Reconsidering IVY-1454

[Charles Duffy <ch...@dyfis.net>]

Howdy, all --

I've been seeing race conditions around mutating methods in IvySettings
resulting in occasional NPEs during IVY-1454 regression testing. We're also
getting issues with ConcurrentModificationExceptions from the logger stack
-- which it seems to me we'd need to move the loggers into thread-local
storage to resolve.

Frankly, there's enough side-effecting code throughout the codebase that
I'm uncomfortable only patching the immediately visible issues. What are
folks' thoughts on putting a big lock around the resolve task, and/or
documenting loudly that Ivy is not supported for concurrent use?

Re: Reconsidering IVY-1454

[Nicolas Lalevée <ni...@hibnet.org>]

Le 26 déc. 2013 à 23:38, Charles Duffy <ch...@dyfis.net> a écrit :

> Howdy, all --
> 
> I've been seeing race conditions around mutating methods in IvySettings
> resulting in occasional NPEs during IVY-1454 regression testing. We're also
> getting issues with ConcurrentModificationExceptions from the logger stack
> -- which it seems to me we'd need to move the loggers into thread-local
> storage to resolve.
> 
> Frankly, there's enough side-effecting code throughout the codebase that
> I'm uncomfortable only patching the immediately visible issues. What are
> folks' thoughts on putting a big lock around the resolve task, and/or
> documenting loudly that Ivy is not supported for concurrent use?

I remember that Xavier wrote once that he wouldn't trust much Ivy's code to be thread safe. I would second that.
So my preference for now would be to "just" document it.

Nicolas


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@ant.apache.org
For additional commands, e-mail: dev-help@ant.apache.org