I think I am mentioning this for the third time: Maybe using "-p" and posting the result may help others (or even yourself) to help you, because seeing your actual make databases should help to understand which targets really depend on which others. Generating a log with "-j1" should be okay in order to get a non-garbled output from which to extrapolate what might be going wrong with the parallel runs. Maybe you can even use my GraphViz script in order to visualise the result.
Another approach would be to strip down your makefiles so as to generate a minimal example showing the same problems. Step by step you can get closer to a solution or post a minimal (set of) makefile (s) showing the problematic behaviour. Good luck! -- Alexander Kriegisch _______________________________________________ Help-make mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-make
