__NHC__ ifdefs in the core libraries

Hi all, There are a couple of hundred ifdefs for __NHC__ in the core libraries, but as far as I know nhc/yhc never switched over to the git repositories for these libraries, and in any case I believe they are now considered unmaintained projects. Even if someone decides to resurrect them, it looks like most of the ifdefs represent bugs that ought to be fixed (for example, Word being a type synonym, [] needing to be im/exported, and an instance export bug). Meanwhile, they make life harder for people working on the libraries: When making a change, one feels that one ought to also update the NHC alternative where appropriate, even though it has probably bitrotted already, and despite not being able to test it. So I propose that we remove all the __NHC__ ifdefs (and also base/NHC/) in the core libraries. What do you think? Thanks Ian

I agree that you should delete all __NHC__ ifdefs in the git repositories for the base library. Even though nhc98 does still continue to build every night, it does not, and never will, use the git repositories. Regards, Malcolm On 11 Feb 2013, at 21:01, Ian Lynagh wrote:
Hi all,
There are a couple of hundred ifdefs for __NHC__ in the core libraries, but as far as I know nhc/yhc never switched over to the git repositories for these libraries, and in any case I believe they are now considered unmaintained projects.
Even if someone decides to resurrect them, it looks like most of the ifdefs represent bugs that ought to be fixed (for example, Word being a type synonym, [] needing to be im/exported, and an instance export bug).
Meanwhile, they make life harder for people working on the libraries: When making a change, one feels that one ought to also update the NHC alternative where appropriate, even though it has probably bitrotted already, and despite not being able to test it.
So I propose that we remove all the __NHC__ ifdefs (and also base/NHC/) in the core libraries. What do you think?
Thanks Ian

+100!
On 11 February 2013 13:01, Ian Lynagh
Hi all,
There are a couple of hundred ifdefs for __NHC__ in the core libraries, but as far as I know nhc/yhc never switched over to the git repositories for these libraries, and in any case I believe they are now considered unmaintained projects.
Even if someone decides to resurrect them, it looks like most of the ifdefs represent bugs that ought to be fixed (for example, Word being a type synonym, [] needing to be im/exported, and an instance export bug).
Meanwhile, they make life harder for people working on the libraries: When making a change, one feels that one ought to also update the NHC alternative where appropriate, even though it has probably bitrotted already, and despite not being able to test it.
So I propose that we remove all the __NHC__ ifdefs (and also base/NHC/) in the core libraries. What do you think?
Thanks Ian
_______________________________________________ Libraries mailing list Libraries@haskell.org http://www.haskell.org/mailman/listinfo/libraries

On Mon, Feb 11, 2013 at 1:01 PM, Ian Lynagh
So I propose that we remove all the __NHC__ ifdefs (and also base/NHC/) in the core libraries. What do you think?
+1 I deleted all the __HUGS__ ifdefs in network recently. If no one is testing a given ifdef branch on a semi-regular basis it will bit-rot. Such branches should be removed. To paraphrase Beyonce: if you liked it you should have written a test* for it. ;) * Or have a continuous build, but that doesn't make for a very catchy sentence.

Hi all,
-----Original message----- From: Ian Lynagh
Sent: 11 Feb 2013, 21:01 Hi all,
There are a couple of hundred ifdefs for __NHC__ in the core libraries, but as far as I know nhc/yhc never switched over to the git repositories for these libraries, and in any case I believe they are now considered unmaintained projects.
Even if someone decides to resurrect them, it looks like most of the ifdefs represent bugs that ought to be fixed (for example, Word being a type synonym, [] needing to be im/exported, and an instance export bug).
Meanwhile, they make life harder for people working on the libraries: When making a change, one feels that one ought to also update the NHC alternative where appropriate, even though it has probably bitrotted already, and despite not being able to test it.
So I propose that we remove all the __NHC__ ifdefs (and also base/NHC/) in the core libraries. What do you think?
+1. I never really touch any __NHC__ (or __HUGS__) as I cannot test is. Moreover, we have all the fancy versioning systems -- if the need ever arises, we can still find the content of the deleted ifdefs... Cheers, Milan
participants (5)
-
David Terei
-
Ian Lynagh
-
Johan Tibell
-
Malcolm Wallace
-
Milan Straka