[Nix-dev] Re: [PATCH 1/2] t/scons

Eelco Dolstra e.dolstra at tudelft.nl
Wed Dec 22 15:22:14 CET 2010


Hi,

On 12/22/2010 01:40 PM, Marc Weber wrote:

> So I don't see any risks. You fail to tell me about those.

I think in general we shouldn't deviate from the upstream behaviour unless there
is a good reason (because upstream presumably knows best).  In this case, people
who use the SCons from Nixpkgs may unwittingly come to rely on the patched
behaviour, and then be unpleasantly surprised when their SCons scripts fail on
other machines.

I do feel that the default path should be patched in Nixpkgs, because leaving it
as is causes a de facto behaviour change (SCons scripts that rely on standard
Unix tools in /usr/bin won't work on NixOS).  However, rather than setting it to
the caller's $PATH (which is unpredictable), it should be initialised to a
deterministic build-time default, e.g.

 "${coreutils}/bin:${gcc}/bin:..."

That is, most of the tools provided by stdenv.  This way, we actually get *more*
predictable behaviour than upstream, so it's at least in the spirit of the
upstream behaviour of not using the caller's environment.  What do you think?

-- 
Eelco Dolstra | http://www.st.ewi.tudelft.nl/~dolstra/



More information about the nix-dev mailing list