[gentoo-user] To emerge -e world or not to emerge -e world?

[Matt Randolph]

I know that upgrading glibc can cause some programs to break if they 
were built against the previous glibc.  This happens to me all the time 
and I have gotten in the habit of simply re-emerging any packages that 
misbehave since a glibc upgrade.

Well, I have upgraded both glibc and gcc within the last week or so.  
And I've been contemplating a kernel upgrade too.  I looked at genlop 
and it said it will take a mere fourteen hours to re-emerge everything 
with an emerge -e world.  I'm tempted to do it, but I'm wary of making 
major changes to a system that currently seems to be working perfectly. 

However, I've only tested a handful of packages (the ones that I use 
every day) since the glibc upgrade, and I did have to rebuild a few of 
them.  For this reason, I'm guessing that a significant number of the 
packages that I haven't tested are actually broken too.  So when I say 
my system seems to be working perfectly, I think that only applies to 
the packages that I interact with daily and probably not to some of the 
ones that I don't.

When does it make sense to re-emerge everything?  I've heard some people 
say never but that others do it perhaps monthly or even more often.

Is there a (significant) risk that something will go wrong?  Even 
terribly wrong?

Is it possible that some important programs aren't working right now due 
to having been built against an older glibc, and that I'm simply 
oblivious to the fact that they aren't working?  I'm worried 
specifically about system programs that I don't usually have reason to 
interact with, yet may be vitally important to the security and 
stability of my system.
-- 
gentoo-user@g.o mailing list