VM fix for 4.0 - what was it?

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

VM fix for 4.0 - what was it?

Bill Schwab
Andy,

Do you recall the nature of the VM fix that you produced for 4.0?  Sorry to
have to ask, but, it appears not to have hit the archive yet.  Any pointers
to the updated VM alone would be appreciated; I remember that Blair put it
on the web, but, can't find the URL now.  I've downloaded the updated MSI
file, but, it would be nice to patch the VM first to see if that helps.

The VM bug/fix is one of several possible factors that I'm considering in a
strange problem that I'm having.  The good news is that I have a recent
backup that's stable, at least on my work machine.  The bad news is that the
"good" image still seems to have problems on my machine at home.  One thing
that I'm considering is bad RAM, though I'd have much sooner believed it of
the box at work as this machine is remarkably stable.  However, a hardware
problem of some sort comes to mind especially given that I suffered some
kind of image and/or change log problem in 3.06 just some weeks ago.  In
that case though, one could argue that it might have resulted from a
partularly ugly lockup of my machine at work.

Whatever the cause this time, the image seems to work fine but deployed
executables produced from it don't run.  Interestingly, the failure is
severe enough that it's necessary to reboot the machine before anything
using the VM can run - I observed that here (home) and with an app that I
built here and ran at the office.  Sadly, I didn't think to bring an app
from there to here.

It appears that all of my packages have saved to disk, so it might be time
to backup and build a new image.  Any other suggestions?

Have a good one,

Bill



--
Wilhelm K. Schwab, Ph.D.
[hidden email]


Reply | Threaded
Open this post in threaded view
|

Re: VM fix for 4.0 - what was it?

Blair McGlashan
Bill

You wrote in message news:921b4p$s1p$[hidden email]...
>
> Do you recall the nature of the VM fix that you produced for 4.0?

It was a fix for a GPF that occurred when attempting to evaluate a block
with a ^-return when execution had already returned below the block's home
context in the stack. This is unlikely to show up in practice at is is a
pretty unusual error condition that one typically sees only in development,
and even then only rarely since it is a coding error. It's easy to spot in
the crash dump as a GPF in the middle of #cannotReturn:

>...Sorry to
> have to ask, but, it appears not to have hit the archive yet.  Any
pointers
> to the updated VM alone would be appreciated;

Try http://homepage.ntlworld.com/blairmcg/DolphinVM004.zip

Regards

Blair