Recovering from a bad crash...

Rich Warren rwmlist at gmail.com
Mon Aug 28 09:16:01 UTC 2006


I looked at monticello, as I understand it, I don't think it will be  
a big help (at least not that I'm now doing nightly backups of the  
image to my .mac account). My current setup is automated, and  
monticello would force me to manually commit changes, so it would  
actually be a step backwards.

But, I may have missed something important.

On another note, is there any way to control the amount of memory the  
VM is using? Particularly on the Mac?

Thanks,

-Rich-

On Aug 27, 2006, at 10:07 PM, stéphane ducasse wrote:

> rich
>
> take 5 min to understand how to use monticello.
> This is the best advice I can give you.
> Have a look at my videos
> 	http://www.univ-savoie.fr/Portail/Groupes/LISTIC/membres/ 
> Stephane.Ducasse/Resources.html
>  and the chapter explaining MC
> 	http://www.iam.unibe.ch/~ducasse/Teaching/CoursAnnecy/0506-M1-OOP/ 
> OOPAvance.pdf (in english)
>
> Stef
>
>
> On 28 août 06, at 03:25, Rich Warren wrote:
>
>> I had to get rid of the image, but I was able to load the changes  
>> from the old image to the new image, so I didn't lose any code.  
>> (I'm not sure if I did it the best way, but it seems to have  
>> worked--was a bit of a pain, however). I'm also backing up my  
>> image nightly now to a remote server--so I hope there are no  
>> future problems.
>>
>> I was modifying the Balloon3D tutorial to make it print function  
>> classes (I will be using 30-dimension functions to test a ML  
>> implementation--but I wanted to print out the 2D versions to  
>> visually inspect the surfaces).
>>
>> It may have been a memory issue. Previously I had tried to create  
>> a 500 x 500 mesh, and bad things started to happen. That locked up  
>> the UI and command-period didn't save me. I force quit squeak and  
>> restarted, and everything seemed fine.
>>
>> I was then debugging the function class code. I don't remember  
>> exactly what I was doing, but I was in the debugger at the time. I  
>> started getting error messages, something about an "error when  
>> unwinding?" Anyway, the message kept popping up, and nothing I did  
>> would get rid of them. Then everything went black.
>>
>> Anyway, thanks to everyone for the advice and support!
>>
>> -Rich-
>>
>>
>> On Aug 26, 2006, at 3:41 AM, stéphane ducasse wrote:
>>
>>> Basically it is really rare that you lose your code in smalltalk  
>>> (may if you disc is totally full).
>>> never throw away an image and its change file after a crash. You can
>>> do a recover from last changes or I never remeber the item name  
>>> in the changes menu.
>>> The changes is a log of all the actions you performed.
>>>
>>> In addtion I suggest you to sue monticello so that you publish  
>>> you code on a remote server (and you house can burn
>>> and the code is still saved) and also automatically backed up on  
>>> your local disc.
>>>
>>> So with changes + MC
>>> normally you should never lose code
>>>
>>> Stef
>>>
>>> Could you let us know what exactly crahsed your system because  
>>> normally squeak is quite robust.
>>>
>>>
>>> On 26 août 06, at 10:36, Rich Warren wrote:
>>>
>>>> I was trying to run test cases for a project I'm working on,  
>>>> when I started getting a lot of errors (Unwind errors, or  
>>>> something like that). I eventually got them to stop, but a few  
>>>> seconds later the entire squeak environment went black. I had to  
>>>> force quit (working on a mac here), and when I tried to restart  
>>>> squeak, it appears to load, but nothing shows up.
>>>>
>>>> Is there any way to recover from something like this, or will I  
>>>> need to go back to a fresh image? I have a fair bit of code  
>>>> already written, and I'd rather not lose it.
>>>>
>>>> I've had intermittent trouble with squeak locking up (usually as  
>>>> a result of something I've done--but sometimes command-. doesn't  
>>>> fix things and I have to force quit). Are there any steps I can  
>>>> take to make squeak more stable?
>>>>
>>>> Thanks,
>>>>
>>>> -Rich-
>>>>
>>>
>>>
>>
>>
>>
>
>




More information about the Squeak-dev mailing list