I did not see this message, grrr... Here's the stack trace: #0 syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38 #1 0x00007f60d833ea84 in g_cond_wait_impl (cond=0x55a694e3ab80, mutex=0x55a694e3ab78) at ../../../glib/gthread-posix.c:1007 #2 g_cond_wait (cond=cond@entry=0x55a694e3ab80, mutex=mutex@entry=0x55a694e3ab78) at ../../../glib/gthread.c:1665 #3 0x00007f60d77a42e3 in pango_fc_patterns_get_font_pattern (pats=0x55a694e3ab70, i=0, prepare=<synthetic pointer>) at ../pango/pangofc-fontmap.c:1136...
maxima.el triggers a warning about deprecated positional arguments to define-minor-mode
So, a simple command to reproduce the issue: echo 'set terminal wxt; plot x' | gnuplot -persist
This problem actually appears even without FVWM's ManualPlacement, i.e. if the window is displayed immediately. So I suppose that it is likely to occur with other window managers.
I've identified the "problematic" commit in Pango: https://gitlab.gnome.org/GNOME/pango/-/commit/89442dae443eba2aa0f0a526b4d6d39c0c9b13c6 The commit message is Just use a single fontconfig thread In ac4d8f2964299d14046cb, we switched from using GTask to using explicit threading for our fontconfig calls, since GTask assumes a mainloop and leaks memory otherwise. But now we are creating a new thread for every single fontconfig call, which is a bit excessive. As a compromise, create a single fontconfig...
Actually, the problem also occurs with gnuplot 5. It has actually appeared with the 1.52 pango library: downgrading the pango packages to 1.51 makes the problem disappear, both with gnuplot 5 and 6. Now, I don't know whether this is a bug in pango or a bug in gnuplot that is triggered only with the new version of pango (for instance, it might be possible that pango 1.52 is faster than 1.51, making the problem appear, since it seems to be a race condition).
With wxt, "gnuplot -persist" hangs if gnuplot cannot draw before stdin is closed
I can't see any reference to doc2web in gnuplot 5.4.4. Do you mean that it will be in the next gnuplot release? (This is actually a way to make the issue disappear.)
Meaningless crossref generation for the HTML manual
The Info manual contains a spurious Texinfo comment "@c <3 [...]"
FYI, I use the "info" format preferably because I can use it in a text terminal and do a full search on it (which is not possible with split HTML files). And PDF is only very readable on a big 4K screen (or on paper, which has its own issues).
The Debian bug, with the details about this issue: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=989029
With a separate build directory, the list of terminals is missing in the info manual
The broken navigation is actually caused by a different bug in doc2texi.el; see bug #2575 I've just reported. This one, about the incorrect relative path, is just about the missing list of terminals, and the corresponding Debian bug is finally: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=989028
Spurious menu at the end of Section 4 of gnuplot.info breaks navigation in info
with a separate build directory, the info manual is not built correctly
Well, after looking at the issue more closely, I think that what I really need is "pause mouse close" as suggested. This works as long as there are no temporary files. For the issue with my Perl script and temporary files, I think that the solution is that the script should wait until gnuplot terminates, so that the temporary files are not removed until then (if need be, the script could optionally fork before running gnuplot so that it could put itself into background just before waiting for gnuplot...
No, terminals like GNOME Terminal, Konsole and the Xfce terminal are text terminals, and the ioctl will work.
Under Unix, environment variables are not shared among processes, they are just inherited by new processes. So, if the parent changes the value of an environment variable, the child won't see it. The size of the terminal can be obtained at any time by using a system call (ioctl(..., TIOCGWINSZ, ...)) and languages like LISP may provide an interface for that (for a shell, one of the interfaces is the COLUMNS shell variable, which is not even an environment variable). It is true that listening to the...
There is no COLUMNS environment variable in my shell (zsh); this is just a shell variable (despite the uppercase letters). The reason is that its value becomes incorrect whenever the window is resized (zsh updates it, but this would not be the case of the child processes in general), thus it must not be present in the environment. I don't know whether this can be used, but I've found this: Get the size of the terminal from Common Lisp.
Maxima should honor the number of columns provided by the terminal
Thanks. Well, this is more or less what I was looking for. But perhaps I'm asking too much. The (minor) issue with this solution is that it blocks the script until gnuplot terminates. Anyway it seems that I do not have the choice: I'm using gnuplot to plot data from temporary files (generated by the script and removed when the script terminates), but gnuplot neither saves the data in memory nor keeps the files open, so that if these files are removed (because the script terminates) while gnuplot...
option so that gnuplot runs in background instead of exiting, until the persistent window is closed
Yes, I later saw that there were other problems than just the paths. If you say that...
docs/Makefile.am mixes up srcdir and builddir: gnuplot executable not found when they are different
Using the upstream gnuplot.pdf may not be a good idea because one may want to regenerate...
please no longer depend on the non-free picins package to build the PS/PDF doc with images
Also, note the full rule about gnuplot.info currently in docs/Makefile.am: gnuplot.info:...
Currently on http://gnuplot.cvs.sourceforge.net/viewvc/gnuplot/gnuplot/docs/Makefile.am?view=markup...
The rule was already written by a gnuplot developer in docs/Makefile.am. You just...
This is working for you only because you use the same build dir as srcdir. Try to...
No, I have both of these tools (gnuplot.texi has successfully been created and I...
Indeed, the rule seems to take gnuplot.texi from the source directory instead of...
"make info" fails
I thought this had been tested. BTW, wxwidgets3.0 also yields problems with xmlcopyeditor....
Grrr... SourceForge (and Firefox) lost again my reply! OK. This is rather unintuitive:...
zooming doesn't work with the 'qt' terminal
Sorry, because I didn't know that there were 2 different issues, this comment was...
After further tests, there actually seem to be 2 issues, one with -persist (which...
Thanks. The patch against 4.6 works. And of course, no need to use -persist.
See https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=750045#53
The zoom buttons are also affected. With wxt and -persist, gnuplot is put in the...
Concerning the documentation, the --persist description in the gnuplot(1) man page...
There are several problems with the "pause mouse key" solution. First, the "show...
The behavior is not satisfactory since in non-interactive mode (as above), gnuplot...
The problem only occurs when -persist is used and the main gnuplot program has terminated,...
The problem only occurs when -persist is used and the main gnuplot program has terminated,...
I don't know, the code has significantly changed. There's no !parent->isVisible()...
Well, the initial size is approximately 243×113, so that these "width=243 height=113"...
gnuplot opens the window with some size, then resizes it to its final size, leading...
zooming doesn't work with the 'qt' terminal
The "show grid" button doesn't work with the 'qt' terminal
I forgot to give the reference of the Debian bug: https://bugs.debian.org/cgi-bi...
With the 'qt' terminal, the initial window size is incorrect