1.9.2

[Ryan Schmidt]

I don't disagree that the correct solution may be complicated, and that what you put in is a simpler compromise that does avoid the race condition, but at such a great expense that I'm pretty sure I'm going to have to revert it in my personal installations of MacPorts at least. It's completely common for me to be building a large port (gcc45, qt4-mac, mysql5, etc.), and during the hours it takes to do that, I'll want to also work on other MacPorts things, which your change now prevents.