[Chart] upcoming minor release

Doug Burke dburke.gw at gmail.com
Fri Mar 21 12:42:20 GMT 2014


On Fri, Mar 21, 2014 at 3:02 AM, Tim Docker <tim at dockerz.net> wrote:

>  Thanks for the patches.
>
> I'm in two minds about -Wall. There's definitely some good come of it:
>
>      - fixes for the unintended partial functions
>      - elimination of redundant imports
>
> but there are some conventions in -Wall that are not really to my taste:
>
>      - I don't mind redundant do keywords - if I have a collection of
> monadic blocks, I'd often rather they looked the same, even if for the
> single statement blocks the do is redundant.
>
>      - I don't really like -fwarn-name-shadowing. I like to choose my
> variable names without having to think about what is defined in every
> enclosing scope.
>
>
I do agree that a strict adherence to some form of a style guide can be
stifling, and that having r, r', r'', ..., r''_'__', ... isn't always
clear. However, some of the renames make it easier for me (at least) to
work out what's going on, so I don't think there's an easy answer.

The one Haskell style guide I know of is Johan's at
https://github.com/tibbe/haskell-style-guide - are there any others?


> Do you think these changes are for the better?
>
>
I like the warm fuzzy feeling I get when doing a build and not see any
warnings swoosh by, and feel that overall it's an improvement, but I can't
honestly say that some of the changes I've made - in particular the ones
you've called out - make the code better. You could pick and choose the
warnings rather than go for -Wall, but then that's a maintenance overhead
since these flags change with compiler releases.

Doug


Tim
>
>
> On 18/03/14 00:07, Doug Burke wrote:
>
>
>  Tim,
>
>  With regards to the 1.2.1 release, I have a PR that improves my rlabel
> fix from changeset  355edca6f656836f4d1a66f548cd3b2fdbd5f286 . It also
> includes additional "cleanup" changes to avoid warnings when built with
> -Wall (and adds this to the cabal file), since that's how I found the
> partial functions in the first place. This also showed up some missing
> exports (mainly lens accessors, although there are a few types that are now
> exported). There are also some hlint changes; I'm not sure of the coding
> standards for chart, so these may or may not be appropriate. Let me know if
> you want things broken up/changed.
>
>  I am also playing around with adding support for polar plots (i.e.
> r,theta), but I don't think that it's at a state worth releasing at this
> time. For those that are interested, it's in my polar branch at
> https://github.com/DougBurke/haskell-chart/tree/polar
>
>  Doug
>
>
>
> _______________________________________________
> Chart mailing listChart at projects.haskell.orghttp://projects.haskell.org/cgi-bin/mailman/listinfo/chart
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://projects.haskell.org/pipermail/chart/attachments/20140321/86ea2a09/attachment.htm>


More information about the Chart mailing list