Latest VM from Zeroconf gives: 'NBCoInterpreter NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Feb 25 2014 NBCogit NativeBoost-CogPlugin-GuillermoPolito.19 uuid: acc98e51-2fba-4841-a965-2975997bba66 Feb 25 2014 https://github.com/pharo-project/pharo-vm.git Commit: 66f4cf9345ba43bc232646d59b5e9369c369caba Date: 2014-02-25 11:45:12 +0100 By: Esteban Lorenzano <estebanlm@gmail.com> Jenkins build #14806 ' Yours appear to be much older. I have a spare XP laptop on my other desk, I'll give it a shot. Phil On Wed, Feb 26, 2014 at 4:51 PM, <btc@openinworld.com> wrote:
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). (432.0@3.0). (30@ -2.0)}) WindowEdgeGripMorph(EdgeGripMorph)>>mouseMove: <-- WindowEdgeGripMorph>>mouseMove: ({(641@663). (432.0@3.0). (30@ -2.0)}) WindowEdgeGripMorph(EdgeGripMorph)>>mouseMove: <-- WindowEdgeGripMorph>>mouseMove: ({(641@656). (432.0@3.0). (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 <philippeback@gmail.com><philippeback@gmail.com>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 <sven@stfx.eu> wrote:
On 26 Feb 2014, at 01:57, Robert Shiplett <grshiplett@gmail.com> 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
------------------------------
