Re: [boost] [build] bjam in boost/tools/build/v2/test doesn't work

Subject: Re: [boost] [build] bjam in boost/tools/build/v2/test doesn't work
From: Jim Bell (Jim_at_[hidden])
Date: 2010-12-04 11:04:42

I'm re-posting this (from 11/22). I'd really like to reconcile command-line vs run.py...

On 1:59 PM, Vladimir Prus wrote:

> > Jim Bell wrote:
>> >> Running bjam in boost/tools/build/v2/test doesn't (seem to) run build's
>> >> regressions. Instead, I get:
>> >>
>> >> jamfile.jam:8: in modules.load
>> >> rule project-root unknown in module
> > You should be using 'python test_all.py'. Running 'bjam' there is not really
> > supposed to do anything.
Thanks!

This works and gives me results like this...
    unit_tests                :  FAILED
    module_actions            :  PASSED
    startup_v1                :  PASSED
    ...

I see, too, that '--verbose' enables detailed output, from
BoostBuild.Tester's docstring (and from trying it). But it says "Note
that it does not also enable more verbose build system output like the
--verbose command line option does."  What does that mean? If I run
"python test_all.py --verbose verbose" it doesn't seem to like that
(attempts to run verbose.init).

Also, I'm seeing a different pass/fail pattern between when I run it
like this and when it runs from run.py. I'm running it with mingw-4.4.3
on the same trunk revision that the "MinGW-32 jc-bell.com" regression
has reported.

Picking "absolute_sources", for example, I see
    absolute_sources          :  PASSED

But the regression shows it failed.

How do I reconcile these?