Again (think I posted this before already), a very good source for java mem spec behaviour is the comment of Brian Goetz
http://www.cs.umd.edu/~pugh/java/memoryModel/jsr-133-faq.html This also explains why you need to declare the field volatile - don't wanna stress this, but I told this also before... ;) LieGrue, strub ----- Original Message ----- > From: Leonardo Uribe <[email protected]> > To: MyFaces Development <[email protected]>; Mark Struberg > <[email protected]> > Cc: > Sent: Wednesday, October 26, 2011 8:40 PM > Subject: Re: MYFACES-3368 status > > Hi > > I'm keeping an eye on the changes done. I don't remember any problem > about 'double lock' on myfaces. Could you please be more specific? In > my understanding everything works as expected. > > regards, > > Leonardo Uribe > > 2011/10/26 Mark Struberg <[email protected]>: >> Hi folks! >> >> I'm down to 24 checkstyle issues (with LineLength=160, otherwise ~600 > more) which are really non-trivial. >> >> A few of them are 'equals() without hashCode()' which are > definitely problematic >> Then we have a few variable namings which I didn't want to change > without talking to you. >> And of course there are some 'double lock idiom' issues. >> It's safe to use double-locking as of Java5, but ONLY if the locked > variable is declared volatile! We barely have this, so I left the check > enabled... >> >> >> Who is taking care of those remaining issues? Leo? >> >> >> LieGrue, >> strub >> >> >
