[e-lang] E 0.9.3 release?

Mark S. Miller erights at google.com
Fri Nov 12 09:41:35 PST 2010

On Tue, Nov 9, 2010 at 4:01 AM, Thomas Leonard <
tal at it-innovation.soton.ac.uk> wrote:

> On Sat, 2010-11-06 at 20:11 -0700, Mark S. Miller wrote:
> > On Sat, Nov 6, 2010 at 10:54 AM, Thomas Leonard <talex5 at gmail.com>
> > wrote:
> >         The 0.9.2a release currently on the main web-site is rather
> >         buggy. In
> >         particular:
> >
> >         - it often fails to start (fixed in r718)
> >         - it contains a known security flaw (fixed in r728)
> >         - it suffers from various race conditions while running
> >
> >         Any chance of a 0.9.3 release?
> >
> >
> > Yes, that's a good idea. I should be able to get to that in early
> > December.
> >
> >
> > Sorry I haven't been finding the time to review your outstanding
> > changes. Could you maintain a page at wiki.erights.org pointing to
> > outstanding changes, so that as I and others do find time to review
> > these, we can easily find them? Thanks.
> Done: http://wiki.erights.org/wiki/User:ThomasLeonard
> I think the CapTP shutdown patches are the only ones that might be
> useful in the next release.
> The Scope/EEnv split may be an API change, so it would make sense to get
> a new stable release out first.

Much appreciated!

 1. Fixed infinite-loop processing unknown updoc directives

Looks good, yes.

2. Removed UserThread that was created per ListenThread
(unless someone can explain what this was for)

Assuming I understand the outcome of the conversation between you and Bill
-- that this is no longer needed -- then yes, looks good.

3. Allow result of orderlyShutdown to resolve
The result is now a promise in the sending vat, which is resolved by
the VatRedirector. Before, it returned a promise in the vat being shutdown,
which ignored __whenMoreResolved messages.
This should also fix thread-safety bugs in mergeInto, orderlyShutdown and

I think so, but I'm not ready to approve this yet. Need to spend more time
thinking about it and reminding myself of how all this works internally.

4. NearRef.sendMsg propagates exceptions

Please rephrase "Throwing away any problem that is returned here." as
"Ignoring any ...". The term "throwing away", when applied to an exception,
might be confused with throwing the exception. With that changed, looks

5. Added static Runner.whenDead() method to register a deadManSwitch
(not sure whether __reactToLostClient is the right method to call here)

I didn't understand the purpose of this until looking at the next CL.

The synchronous call needs to be wrapped in a try/catch that catches, logs,
and otherwise ignores error thrown by the call.

6. Shut down VatTP automatically when its vat is shut down

I'll have to get back to this another day.

7. Added "?? shutdown vat @vat" updoc directive

Postponed as well.

Thanks again for doing all this!!!

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.eros-os.org/pipermail/e-lang/attachments/20101112/2209159f/attachment.html 

More information about the e-lang mailing list