#14558: Unable to parse integer-gmp's Cabal file -------------------------------------+------------------------------------- Reporter: taylorfausak | Owner: hvr Type: task | Status: closed Priority: normal | Milestone: Component: Core Libraries | Version: 8.2.2 Resolution: fixed | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by taylorfausak): Once again I am sad to see this ticket closed without any fix put in place. I am also disappointed in the behavior of GHC developers who appear to be breaking downstream tools and making absolutely no concessions to them. Needless to say, this entire experience has not been positive. I'll leave you with some assorted thoughts: - As has been stated repeatedly in this thread, one of the problems is that build plans that used to work have stopped working. Anybody trying to use a GHC >= 8.2 snapshot with Stack < 1.6.1 will be unable to do so. - `integer-gmp` does not have its own bug tracker. As I said when I opened this ticket: "I was not able to find an issue tracker for integer-gmp. Someone suggested that I open an issue here instead." I asked @hvr on Twitter where to report `integer-gmp` bugs and he said I should report them to him directly. https://twitter.com/taylorfausak/status/938243186237083649 - In my opinion, Stacakge nightlies aren't "unstable". Furthermore, knowingly breaking them is at best inconsiderate and at worst hostile. - It did not take Stack 5 months to support GHC >= 8.2. I have been using Stack with GHC 8.2 since the release candidates. Stack's support of GHC 8.2 was broken by @hvr's package uploads on December 4. - `integer-gmp`'s Cabal file has been updated on the `ghc-8.2` branch of the GHC repository, but it has not been updated on the `ghc-8.4` branch. Perhaps we'll end up in a situation where the repo doesn't match Hackage again. I find myself once again repeating the crux of this ticket:
it's frustrating that a core library is using a bleeding edge feature for basically no reason
I haven't seen anyone even attempt to explain why `integer-gmp` needs to use `^>=` bounds. The argument seems to be that it can, so why not? I think "not breaking other tools" is a pretty compelling reason not to. Apparently GHC developers think breaking other tools is an acceptable price to pay for saving 8 characters in a single Cabal file. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14558#comment:43 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler