Re: [Plplot-devel] 5.14.0 release status

[Alan W. I. <Ala...@gm...>]

On 2018-12-10 19:04-0800 Alan W. Irwin wrote:

> Hi Phil:
>
> This is mostly addressed to you because it turns out one of your
> commits is causing release-critical trouble with the plsori
> call in example 3 as well as any attempt to use a non-zero -ori
> option.
[...]
> I used the commands
>
> make -j16 x01c ; make -j16 cairo
> examples/c/x01c -dev xcairo -ori 0.001
>
> to help git bisect the issue.
[...]
> @Phil:
> This commit (or its reverse) is not intrusive at all so it should not
> be that difficult to figure out why this change had such a
> catastrophic effect on plsdiori.

Hi Phil (again):

After some investigation I decided to make the
following *local* change to the master tip version.

diff --git a/src/plcore.c b/src/plcore.c
index 8d82b5b0f..33c84eeb7 100644
--- a/src/plcore.c
+++ b/src/plcore.c
@@ -1965,10 +1965,10 @@ calc_didev( void )

  // Set clip limits to conform to new page size

-    plP_sclp( (PLINT) ( plsc->didxax * plsc->phyxmi + plsc->didxb ),
-        (PLINT) ( plsc->didxax * plsc->phyxma + plsc->didxb ),
-        (PLINT) ( plsc->didyay * plsc->phyymi + plsc->didyb ),
-        (PLINT) ( plsc->didyay * plsc->phyyma + plsc->didyb ) );
+    plsc->diclpxmi = (PLINT) ( plsc->didxax * plsc->phyxmi + plsc->didxb );
+    plsc->diclpxma = (PLINT) ( plsc->didxax * plsc->phyxma + plsc->didxb );
+    plsc->diclpymi = (PLINT) ( plsc->didyay * plsc->phyymi + plsc->didyb );
+    plsc->diclpyma = (PLINT) ( plsc->didyay * plsc->phyyma + plsc->didyb );
  }

This change reverses one part of your commit that is obviously incorrect (since
plP_sclp does not set plsc->diclpxmi et al and instead sets struct
elements with almost the same name (clpxmi rather than diclpxmi, etc.))

My quick tests of this change shows it restores proper functioning
of -ori and example 3 while also having good rendering results for

examples/c/x33c -dev wxwidgets

as well as -dev xwin, -dev xcairo, and -dev qtwidget.

So I am hugely relieved to have found this solution for this
regression so quickly. However, I haven't pushed this change yet
because I haven't comprehensively tested it yet.

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
__________________________