de/pipermail/tweak/2005-August/000881.html). See also
http://bugs.impara.de/view.php?id=5233

Cheers,
   - Andreas

Hi andreas

where is the code? How can we put it in squeak if this is not visible  
for us?
I thought you were the maintainer of Graphics so that you would  
include it in there.

Stef

On 14 oct. 06, at 20:57, Andreas Raab wrote:

stephane ducasse wrote:
> where is the code? How can we put it in squeak if this is not visible
> for us?
> I thought you were the maintainer of Graphics so that you would include
> it in there.

Please do me a favor and at least have a peek at the reference I give. I
don't do it for my own good. This bug report explains the problem.

   http://bugs.impara.de/view.php?id=5233

TTCFont and friends are in Multilingual, not Graphics. In other words, I
can't touch them.

Cheers,
   - Andreas

On 15 oct. 06, at 10:08, Andreas Raab wrote:

> stephane ducasse wrote:
>> where is the code? How can we put it in squeak if this is not  
>> visible for us?
>> I thought you were the maintainer of Graphics so that you would  
>> include it in there.
>
> Please do me a favor and at least have a peek at the reference I give.

I look at it but it was not on mantis and I'm on the tweak mailing-list.

> I don't do it for my own good. This bug report explains the problem.
>
>   http://bugs.impara.de/view.php?id=5233
>
> TTCFont and friends are in Multilingual, not Graphics. In other  
> words, I can't touch them.

But if I follow your usual rant "us stupid harvesters do not  
understand what are package maintainers", we
could not move anything to the graphics package either because you  
are the maintainer of it.

So instead of ranting on our stupid practices, if you could admit  
that a bit of communication and synchronization
are necessary in the process and that with the merge mechanism of MC  
we can publish items in a package and that the
maintainer of the package can merge it back (and possibly reject the  
changes) because else we are stuck.

So now concretely how do we proceed for this changes?
Because you could have done it also, published two new versions of  
the packages and notify us.
I do not know who is maintaining the package multilangual. So we can  
do it if the code works because we are in ReleaseCandidate
status and this is taking far too much time so we could also postpone  
it to 3.10.

Stef

Hi Stef -

By now I *really* want you off the helm for any future Squeak release.
Your behavior of insulting *every last person* who finds any shortcoming
in 3.9 is just unworthy. I find a problem, I post a fix (without btw,
requesting any immediate action like including this in the current
release) I explain what the deeper problem is and request resolution by
filing a bug and you call me ranting?

I'm sorry, but in this kind of atmosphere I really have no interest in
contributing. Consider my fix withdrawn.

Cheers,
   - Andreas


stephane ducasse wrote:

hi andreas

Come on you systematically mentioned that we were not respecting  
package maintainer... Do you want that I really lose
time to go in your emails to show that to you? Of course this should  
be us that do not understand.


> Hi Stef -
>
> By now I *really* want you off the helm for any future Squeak  
> release. Your behavior of insulting *every last person* who finds  
> any shortcoming in 3.9 is just unworthy.

I'm not! I think that you over reacted. For once this is not me.
We are just in release candidate 2. We got in Gamma for at least some  
months. So we should stop.
I was just showing to you that your statement that we should not  
publish code in packages that has a maintainer leads to a
deadlock. Because you cannot touch multilnagual and we cannot touch  
graphics.

> I find a problem, I post a fix (without btw, requesting any  
> immediate action like including this in the current release) I  
> explain what the deeper problem is and request resolution by filing  
> a bug and you call me ranting?

Not in this email. I was referring to some previous ones but not this  
one. :)

>
> I'm sorry, but in this kind of atmosphere I really have no interest  
> in contributing. Consider my fix withdrawn.

Cool. This is the perfect attitude!
Quite constructive.

I just published a new cs and a new image

RC2-7064

        - I fixed problems with the repository (my mistake :()
        - included the SMLoader fix

Next: eval fix of scott
Stef

On Oct 15, 2006, at 5:40 AM, stephane ducasse wrote:

> I just published a new cs and a new image
>
> RC2-7064
>
> - I fixed problems with the repository (my mistake :()
> - included the SMLoader fix

Yep, I verified that the SMLoader fix is working in 7064.  (For some  
reason, the problem was still there in 7062, even though it looked  
like the fix was in the changesorter.)  Also, the windowing weirdness  
in 7062 seems to be fixed.

> Next: eval fix of scott

That's worth a serious look for 3.9, even at this late time.  Also,  
I'd recommend added the window title font fix that I just posted,  
since it's just a cosmetic tweak. :-)

- Doug

I saw and I will apply in the last batch (I hope) of fixes and  
release the steam.

Stef


On 15 oct. 06, at 19:10, Doug Way wrote: