Non-monotonic releases and package clobbering

Hello all, I recently noticed that there is a subset of packages on Hackage which have had a non-monotonic release in their history. Non-monotonic releases occur for various reasons; for example, http://hackage.haskell.org/package/QuickCheck-1.2.0.1 was made in 2010 (2.1 was released in 2008), ostensibly as a minor bugfix release. However, I found at least one case where a non-monotonic release was due to package clobbering: http://hackage.haskell.org/package/Adaptive-0.1 http://hackage.haskell.org/package/Adaptive-0.22 It's easy to see that these are two different packages. 0.1 is the "later" release, but it is completely shadowed by the old 0.22 release. I'm not sure if I have any concrete recommendations (since this problem is mostly subsumed by the "people shouldn't be able to upload arbitrary stuff to Hackage problem); but I thought this might be interesting information to someone. Edward

Clearly the solution for this particular conflict is just to vary the case
of the first letter, like we have with the Numbers vs. numbers packages.
What could be cleaner? ;)
On Sun, Nov 25, 2012 at 12:59 PM, Edward Z. Yang
Hello all,
I recently noticed that there is a subset of packages on Hackage which have had a non-monotonic release in their history. Non-monotonic releases occur for various reasons; for example, http://hackage.haskell.org/package/QuickCheck-1.2.0.1 was made in 2010 (2.1 was released in 2008), ostensibly as a minor bugfix release.
However, I found at least one case where a non-monotonic release was due to package clobbering:
http://hackage.haskell.org/package/Adaptive-0.1 http://hackage.haskell.org/package/Adaptive-0.22
It's easy to see that these are two different packages. 0.1 is the "later" release, but it is completely shadowed by the old 0.22 release.
I'm not sure if I have any concrete recommendations (since this problem is mostly subsumed by the "people shouldn't be able to upload arbitrary stuff to Hackage problem); but I thought this might be interesting information to someone.
Edward
_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries
participants (2)
-
Daniel Peebles
-
Edward Z. Yang