As far as I understand it, env.Detect calls env.WhereIs with path=None, which means it uses env['ENV']['PATH']. By default, a new Environment doesn't inherit the user's shell environment into env['ENV'], but users can modify it as they want. So in a real user's code, the path Detect will use depends on what they put into their SConstruct.
I see your point about the tests being more lenient in where they look; it makes it hard to skip tests reliably. Perhaps the tests that do that should add their executable's dir to env['ENV']['PATH'] when running the tests? If you like that, perhaps we could codify it by making it easy to do from the test framework. On Sun, Mar 2, 2014 at 2:40 PM, Russel Winder <[email protected]> wrote: > I think I have ranted about this before, but the red haze must have been > so strong I failed to note the well reasoned responses you folk will > have provided. > > The core problem is that TestSCons.TestSCons().where_is(x) uses the > user's path for executable look up whereas Environment.Detect(x) does > not. Thus canonical, idiomatic tests (based on the tests in the SCons > codebase) are inconsistent with canonical, idiomatic tools (based on the > tools in the SCons codebase). Basically there is a fundamental > assumption that all executables are in (for Linux and OSX) /usr/bin > without fail, and that anything else is unacceptable. > > If the assumption SCons wants to make is that there is /usr/bin and > nothing else fine, but the tests framework should reflect that - even > though it would be wrong ;-) > > So can someone ignore all the tests of the SCons system and tell me how > to make the decision in a test of whether env.Detect(x) will find the > executable. > > I'm off for more Ricard so all future emails tonight will be > incomprehensible... > > -- > Russel. > > ============================================================================= > Dr Russel Winder t: +44 20 7585 2200 voip: > sip:[email protected] > 41 Buckmaster Road m: +44 7770 465 077 xmpp: [email protected] > London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder > > _______________________________________________ > Scons-dev mailing list > [email protected] > http://two.pairlist.net/mailman/listinfo/scons-dev > -- Gary
_______________________________________________ Scons-dev mailing list [email protected] http://two.pairlist.net/mailman/listinfo/scons-dev
