|
From: sfeam <sf...@us...> - 2018-05-31 01:47:07
|
I do not understand the root cause of the problem, but I can give you an easy fix for your script. The badness all comes from inclusion of that final line: plot ... 1/0 lt -1 notitle # Dummy so I can move above around If you delete that line, it works. I understand why you had that line there, but it has not been needed for a very long time. Since at least version 4.6 you can end a plot command with a comma and at worst you will get a warning message. Why does including an all-undefine plot cause the logscale to fail? Good question. I'll look into it. But meanwhile you should be able to get your log output again. Ethan On Thursday, 31 May 2018 08:00:41 Dave Horsfall wrote: > ---------- Forwarded message ---------- > Date: Wed, 30 May 2018 16:56:28 +1000 (EST) > From: Dave Horsfall <da...@ho...> > To: Gnuplot Mailing List <gnu...@li...> > Subject: I think something in Gnuplot broke... > > (Redirected here from the MacPorts list, which just upgraded to a new > release.) > > Something suddenly busted here; my spam graph is utterly weird now, and I was > wondering whether it was anything that I did... > > See www.horsfall.org/spamlog.pdf for yesterday's graph (nice and pretty!) and > www.horsfall.org/spamlog-busted.pdf (where the hell did my log scale on the > left go?). > > Command and trimmed datafile attached... > > I noticed that Gnuplot itself went from: > > Version 5.2 patchlevel 2 last modified 2017-11-15 > > to: > > Version 5.2 patchlevel 3 last modified 2018-05-04 > > It was posited that not many people would run mixed log/lin y-axes so soon > after the upgrade, but I do it every day... > > Thanks. > > -- Dave |