Which one should I download?

Pharo 3.0 on Windows 8.1

Installing and running Pharo 3.0 on a blank Windows 8.1 machine in under 1 minute.

http://www.youtube.com/watch?v=FKokxrQtrg0

With an extra minute to demo some basic interactions.

On 26 Feb 2014, at 14:07, Pharo4Stef <[hidden email]> wrote:

> robert
>
> Are you sure that you really use the latest version?
> Because I do not see why opening a window would be platform specific and I could not reproduce this problem.
> So do you have a real way to reproduce the problem? can you send us the debuglog?
> And BTW Pharo is our and your system and we work all to make it better so no need to be pushy it does not help.
>
> Stef

Robert Shiplett wrote:

I did a fresh 3.0 install when I got the latest walkback ... I had not even dropped in the files from latest.zip yet.

I have insomnia, so please bear with me ... I should be asleep at this hour ...

debug log attached 

BTW  - Why are we not naming these log files for the image that was running when the log was written ?

Well at least in the first few lines it says "Jenkins build #14776 Image: Pharo3.0 [Latest update: #30782]"
cheers -ben

On 26 February 2014 04:36, Sven Van Caekenberghe <[hidden email]> wrote:

On 26 Feb 2014, at 01:57, Robert Shiplett <[hidden email]> wrote:

> I am running from that latest.zip tonight and PackageTreeNautilusUI >>update is getting a walkback messageNoUnderstood for
> receiver of >>protocolsFor: is nil
>
> I simply flipped to class view in Announcements and then flipped that toggle back.
>
> Does this need to be reported, or is Nautilus still known to be unstable ?

I can't reproduce this either, please be more specific.

BTW, Nautilus is stable and used daily by the whole community. Of course, it is a complex piece of code with lots of UI and system interactions, so issues are always possible.

Sven

Hi Robert,

I am on Windows 7 SP1 so am probably in a good position to try to replicate.  Trying to align with your screen snapshot this is what I've done:
1. Downloaded and launched fresh image build 30782 with PharoLauncher.
2. Clicked World > System Browser
3. Clicked on Announcements-View, AnnouncementSpy, [class-side] checkbox, instance creation, openOn:, [class-side] checkbox
and all is fine. 
Can you detail where your test procedure differs from that?
-----

Do you have any startup scripts under C:\Users\$yourAccount\AppData\Roaming\pharo
or have you saved any preferences. (btw, if there a command line option to avoid processing startup and preference files, it would be useful in this situation to remove doubt?)

I can't think where a difference may lie except the operating system version.  Can you know someone with Windows 7 or 8 your can test Pharo from a USB stick.
-----

If I put (self haltOnce) at the top of EdgeGripMorph>>mouseMove: a large amount of clicking around the <Class-side> cannot trigger the breakpoint.  This only triggers if I drag edge of the Browser window to change its size.  So is there some interaction that occurs with that in your testing?
-----

Now it seems the problem is that _lastMouse_ instance is nil. Investigating shows this has only three relevant references (from the debugger, right-click on the variable in the bottom left pane and choose <List Methods Using "lastMouse". These are:
* AbstractResizerMorph>>mouseDown:
* EdgeGripMorph>>mouseMove:
* EdgeGripMorph>>mouseDown: - which overrides AbstractResizerMorph, so for our purposes there are only two relevant references to _lastMouse_
#mouseDown is the only method that stores into _lastMouse_ which is guaranteed to be non-nil since the collection is hardcoded. 
So it appears that #mouseDown should be called before #mouseMove and it is not.

Try putting the following code at the top of #mouseDown and #mouseMove
    Transcript crShow: thisContext printString , ' <-- ' , thisContext sender printString , '  (',  lastMouse printString , ')'.

Then doing the following:
1. Clicked World > System Browser
2. Grabbed bottom edge and dragged up to make window smaller.
I get...
    WindowEdgeGripMorph(EdgeGripMorph)>>mouseDown: <-- WindowEdgeGripMorph>>mouseDown:  (nil)
    WindowEdgeGripMorph(EdgeGripMorph)>>mouseMove: <-- WindowEdgeGripMorph>>mouseMove:  ({(641@664). ([hidden email]). (30@ -2.0)})
    WindowEdgeGripMorph(EdgeGripMorph)>>mouseMove: <-- WindowEdgeGripMorph>>mouseMove:  ({(641@663). ([hidden email]). (30@ -2.0)})
    WindowEdgeGripMorph(EdgeGripMorph)>>mouseMove: <-- WindowEdgeGripMorph>>mouseMove:  ({(641@656). ([hidden email]). (30@ -2.0)})

-----
btw, my VM is same as yours...
NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Dec  7 2013
NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Dec  7 2013
https://github.com/pharo-project/pharo-vm.git Commit: 71adca113a9aff2876288927e5c1c86bf7ac13e2
Date: 2013-12-07 05:00:47 -0800 By: Philippe Back [hidden email] Jenkins build #14776

cheers -ben
P.S. I've not encountered a problem like that with Nautilus in the past 3 months I've been using it - but different environments can dig out corner cases. Good luck.

Robert Shiplett wrote:

hit another walkback first

screenshot attached

this must be Windows XP 3.1 issue ...

On 26 February 2014 04:36, Sven Van Caekenberghe <[hidden email]> wrote:

On 26 Feb 2014, at 01:57, Robert Shiplett <[hidden email]> wrote:

> I am running from that latest.zip tonight and PackageTreeNautilusUI >>update is getting a walkback messageNoUnderstood for
> receiver of >>protocolsFor: is nil
>
> I simply flipped to class view in Announcements and then flipped that toggle back.
>
> Does this need to be reported, or is Nautilus still known to be unstable ?

I can't reproduce this either, please be more specific.

BTW, Nautilus is stable and used daily by the whole community. Of course, it is a complex piece of code with lots of UI and system interactions, so issues are always possible.

Sven

---

Robert Shiplett wrote:

"BTW, Nautilus is stable and used daily by the whole community."

This still rankles ... I have been a Smalltalker since long before Squeak and have been using Pharo since Seaside first ran on it - if not earlier than that ...

I also received from another on the mail list

"and in fact I cannot reproduce it."

"And in fact" is such a wicked rhetorical move ... in a game I have lost all interest in playing now that I am retired.

It may take me a lot of effort to track these 2 bugs down as far as "Follow these steps to reproduce" .. but then I am not a CONSORTIUM member ( I used to pay big bucks for Smalltalk, back in the day ) so may be I am too far outside the "whole community".... btw.

I will spare you my plot's view of the term "stable" and the notion of stability ... and will be sure I rename key files as I go along ... [ why are we still having to rename files such as these when able to run multiple images in the same dir ? ]

ciao

btw, have you tried PharoLauncher?
It is a very cool way of managing multiple images.
https://ci.inria.fr/pharo-contribution/job/PharoLauncher-Win-Package/lastSuccessfulBuild/artifact/pharo-ci/pharo_installer.exe
cheers -ben

On 26 February 2014 04:36, Sven Van Caekenberghe <[hidden email]> wrote:

On 26 Feb 2014, at 01:57, Robert Shiplett <[hidden email]> wrote:

> I am running from that latest.zip tonight and PackageTreeNautilusUI >>update is getting a walkback messageNoUnderstood for
> receiver of >>protocolsFor: is nil
>
> I simply flipped to class view in Announcements and then flipped that toggle back.
>
> Does this need to be reported, or is Nautilus still known to be unstable ?

I can't reproduce this either, please be more specific.

BTW, Nautilus is stable and used daily by the whole community. Of course, it is a complex piece of code with lots of UI and system interactions, so issues are always possible.

Sven

Hi Robert,

Thanks for your report.  A few general things that you are probably already familiar with - comments inline...


Robert Shiplett wrote:

regardless ...

I did a sanity check on the contents of yesterday's latest.zip and hit a walkback

I then did a sanity check on the latest 3.0 installer and hit a different walkback.

At least on Win XP 3.1 that does not feel like Beta but like alpha, in my experience.  Unstable.

I'm glad you limited that comment for XP.  My understanding is that the Beta phase began when features stopped being added and all 11000 daily run automated tests were green.  I've been using Pharo 3 on Windows 7 for two months and found it very stable.  However different environments can dig out corner cases, and anyway your UI problem is one that perhaps would not be found by the current automated testing.  I think none of the main developers use XP, so its understandable that something XP related might slip through to Beta phase.  Even though there might be a lot of XP installations out there, with last sale date circa 2008, end of mainstream support 2009 and extended support ending this April, I can appreciate that with limited resources XP is not a prime supported platform.  However hopefully now you've identified an issue, it can be resolved fairly quickly.

Now I notice in EdgeGripMorph>>mouseMove: that there is the line...
    target ifNil: [^self].

A simple fix might be adding a line...
    lastMouse ifNil: [ ^self ].
Can you observe any negative consequences of that in operation?

Sanity check for a Smalltalk image

  open a browser 

  open a workspace

  flip about in the browser  [across packages, classes, instance view, class view, implementors and callers

Even though you are able to trigger this problem quite easily, the difficulty is that "flip about in the browser" is not a reproducible case.  You might be consistently performing some small unrecognized action that we happen to avoid.  These "unreproducible" cases (in the strictest sense of having a minimal step-wise set of actions guaranteed to reproduce the problem, rather than just frequency of occurrence) are difficult issues to troubleshoot remotely.  We will be relying on your intuition to help get this so it can be reproduced by other.  Until that time there is not much developers can do.

  DoIt from workspace [never got here either time ; I would do smthg like a ProfStef or call up a browser ; no set routine as that way lies madness ]

"No set routine" is a good way rattle-the-cage to identify new bugs, but once found , narrowing it a reproducible case is immensely useful.

AND while doing the above, flip between windows, max /min [ I have been doing this for years ... it takes only a couple of minutes - it is not a step-wise protocol .... that would result in missing things that you wouldn't expect - and it is in NO way exhaustive ... it is just a sanity check ( Do you know what year it is? Who is the President of USA ?  What is today's date ? How did you get here ? Do you know why you were [arrested/admitted/removed from the aircraft/booted outa the bar ... ) ]

Hunch ONLY :

Likely culprit ... each time I also opened for future use (and to have some frames open

   Metacello browser

   the 2 help items off the main popup menu (one general; one Tutorials - a new user is likely to do so )

Hunches have a lot going for them, but if its hard for you to make it "reproducible" then its double-hard for other to reproduce it so they can address it.
cheers -ben

I suspect one of the latter, based on Last Class Changed for that image ... but that will be my hunch for now ...

ciao

r

  

   

On 26 February 2014 07:18, Esteban Lorenzano <[hidden email]> wrote:

On 26 Feb 2014, at 11:30, Robert Shiplett <[hidden email]> wrote:

"BTW, Nautilus is stable and used daily by the whole community."

This still rankles ... I have been a Smalltalker since long before Squeak and have been using Pharo since Seaside first ran on it - if not earlier than that ...

I also received from another on the mail list

"and in fact I cannot reproduce it."

"And in fact" is such a wicked rhetorical move ... in a game I have lost all interest in playing now that I am retired.

it is not. 

it was just saying that I tried your problem and I could not reproduce it. It was not to been rethorical or anything, just pointing a fact. 

I’m interested (and is also my job) on having the best and stable version of Pharo possible. 

I’m so interested that I was at 10min from leaving a my house to take an 10h fly and I sit to try to reproduce your problem. 

So, please… do not misinterpret the contents of the mail. It was an attempt to have an understanding of the problem, in an effort to help.

An effort, btw, that sometimes is not easy, because we do not have all the elements to make the diagnostic… 

It may take me a lot of effort to track these 2 bugs down as far as "Follow these steps to reproduce" .. but then I am not a CONSORTIUM member ( I used to pay big bucks for Smalltalk, back in the day ) so may be I am too far outside the "whole community".... btw.

The community is here. In this list. Around the list. 

Consortium is for companies willing to contribute with some money support (and so is the association for individuals), because things cost effort and effort implies time and time implies money. 

We always try to solve the problems of everybody (we, the community, including yourself), because that’s the best for all… so do not put yourself outside in a way we never say/act/even think about. 

We want to help you. 

I will spare you my plot's view of the term "stable" and the notion of stability ... and will be sure I rename key files as I go along ... [ why are we still having to rename files such as these when able to run multiple images in the same dir ? ]

Your pov about stability can be ok. What is not ok is such amount of negativity over a simple fact: we has not been able to reproduce your problem (but you avoid the other obvious fact: we tried to reproduce it. So we did not ignored you).

Again, I try to help, as everybody in the community… but such a negative over-reaction is probably just compared to my own overeacted answer :)

Esteban 

  

ciao

On 26 February 2014 04:36, Sven Van Caekenberghe <[hidden email]> wrote:

On 26 Feb 2014, at 01:57, Robert Shiplett <[hidden email]> wrote:

> I am running from that latest.zip tonight and PackageTreeNautilusUI >>update is getting a walkback messageNoUnderstood for
> receiver of >>protocolsFor: is nil
>
> I simply flipped to class view in Announcements and then flipped that toggle back.
>
> Does this need to be reported, or is Nautilus still known to be unstable ?

I can't reproduce this either, please be more specific.

BTW, Nautilus is stable and used daily by the whole community. Of course, it is a complex piece of code with lots of UI and system interactions, so issues are always possible.

Sven

Robert wrote
>I'm dragging and dropping in a fresh install

No, you are not.
You are dropping a new image on an old vm.
Replace the vm and retry.

Stephan

Robert wrote:
>I am unzipping from our zip on pharo-project.org page for 3.0 Beta
>
>If that is wrong, let's have someone fix that zip, that page, what-ever

Thank you, sounds like you found the problem. I support your demand.

Get a newer vm from
  http://files.pharo.org/vm/pharo/win/
until the problem is fixed.

Stephan

On 26 Feb 2014, at 21:16, Robert Shiplett <[hidden email]> wrote:

> Btw, desktop menu move top window to back seems to have no effect ( unless fired from a given window and not the Pharo world area )

https://pharo.fogbugz.com/f/cases/13001/SystemWindow-sendTopWindowToBack-does-not-work-when-invoked-from-World-menu-or-global-shortcut