Suggestion: A new variable for a few Makefiles: IS_BINARY

Doug Barton dougb at
Fri Jan 22 01:22:56 UTC 2010

I'm sorry, you still haven't answered my questions:

1. What dangers are you trying to protect users from?

2. Why do you feel that existing safeguards for what goes into the ports
tree are not adequate?

Responding indirectly to your last post, the ports infrastructure is
already VERY complex. Complex systems by their very nature are fragile.
Adding more complexity (and therefore more fragility) without a very
good reason is a very bad idea. You are asserting that a change is
necessary. Therefore it's on you to prove that the benefits of the
proposed change outweigh the costs.

Some of the costs that come immediately to mind:
1. Someone has to write the patch
2. Someone has to test it
3. Someone has to identify all existing ports that install binaries.
3. Someone has to add flags to all existing ports that install binaries
4. Someone has to verify that new ports that are added to the tree
either do or do not have binaries, and that flags are set accordingly
5. Someone has to maintain the code once it's in the tree

Your assertion from your previous message seems to be that "it should be
done because it will help some people, and won't hurt anyone who doesn't
use it." I disagree with your analysis. I see significant "costs" in
terms of person-hours to do the things that are in that list above, and
I'm sure there are other costs that I'm not thinking of.

So what I'm asking you to do is outline in detail what benefits your
proposed change will bring to the community that will justify the cost
of making the change.


More information about the freebsd-ports mailing list