[Cuis-dev] White VM window

Hilaire Fernandes hfern at free.fr
Wed Aug 17 07:49:39 PDT 2022


Hi Juan,

At first I got a DNU when executing this snippet, but now ok. Whatever.

In a freshly updated Cuis image, installed with DrGeo code, in developer 
set, I can execute this code now:

*Smalltalk at: #LogStartupDebugAids put: true*

When I save the image without quitting, the mouse cursor turns to the 
writing cursor, image stalled, no output on the console from where was 
started the VM

I also test with my bundle build procedure where I also turn on this log 
startup flag. At the end the image is automatically saved and quit. 
There is no issue.

When the bundle is  started, DrGeo hangs, console reports:

2022-08-17T16:35:16.171 process:79 926479 To #openSourceFiles
2022-08-17T16:35:16.174 process:79 926479 To #initClassCachedState
2022-08-17T16:35:16.184 process:79 926479 To #doStartUp:

When I run the image with the VM -tracestores option, I have more details:

2022-08-17T16:44:09.937 process:79 926479 To #openSourceFiles
2022-08-17T16:44:09.940 process:79 926479 To #initClassCachedState
SmallFloat64>+
SmallFloat64>+
2022-08-17T16:44:09.950 process:79 926479 To #doStartUp:
Delay class(Behavior)>startUp:
DisplayScreen class(Behavior)>startUp:
Cursor class(Behavior)>startUp:
InputSensor class(Behavior)>startUp:
ProcessorScheduler class(Behavior)>startUp:
WeakArray class(Behavior)>startUp:
Locale class>startUp:
NaturalLanguageTranslator class>startUp:
DrGeoSystem class(Behavior)>startUp:
SmallFloat64>/
SmallFloat64>>=
SmallFloat64>-
Point>+
Point>+
SmallFloat64>=

All in all, -tracestores gives complementary information to your 
changeset. But it is still too little information.

I have an issue in my startup sequence, I think there is a weakness 
somewhere in the font handling. It is possible my startup sequence 
execute unproperly with a wrong path name to font or/and a wrong font 
name. I have observed Cuis does not like that kind of things.

Oh, if I don't kill quickly the stalled image, it also eat up all the 24 
GB of my workstation, forcing me to hard reset it.

Thanks

Hilaire

Le 17/08/2022 à 15:26, Juan Vuletich a écrit :
> Yes. Exactly that. (Please note how I describe in detail what I'm 
> doing, this is what I ask you everybody to do, so I have information 
> to work with) After pulling the repo, I start Cuis as usual, install 
> updates, evaluate the snippet you say, save and quit the image. When I 
> restart it, I get this:
>
> 2022-08-17T10:18:45.979 process:79 2121406 To #openSourceFiles
> 2022-08-17T10:18:45.990 process:79 2121406 To #initClassCachedState
> 2022-08-17T10:18:45.999 process:79 2121406 To #doStartUp:
> 2022-08-17T10:18:46.004 process:79 2121406 To 
> #spawnNewMorphicProcessFor:  [world]
> 2022-08-17T10:18:46.007 process:79 2121406 To 
> #restoreLostChangesIfNecessary
> 2022-08-17T10:18:46.073 process:40 1661169 To #restoreDisplay
> 2022-08-17T10:18:46.077 process:40 1661169 To #processCommandLineArguments
> 2022-08-17T10:18:46.080 process:40 1661169 To #launchApp
>
> This shows in the terminal from where I started Cuis (this is stdout), 
> it also appears in the Transcript once the UI is working, and it also 
> appears on file 'transcript.txt' in the current directory of the image 
> (from where I started it).
>
> What do you get? Does the image start at all? If you do, for example, 
> senders of #*LogStartupDebugAids,* and follow from there, do you see 
> stuff that makes sense?
>
> Thanks,

-- 
GNU Dr. Geo
http://drgeo.eu
http://blog.drgeo.eu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20220817/73059619/attachment.htm>


More information about the Cuis-dev mailing list