On Mon, Dec 11, 2017 at 1:20 PM, Eric S. Raymond <e...@thyrsus.com> wrote: > Jason Merrill <ja...@redhat.com>: >> I've pushed to the thyrsus gcc-conversion repository the rest of my >> work from last year, including a few changes to the attribution map >> and a post-processing script. > > Got it. I will, as before, be pushing my lift file mods to that repository. > >> It has occurred to me that for the period since the creation of >> gcc.gnu.org we might just use @gcc.gnu.org addresses (which are likely >> to forward) rather than try to nail down everyone's current address. >> But using the address from the ChangeLog sounds better. > > That's your policy choice to make, I'm just the mechanic. The map file > can be replaced or updated at any time. > > What do your repo load times look like, and how much RAM is in your > surgery machine? My loads are currently hanging in D state after 13 hours on > a 64GiB hexacore Xeon 3 running at nominally 3.50GHz. I'm beginning > to wonder if I have some kind of high-memory error - I don't recall this > happening last time I took a swing at the conversion.
I think it was on the order of 18-20 hours, but it's been a while. The machine I was using before actually has more memory now than it did then; free -g reports 251G total memory. Jason
