I am really impressed at the winner of the biggest error. It's incredible how much it produces. Right now I'm testing it and running it through pv to see how many errors it makes. Right now on my machine it's producing errors at a rate of 74MB/s. I never thought I'd look at errors in rates of bytes/time.
I ran it for a couple of minutes piped to a log file to catch the errors, I had to kill it because the file was becoming too big. When I closed it, it was 4.9GB (and still growing).
That actually sounds reasonable to me. GCC producing 230GB of error messages is definitely excessive. If it ever completes I'll update here with the actual final count of the length of the error message.
Ok, it's been running all day while I've been working. I don't think it's going to ever complete. I'm at 1.32 TB of output from it. I can't overstate how impressed I am by that. If it completes before I leave then I'll start it again but run it through a compressor and hope I don't fill up my disk. It was large strings of //.//. so i think it should compress well and I'll upload it somewhere for someone to look at.
Yes, thanks! I was originally searching for "c++ pv", which led to lots of c++ examples implementing "present value" calculators, or snippets of code using "pv" as a variable name. I later tried "pv tarball" (figuring that a useful command-line utility would probably be downloadable that way), which led me to http://www.catonmat.net/blog/unix-utilities-pipe-viewer/ and so to http://www.ivarch.com/programs/pv.shtml.
