Re: Proposal: Add readMaybe (and possibly readEither) to Prelude, make Haddocks for read more cautionary

When updating the documentation for Read, one can at the same time stress that this class is for parsing values in *Haskell* syntax. Same for Show, it should be stressed that it is for printing values into their *Haskell* linearization. I'd guess "Show" is the most abused type class. (Our own code base, Agda, is a prime example for this.) Currently we have: https://hackage.haskell.org/package/base-4.9.0.0/docs/Text-Show.html Converting values to readable strings: the Show class and associated functions. https://hackage.haskell.org/package/base-4.9.0.0/docs/Text-Read.html Converting strings to values. This description is way too generic/generous... Cheers (and a happy 2017), Andreas On 30.12.2016 09:43, David Menendez wrote:

On Fri, Dec 30, 2016 at 2:13 AM, Sven Panne mailto:svenpanne@gmail.com> wrote:

2016-12-30 5:50 GMT+01:00 David Menendez mailto:dave@zednenem.com>:

[...] I don't think making life easy for -Wall clean people should be a goal. The whole point of warnings is that they indicate things that might not be a problem. Otherwise, they’d be errors. This is especially true for warnings that only show up if you use -Wall instead of -W.

This is largely a matter of personal preference, and this is probably even changing over time: 10-20 years ago, I didn't care much about -Wall (in various languages/compilers) too much, but this has changed with experience in tons of projects: Basically each and every warning turned into a bug sooner or later, with very, very few exceptions. So I'm basically a hardcore -Wall-clean-fanatic nowaydays. :-) Not using -Wall doesn't make the problems go away, you only discover them much, much later, probably when your SW is shipped to your client.

So whatever is done, it should be easily be possible to be -Wall-clean, which basically means more control over warnings. Especially important are one-shot things like the usual C/C++'s NOLINT ("I know what I'm doing here, really!") comments, which make it possible to be extremely fine-grained about warnings.

Warnings from compilers are just like people crying for help: If you see them too often, you get used to them and ignore them, which in the long run is bad for all parties involved…

Why -Wall and not -W? If something is almost always an issue, shouldn’t it be in -W?

The point I was trying to make is that we don’t want to prevent compiler developers from adding warnings (or library developers from deprecating things) merely to make life easier for people who want to be -Wall clean.

Having a syntax to disable a warning is not a bad idea, even though it brings to mind the “please” keyword from Intercal.

-- Dave Menendez mailto:dave@zednenem.com> http://www.eyrie.org/~zednenem/

_______________________________________________ Libraries mailing list Libraries@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries

-- Andreas Abel <>< Du bist der geliebte Mensch. Department of Computer Science and Engineering Chalmers and Gothenburg University, Sweden andreas.abel@gu.se http://www2.tcs.ifi.lmu.de/~abel/