Frozen Image

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Frozen Image

Zulq Alam
Hi,

Maybe you can help me, I have an image that is stuck in some form of
infinite loop.

I can interrupt the loop using Alt+. and a debugger appears but the loop
seems to resume again.

Any ideas what's going on?

Or failing that, what's the best way of recovering code from a .changes
file or an external .image file?

Thanks,
Zulq.


Reply | Threaded
Open this post in threaded view
|

Re: Frozen Image

Benjamin Schroeder-2

On Apr 14, 2006, at 6:31 AM, Zulq Alam wrote:

> Hi,
>
> Maybe you can help me, I have an image that is stuck in some form  
> of infinite loop.
>
> I can interrupt the loop using Alt+. and a debugger appears but the  
> loop seems to resume again.
>
> Any ideas what's going on?
>
> Or failing that, what's the best way of recovering code from  
> a .changes file or an external .image file?

Hi Zulq,

I'm not sure from your message if this has happened with a running  
image, or if it happens every time you start up your image. Here's  
some advice in case it's the former, and your image will still start  
up OK.

Go to the World menu (click on the background) and choose  
"changes...". This will bring up a menu with options for browsing the  
changes file. The "recently logged changes" option will let you look  
at all the changes you had made since the last save, and optionally  
load selected ones into your image.

Note that there are "do its" mixed in with the method changes. Some  
of these are from code you executed in a workspace or inspector; some  
of them are from changing class definitions or removing methods. You  
can select which ones of these to file in as well (obviously, the  
class-change and method-removal ones are usually important).

There may be a way to keep from having to shut down your image - when  
I have had that situation, sometimes I have been lucky enough to get  
the thing interacting again by sending multiple interrupts, or just  
waiting out a long series of debugger windows that appeared. Of  
course I've also had times when I had to force-quit.

Hope this helps - good luck.

Benjamin Schroeder


Reply | Threaded
Open this post in threaded view
|

Re: Frozen Image

Zulq Alam
 > I'm not sure from your message if this has happened with a running
image, or if it happens every time you start up your image.

It happens every time I start the image. Your advice is good in the
later case as well though.

I started a new image, used a file list to find the original .changes
file and used this information to recover the important changes using a
changelist browser. This is the same browser that comes up when looking
at recently logged changes just looking at an external .changes file.

 > Note that there are "do its" mixed in with the method changes.

Yes, this was a pain as I couldn't just select the code modifications
and finding the pertinent "do its" was too time consuming. Instead I
just skipped them all and redid as much of the refactoring that was
undone. I'm glad I didn't have more work to recover.

 > Of course I've also had times when I had to force-quit.

I have always been able to interrupt and save before.

Thanks,
Zulq.

Reply | Threaded
Open this post in threaded view
|

Re: Frozen Image

stéphane ducasse-2

On 14 avr. 06, at 14:55, Zulq Alam wrote:

> > I'm not sure from your message if this has happened with a  
> running image, or if it happens every time you start up your image.
>
> It happens every time I start the image. Your advice is good in the  
> later case as well though.
>
> I started a new image, used a file list to find the  
> original .changes file and used this information to recover the  
> important changes using a changelist browser. This is the same  
> browser that comes up when looking at recently logged changes just  
> looking at an external .changes file.
>
> > Note that there are "do its" mixed in with the method changes.
>
> Yes, this was a pain as I couldn't just select the code  
> modifications and finding the pertinent "do its" was too time  
> consuming. Instead I just skipped them all and redid as much of the  
> refactoring that was undone. I'm glad I didn't have more work to  
> recover.

Indeed this is a misconception of the changeset and their underlying  
meta-model.
Class definition and a lot more should not be represented as doit. In  
VW they fixed that.

Now I suggest to use Monticello since you do not get these kinds of  
problems.


>
> > Of course I've also had times when I had to force-quit.
>
> I have always been able to interrupt and save before.
>
> Thanks,
> Zulq.
>