Subject: Re: [Boost-build] using <dependency> in Bjam
From: Robert Ramey (ramey_at_[hidden])
Date: 2009-03-25 19:24:53
you look at the current trunk tests for bronf1 you can see what I mean.
Uh oh - looks like I didn't look closely enough. The results
page shows test fail when in fact it wasn't run because of a failure
in a dependent test. So it seems that contrary to one's first impression
it works as advertised. I was sure I had observed a problem with this on
test setup but I'll double check.
Vladimir Prus wrote:
> On Wednesday 25 March 2009 23:40:36 Robert Ramey wrote:
>> There are a bunch of tests in the serialization library which make no
>> sense to run if the compiler/library doesn't support wide character
>> i/o. In order to skip these, I've included the following rule in the
>> It doesn't seem to be skipping tests.
> Can you provide more precise information? What test is run despite
> what configuration check failing? How can I reproduce it?
> - Volodya
> BTW, Boost.Jam is low-level engine, and does not have <dependency>
> or anything like that -- these are provided by Boost.Build
> Unsubscribe & other changes:
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