On Jan 25, 2014, at 5:49 PM, Marshall Clow <mclow.lists@gmail.com> wrote:

On Jan 25, 2014, at 4:58 AM, Beman Dawes <bdawes@acm.org> wrote:

On Sat, Jan 25, 2014 at 3:42 AM, Joaquin M Lopez Munoz <joaquin@tid.es> wrote:
Hi,

Unless I'm mistaken, the develop testing results page currently
has two testers running *master* revisions:

* marshall-mac: rev f612fc
* DebSidC++: rev f612fc

I don't know whether this is a misconfiguration problem on the testers'
side or a bug in the testing/reporting scripts.

The "How do you interpret the SHA1 revision (outdated results)?" thread came to the came conclusion.

My guess for a fix is marshall-mac and DebSidC++ to:

* Delete the test directory.
* Start fresh in a new test directory as if these runners were brand new.  "Starting fresh" should include downloading a fresh copy of run.py.
* To speed testing, just run for regex. I.E. add --libraries=regex to the run.py command line.

Done.

Here’s the invocation:

cd /Volumes/TwoTB/boost/regression/develop
python run.py --runner=marshall-mac --tag=develop \
--toolsets=darwin,clang,clang-11,clang-tot,clang-tot11,clang-asan,clang-asan11 \
--bjam-toolset=darwin --pjl-toolset=darwin --bjam-options="-j 12”

I came in to work this morning, and my machine was still doing this.
More precisely, “process_jam_log” was running, taking 100% of one core, and apparently has been doing so for many, many hours.
I’d put up “bjam.log” for people to look at, but it is 11.6 GB long.

Suggestions?

— Marshall