> your criticism of Shea's commit is too strong. He changed the default > version of ruby, and the way he renamed those files kept the currently > used naming convention for that expression intact. That is a very > sensible thing to do.
There are two important properties the name can tell about - name - version default.nix tells "I'm the default version", but does no longer tell which version it is 19.nix tells "I'm version 1.9", but does no longer tell that its the default version - so neither is perfect. In any case its "redundancy" which should be avoided if possible cause the default version is not set by filename, but in all-packages.nix. Using sensible names is good habit always. That's why I even proposed default-19.nix (which was kind of joke just to illustrate that both solutions could be accepted). In the end the name of a .nix file doesn't matter its contents do. If there are weak pro and cons I'd vote for keeping as is. And thats what the mail was about - sorry if I didn't talk about this point more clearly. Sorry if my mail sounded strong. I want to avoid noise so that we can focus on things which really matter (such as fixing the ghc-pkg list output) Marc Weber _______________________________________________ nix-dev mailing list [email protected] http://lists.science.uu.nl/mailman/listinfo/nix-dev
