ok, thank you<div><br></div><div>-- Pavel<br><br><div class="gmail_quote">On Sun, May 26, 2013 at 4:49 PM, Camillo Bruni <span dir="ltr"><<a href="mailto:camillobruni@gmail.com" target="_blank">camillobruni@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br>
We had recently the "same" problem.<br>
CompiledMethods are dual objects, storing literals (full oops) in the first<br>
section and storing bytes in the second part. It does not make sense to read<br>
a single byte from an OOPs (and even worse, change a byte)<br>
<br>
<br>
In short you have to use #header and #literalAt: for the first part and<br>
#basicAt: for the byte part.<br>
<div class="HOEnZb"><div class="h5"><br>
<br>
<br>
On 2013-05-26, at 16:34, Pavel Krivanek <<a href="mailto:pavel.krivanek@gmail.com">pavel.krivanek@gmail.com</a>> wrote:<br>
> Hi,<br>
><br>
> for the Pharo Kernel ImageMap job I do something like<br>
><br>
> m := CompiledMethod allInstances anyOne.<br>
> 1 to: m basicSize do:[:i | m basicAt:i ].<br>
><br>
> For previous VMs it worked well without error but now the primitive fails<br>
> and the SubscriptOutOfBounds error is raised at index 1.<br>
> Fails on the latest Pharo VM and VM by Eliot r2714.<br>
><br>
> Cheers,<br>
> -- Pavel<br>
<br>
</div></div></blockquote></div><br></div>