I'm no longer able to run regression tests, and quick look at ftp://fx.meta-comm.com/boost-regression/CVS-HEAD/ reveals that few others have this problem too. Apparently bjam is called from regression.py in wrong way, or Jamfile is broken, as bjam.log ends with: don't know how to make None don't know how to make -sALL_LOCATE_TARGET=C:\MinGW\Bronek\boost\results ...found 2 targets... ...can't find 2 targets... B.
Bronek Kozicki wrote:
I'm no longer able to run regression tests, and quick look at ftp://fx.meta-comm.com/boost-regression/CVS-HEAD/ reveals that few others have this problem too. Apparently bjam is called from regression.py in wrong way, or Jamfile is broken, as bjam.log ends with:
don't know how to make None don't know how to make -sALL_LOCATE_TARGET=C:\MinGW\Bronek\boost\results ...found 2 targets... ...can't find 2 targets...
My fault. Should be fixed now. Please try again. Markus
I forgot one thing. You have to manually get the latest regression.py script from boost cvs. It's in tools/regression/xsl_reports/runner/regression.py. Otherwise, the next run will still fail as it updates itself only after the regression run is done. Sorry for the troubles. Markus
Bronek Kozicki wrote: [...] here is how bjam is started by regression.py : "C:\MinGW\Bronek\boost\bjam.exe" "-sTOOLS=mingw mingw-3.4.1 como-win32-4.3.3-vc7.1 vc7.1" "-sBOOST_ROOT=C:\MinGW\Bronek\boost\boost" -d2 --dump-tests None "-sALL_LOCATE_TARGET=C:\MinGW\Bronek\boost\results" B.
Bronek Kozicki <brok@rubikon.pl> writes:
I'm no longer able to run regression tests, and quick look at ftp://fx.meta-comm.com/boost-regression/CVS-HEAD/ reveals that few others have this problem too. Apparently bjam is called from regression.py in wrong way, or Jamfile is broken, as bjam.log ends with:
don't know how to make None ^^^^ This seems to indicate that the problem is in regression.py Can you analyze the changes to that file from the past 24 hours?
-- Dave Abrahams Boost Consulting http://www.boost-consulting.com
David Abrahams wrote:
This seems to indicate that the problem is in regression.py Can you analyze the changes to that file from the past 24 hours?
Markus fixed it and now it works - however I had to manually replace broken regression.py with the one from tools/regression/xsl_reports/runner . I think that other runners should take care of it, too (Cc-ed). BTW the change was adding "bjam_options", apparently not set before being used when creating bjam comand line "test_cmd = ...", line 438. B.
participants (3)
-
Bronek Kozicki -
David Abrahams -
Markus Schöpflin