Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [wtp-dev] Bug 306757 - ModelManagerImpl is highly susceptible to deadlock

Hi Nitin,

I have added a comment to 306757 to continue discussion.



----- Original Message -----
From: nitind@xxxxxxxxxx
To: wtp-dev@xxxxxxxxxxx
Sent: Monday, March 29, 2010 6:54:10 PM GMT -08:00 US/Canada Pacific
Subject: [wtp-dev] Bug 306757 - ModelManagerImpl is highly susceptible to deadlock

One of the things I recall discussing with JSF committers at EclipseCon 
had to do with a timeout on the wait() call that's used to synchronize 
(and avoid duplicated effort during) model oads--it normally had no 
timeout and we'd added the ability to set one into 3.1.x.  We'd actually 
only put it in that controllable timeout for WTP 3.1.x, planning to have a 
better aolution for 3.2.  Well, that solution is finally upon us in the 
form of bug 306757 and the changes to the Jobs system of which it makes 
use. After quite a bit of review, it's been dropped into the build.  I 
wanted to give the JSF team (and anyone else making use of that settable 
timeout) a very visible request to keep a lookout for any new issues that 
might come up as a result of this change.  This should eliminate the need 
for a controllable wait() timeout, but as always we'd appreciate you 
letting us know if you run into more problems as a result of this change.

Nitin Dahyabhai
Eclipse WTP Source Editing (and JSDT, too)
IBM Rational

wtp-dev mailing list

Back to the top