From: David Abrahams (dave_at_[hidden])
Date: 2002-08-21 14:07:41
From: "bill_kempf" <williamkempf_at_[hidden]>
> What you describe as the actual behavior, and what I see when I use
> the -a option (or do a clean first), are in fact quite usable.
> However, I'm experiencing several issues, some of which were the
> result of the confusion here. Let me describe them.
> Just modifying a test to produce an error and then invoking:
> bjam -sTOOLS=vc7
> appeared to run, with out any indication of failed targets.
None whatsoever? Please give me instructions for reproducing this.
> inspection did reveal there was no "capture-run-output", however, so
> I'm not sure that it ever tried to run the test.
It didn't. It sounds like it failed to build some dependency of the run
test and stopped there.
> I'm guessing it
> only rebuilt the tests with out running.
If the test didn't run, it should only be because the .exe or one of the
other dependencies failed to build.
> Deleting the bin directory entirely results in the same behavior:
> i.e. everything compiles, no failed targets are reported, and
> there's no "capture-run-output" in the output.
> I can't explain any of that behavior.
Me neither. Isn't that the same as what happens when you do a clean first?
BTW, are you explicitly requesting the "test" target? Tests don't run as a
result of building "all"
> Another thing I notice is that only one target (including it's 4
> variants) is being built, and is named test_base.test. It appears
> that this is the target based on test_thread.cpp. The other 5
> targets aren't built or run at all.
Perhaps because they're up-to-date?
David Abrahams * Boost Consulting
dave_at_[hidden] * http://www.boost-consulting.com
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk