Martin HS

In my test: - have a trie (`root1`), - add a thing (`root2`), - delete that thing again (`root3` == `root1`). This is the output: ``` nodeset owner: 0x0000000000000000000000000000000000000000000000000000000000000000 [+]:...

Neat. Do you think this is something generic enough that we should include it in the 'production client' ? I I'm thinking this is one fo those one-off things which...

Good point. I was actually kind of wondering the same. After a bit more digging, I found that _preceding_ the `xss-lock` crash, there was another crash. So the first crash...

I think the dump above is a misleading. It came from the `Qubes Manager` logs from dom0, which I believe is the hypervisor logs(?). I later checked the actual `journalctl...

Well, I came to the conclusion that I had hit upon an old bug for which there was a "Solution". I don't mind keeping it open though.

Right, my bad. 

On another note: the immediate is defined as uint16, so should not be referenced to as '-1'. Also, there needs to be not one but two bytes of immediate data...

I don't really feel comfortable being listed as one. This eip has changed a lot. 

@gballet why not just we merge this -- seems to fix an error, as @willianpaixao has shown. The upstream "bug" is not truly a bug -- geth also didn't always...

> @gballet why not just we merge this -- seems to fix an error, as @willianpaixao has shown. The upstream "bug" is not truly a bug -- geth also didn't...