On 02/26/2014 06:24 PM, Andrew Hughes wrote: > > > ----- Original Message ----- >> On 02/26/2014 02:47 PM, Andrew Hughes wrote: >>> This should be backported. It fixes this along with other such issues: >>> >>> changeset: 3029:871cffb21423 >>> user: alanb >>> date: Thu Oct 07 14:36:17 2010 +0100 >>> summary: 6989466: Miscellaneous compiler warnings in java/lang, >>> java/util, java/io, sun/misc native code >>> >>> http://hg.openjdk.java.net/jdk7/jdk7/jdk/rev/871cffb21423 >> >> Heavens, that's a big patch. Why do we want it all? > > Fairly long, but nothing controversial. > > I'd rather keep us in sync with 7 than creating new bugs & patches that > make it harder to trace things later.
OK, but that is a really important semantic change buried in a patch that is just a ton of warning fixes. As an example of bad process that is pretty hard to beat, even if it does tick all the boxes. Warning fix patches have a horrible history of breaking things. We can hope that these patches will all work when back ported, but how do you know that they do? Andrew.
