Hi Alan
On 31 July 2018 at 19:21, Alan W. Irwin <Ala...@gm...> wrote:
> Hi Phil:
>
> I am moving this discussion to plplot-devel where it belongs.
Sorry, I hadn't realised we'd gotten off list. Thanks for moving it back.
>
> I think your backwards-incompatible changes [h,l,s ==> c1,c2,c3
> changed member names] in the PLControlPt struct are likely a good idea
> since h,l,s was a misnomer. Nevertheless, those changes did cause
> important build problems in the Tk-related part of the build which I
> have now fixed (commit 94e2e5b01).
Is this struct really part of our public API? It is not mentioned
anywhere in the docs as far as I can tell so I presumed this was an
internal use only structure.
>
> Example 12 (which uses RGB interpolation) shows the effect of your
> change. The old code produced
> <http://plplot.org/examples-data/demo12/x12.01.png> which shows fairly
> discontinous colour results from 1982 to 1986, while the new code
> produces much smoother colour changes in that date range. So from
> this evidence I feel your change is a nice win.
>
> I previously said (in response to your question about how you
> should edit README.release):
>
>> I assume you will want to insert a paragraph concerning
>
> your change in a new section 2.8 of "Improvements relative to the
> previous release". Also, since you changed the semantics of the two
> functions (although not their API) for the RGB case you should
> probably specifically warn the users of this forthcoming release about
> that semantics change by adding a one-sentence new section 1.10 in
> "OFFICIAL NOTICES FOR USERS" that refers to your new section 2.8 for
> the details.
>
> Since your backwards-incompatible API change to the PLControlPt struct is
> going
> to cause build problems (just like it did in our Tk-related code)
> for the minority of our users who actually use that struct
> in their code, you should expand that suggested single sentence in 1.10 to
> warn users of this specific change to PLControlPt.
>
> I look forward to your commit with the suggested changes in README.release.
They'll arrive this afternoon all being well.
>
>
> Alan
> __________________________
> Alan W. Irwin
>
> Programming affiliations with the FreeEOS equation-of-state
> implementation for stellar interiors (freeeos.sf.net); the Time
> Ephemerides project (timeephem.sf.net); PLplot scientific plotting
> software package (plplot.sf.net); the libLASi project
> (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
> and the Linux Brochure Project (lbproject.sf.net).
> __________________________
>
> Linux-powered Science
> __________________________
|
