octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Help backing out a regression to Gnuplot on stable


From: Ben Abbott
Subject: Re: Help backing out a regression to Gnuplot on stable
Date: Tue, 15 Dec 2015 11:29:15 -0500

> On Dec 15, 2015, at 9:44 AM, Ben Abbott <address@hidden> wrote:
> 
> A regression to print() occurred as a result of a proposed fix for bug 46122.
> 
> https://savannah.gnu.org/bugs/index.php?46122
> 
> The changeset is af5591ef9790
> 
> http://hg.savannah.gnu.org/hgweb/octave/rev/af5591ef9790
> 
> The original bug report was’t a problem with Octave, but with the clumsy way 
> which Mathworks setup the Paper* properties and hence the common confusion 
> which follows. A second bug report covering the same issue for OpenGL was 
> also reported, #46292.
> 
> https://savannah.gnu.org/bugs/?46292
> 
> I’ve responded on both reports and provide syntax to produce the desired 
> result.
> 
> In any event, this change will need to be backedout. I assume it is desired 
> to use “hg backout -r af5591ef9790” and some other modifiers to property 
> merge everything. This change should also be merged into the default branch. 
> I’m not familiar with using “backout” or with the process of merging a change 
> from stable to default.
> 
> John/Rik, is this something one of you can quickly take care of for me.
> 
> I have additional changesets for Gnuplot to follow. As far as I can tell, 
> they will all be pushed to default.
> 
> Ben
> 
> p.s. I'm using web-mail to get around a firewall which blocks email. 
> Apologies if this isn't plain text.

Is pushing a changeset like below acceptable?

—————————————
# HG changeset patch
# User Ben Abbott  <address@hidden>
# Date 1450192520 18000
#      Tue Dec 15 10:15:20 2015 -0500
# Branch stable
# Node ID 698e91605f17929f2edcaeeab7be93b2ad268645
# Parent  af5591ef9790fd0466c11c8fd4fcf7ec5c86a4e7
Backout rev af5591ef9790.

* scripts/plot/util/__gnuplot_drawnow__.m: When printing, the paper properties
are modified to size all objects to 1 pixel/point. This is needed to support
the design of GL2PS. This is also done for Gnuplot to avoid toolkit specific
code in print.m and __print_parse_opts__.m. Toolkit specific code shoudl be
limited to toolkit specific m-files, i.e.__opengl_print__.m, and
__gnuplot_print__.m. The changeset being backedout is associated with bug
report #46122.

diff --git a/scripts/plot/util/__gnuplot_drawnow__.m 
b/scripts/plot/util/__gnuplot_drawnow__.m
--- a/scripts/plot/util/__gnuplot_drawnow__.m
+++ b/scripts/plot/util/__gnuplot_drawnow__.m
@@ -185,8 +185,8 @@ function enhanced = gnuplot_set_term (pl
                || any (strcmp (term, {"canvas", "emf", "gif", "jpeg", ...
                                       "pbm", "png", "pngcairo", "svg"}))))
           ## Convert to inches
-          gnuplot_pos = gnuplot_pos / get (0, "screenpixelsperinch");
-          gnuplot_size = gnuplot_size / get (0, "screenpixelsperinch");
+          gnuplot_pos = gnuplot_pos / 72;
+          gnuplot_size = gnuplot_size / 72;
         endif
         if (all (gnuplot_size > 0))
           terminals_with_size = {"canvas", "emf", "epslatex", "fig", …
—————————————

Meaning is this equivalent to ...

hg backout --merge --logfile ChangeLog.txt --rev af5591ef9790

If so I can push this change, but still request help with merging to default.

Ben


reply via email to

[Prev in Thread] Current Thread [Next in Thread]