3.0.3 and 3.2 releasing

[Sascha Wildner]

On Fri, 24 Aug 2012 16:03:09 +0200, John Marino <dragonflybsd at marino.st>  
wrote:

> My point is that it was tagged in the middle of active bug fixing that  
> applies to 3.0.x series.  Why not wait until the bug fixing is over?  
> We're talking about a day or two.  Again, I see no point in releasing  
> (updating) damaged goods with the known fix at hand.

I see b148267406ef2d0543d5d87d15c283b2d314516f ("kernel - Attempt to fix  
i386 wire_count panic (98)") which says it's an attempt to fix the wire  
count bug and I see 554cf9ac8ff4af8eedd0f5896dfa5d2e5f61038d ("kernel -  
Fix pmap_remove() issue.") which says "I don't know if this had anything  
to do with the wire_count panic." in the message. Comments in the tracker  
are not definitive either. Last one says that  
b148267406ef2d0543d5d87d15c283b2d314516f is "promising". Nothing of it has  
been MFC'd either.

It really should get all the time it needs (in master) but I still think  
nothing should block us from updating our ISOs with fixes that have been  
in the branch for months now, some of them important. And I still don't  
see why that should depend on unfixed issues. It's not like rolling a new  
ISO is a huge amount of work (most of the time spent is waiting until it's  
built and uploading it), that can't be done again as new (important) fixes  
come in.

But I guess we can only agree to disagree here.

Sascha