regression in CONTOUR, /iso

[alaingdl]

A regression appeared in CONTOUR between Feb 16 & 19

CONTOUR, dist(50,25),/iso

To avoid regressions of such kind, I suggest : 1/ we generalize what I did for the Coyote test suite (having references, generating outputs in PNG using TVRD(), generating a PDF through LaTeX, giving a look to the PDF) 2/ working in branch(es) and testing in branch before merging to avoid to put regression(s) in the main code (it is difficult for me to recommend to colleagues to use the Git Version if the quality of the Git version is oscillating too much)

[GillesDuvert]

works for me, apparently. More details?

[alaingdl]

old (as is 2023-02-15)

new (as is today 2023-03-10)

I have the same result on three different machines , all compilations done with the script.

And sorry, maybe I did not checked enough the corrections for #1495 but we need to improve the way to check consequences of changes in graphical outputs --> production of list of figures + processing of PDF

[GillesDuvert]

Still, worksforme, I do not have this. However, I cannot find the PR with the patch https://github.com/gnudatalanguage/gdl/commit/539100cbef88e371a2c55076dec3e8667749255a that solve it indicated in #1495 Wonder if this patch has really been merged

[alaingdl]

re-tested, same problem, it is broken for Y between 17 and 26 but ok below 16 (included) and above 27 (incl.)

CONTOUR, dist(50,y),/iso

The example KO in #1495 is now OK for me

[alaingdl]

Another example which was OK in January 2022 but is not OK now :

GDL> window, xs=200, ysize=400
GDL> plot, findgen(10),/iso

(just for the record, no problem with the title, if any)

[alaingdl]

More informations :

• In fact the behavior on 2 different computers is different
• !d seems to not be upgrade if the window is resized (by mouse). It should.
• even simpler way to bad plot :

GDL> window, xs=200, ysize=400
GDL> plot, findgen(10),/iso
GDL> plot, findgen(10)^2,/iso

[GillesDuvert]

indeed.

[GillesDuvert]

solved. quick test: window,0, xs=400, ysize=200 & plot, findgen(10),/iso & window,1, xs=400, ysize=200 & plot, findgen(10)^2,/iso &window,2, xs=200, ysize=400 & plot, findgen(10),/iso & window,3, xs=200, ysize=400 & plot, findgen(10)^2,/iso

[alaingdl]

Thanks Gilles, no more energy today to test, but tomorrow ...