On Thu, Jun 5, 2014 at 3:55 PM, Rene Rivera <email@example.com> wrote:
On Thu, Jun 5, 2014 at 3:17 PM, Niklas Angare <firstname.lastname@example.org> wrote:
"Tom Kent" <email@example.com> wrote:Unless you think you'll have to run the debugger on the output binaries frequently, running run.py with --bjam-options="debug-symbols=off" seems like a decent work-around. I do it for my runners to save SSD life.
I recently got access to a machine to setup as a tester, but it only
has just under 50GB of space available. It would be great to get
another tester running, but I can't unless we fix the cleanup problem.
Does anyone have any further information about this? Ideas how to fix
Of course, a real solution wouldn't hurt :)Can't promise anything.. But I'll look into removing more test binaries. Have to be careful because loosing dependencies could cause test to get run all the time (or not at all).OK, I've implemented some additional test targets removal code. Because I did not want to impact current testing the additional removal is not enabled. To use add a "--remove-test-targets" option to the b2 invocation. This will remove all executable and objects for all test targets (it doesn't remove them for dependent library targets though). I'd appreciate some feedback as to how it works and if there are any problems since it's not obvious code.