Deprecating fromIntegral

[David Feuer]

I wonder if this is something HLint could help with. I imagine many bugs
could be avoided if every use of fromIntegral used visible type application
to indicate explicitly what type was being converted to what type.

On Mon, Aug 10, 2020, 4:34 AM Sven Panne <svenpanne at gmail.com> wrote:

> Am Mo., 10. Aug. 2020 um 09:15 Uhr schrieb Bardur Arantsson <
> spam at scientician.net>:
>
>> On 08/08/2020 15.44, Vanessa McHale wrote:
>> > -1 from me, massive work to overhaul the ecosystem.
>>
>> Why would a massive overhaul be necessary for deprecation? If that's the
>> case then there's a deeper more serious underlying issue around
>> deprecation, IMO.
>>
>
> Two things come to my mind here:
>
>    * You'll probably break quite a few projects which use -Werror. I know
> that there are different opinions regarding -Werror in general, but in any
> case there *will* be breakage.
>
>    * If you consider books and tutorials a part of the ecosystem (which I
> definitely do), there is even more "breakage": From the top of my head I
> would say that quite a few of them use fromIntegral, so deprecation will
> cause confusion.
>
> All these things are definitely fixable, but neither quickly nor without a
> negligible cost. Deprecations should not be done lightly.
>
> Regarding the deprecation itself: I fail to see why fromIntegral is worse
> than (+), (-), (*), ..., and nobody is proposing to remove these. The real
> problem is using fixed-sized numbers where they shouldn't be used, so a -1
> from me.
>
> Cheers,
>    S.
>
> _______________________________________________
> Libraries mailing list
> Libraries at haskell.org
> http://mail.haskell.org/cgi-bin/mailman/listinfo/libraries
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/libraries/attachments/20200810/f70f1494/attachment.html>