(Please, next time, upload your gp script as a text file.)
Has nothing to do with multiplot. The problem is
set ytics font "Times-Roman, 6"
I say that font is not available on your computer, it neither is on mine. That seems to crash with the libgd terminals (at least on gp466/windows. In gp50 gd gives an error message about the missing font, without crashing.)
Last edit: Karl Ratzsch 2015-01-22
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Nothing changed between gp466 and gp500 with regard to font handling in the gd terminal. Could this instead be due to a different version of libgd used by the two gnuplot executables?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Most likely, and/or different compile options for libgd. My home-compiled version (mingw32, with libgd-2.1.0) simply ignores the nonexistent font, both in gp46csv and 50csv.
The official windows binary of gp46pl6 crashes, gp50pl0 shows warnings. I´m not sure which libgd version the 466 binary uses, the dll is called "libgd-2-733361a31aab" (?). 46pl4, which uses the old "bgd.dll" binary version of libgb, also ignores the wrong font.
I was the creator of the bug but posted it without my account xD
Thanks for the answers, as you said the times-roman wasn't on my libgd and well.. i changed it to arial and it worked!
The bad thing is that in png format it pixels a lot, i tried "emf" and its awesome, has a great resolution even zoomed in when the image is inserted in microsoft word.
Thanks!
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
This is expected, and should be in any FAQ or manual.
Some file formats are pixel-based (e.g. png gif jpeg some embedded *.eps images) and others are vector-based (e.g. PostScript pdf emf svg).
Vector-based formats can be zoomed with good results, pixel formats cannot. This is a property of the file format itself, and has nothing to do with gnuplot per se.
It is always better to choose a vector-based format when you need high resolution or scalable output.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
(Naturally, as the emf plot is a vector graphics.)
One thing puzzled me: The png terminal (gp50pl0/official windows binary) reacts differently to the font name "Times-Roman" than to a random nonexisting "Notafontname". The former throws a warning
"gdImageStringFT: fontconfig: Didn't find expected font family. Perhaps URW Type 1 fonts need installing? while printing string hereTeststring with font Times-Roman"
and prints nothing, while the latter prints a bunch of random symbols.
Existing windows fonts work ("Arial", "Times New Roman").
"Times-Roman" is an Adobe Type 1 font, whereas nearly everything else on a Windows machine is probably a TrueType or OpenType font. Both should work, but different font handling code may be involved.
Printing random symbols sounds like a bug. Are they really random or could it be an issue of the encoding?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
OK, so support for "Times-Roman" is compiled into but the font not distributed with the gp binary package? Which results in a crash or the warning with no output at all, depending on libgd version. That makes sense.
--
The symbols are indeed not random, they are reproducible characters from some symbol font. I haven´t been able to find out what font that is.
The only fonts included with gnuplot itself are the vectorized Hershey fonts used by the canvas terminal. Other than that, we depend on the arious support libraries and the operating system to find and render something reasonable in response to a requested font. In the case of the png terminal, the library involved is libgd, which looks for fonts in the search path defined by environmental variable GDFONTPATH. "Times-Roman" is one of the core PostScript fonts guaranteed to be present when PostScript-based Adobe products are installed. That doesn't mean there is any standard directory to look in, however. Adobe fonts use a non-standard "Adobe Custom" encoding.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am closing this as "out of date" since everything mentioned now has newer versions. If the problem is still present, please re-open or start a new report.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
(Please, next time, upload your gp script as a text file.)
Has nothing to do with multiplot. The problem is
I say that font is not available on your computer, it neither is on mine. That seems to crash with the libgd terminals (at least on gp466/windows. In gp50 gd gives an error message about the missing font, without crashing.)
Last edit: Karl Ratzsch 2015-01-22
Nothing changed between gp466 and gp500 with regard to font handling in the gd terminal. Could this instead be due to a different version of libgd used by the two gnuplot executables?
Most likely, and/or different compile options for libgd. My home-compiled version (mingw32, with libgd-2.1.0) simply ignores the nonexistent font, both in gp46csv and 50csv.
The official windows binary of gp46pl6 crashes, gp50pl0 shows warnings. I´m not sure which libgd version the 466 binary uses, the dll is called "libgd-2-733361a31aab" (?). 46pl4, which uses the old "bgd.dll" binary version of libgb, also ignores the wrong font.
(Btw.: Just saw that, as of last week, there´s a new 2.1.1 version of libgd, https://bitbucket.org/libgd/gd-libgd/downloads)
I was the creator of the bug but posted it without my account xD
Thanks for the answers, as you said the times-roman wasn't on my libgd and well.. i changed it to arial and it worked!
The bad thing is that in png format it pixels a lot, i tried "emf" and its awesome, has a great resolution even zoomed in when the image is inserted in microsoft word.
Thanks!
This is expected, and should be in any FAQ or manual.
Some file formats are pixel-based (e.g. png gif jpeg some embedded *.eps images) and others are vector-based (e.g. PostScript pdf emf svg).
Vector-based formats can be zoomed with good results, pixel formats cannot. This is a property of the file format itself, and has nothing to do with gnuplot per se.
It is always better to choose a vector-based format when you need high resolution or scalable output.
(Naturally, as the emf plot is a vector graphics.)
One thing puzzled me: The png terminal (gp50pl0/official windows binary) reacts differently to the font name "Times-Roman" than to a random nonexisting "Notafontname". The former throws a warning
and prints nothing, while the latter prints a bunch of random symbols.
Existing windows fonts work ("Arial", "Times New Roman").
Last edit: Karl Ratzsch 2015-01-23
"Times-Roman" is an Adobe Type 1 font, whereas nearly everything else on a Windows machine is probably a TrueType or OpenType font. Both should work, but different font handling code may be involved.
Printing random symbols sounds like a bug. Are they really random or could it be an issue of the encoding?
OK, so support for "Times-Roman" is compiled into but the font not distributed with the gp binary package? Which results in a crash or the warning with no output at all, depending on libgd version. That makes sense.
--
The symbols are indeed not random, they are reproducible characters from some symbol font. I haven´t been able to find out what font that is.
The only fonts included with gnuplot itself are the vectorized Hershey fonts used by the canvas terminal. Other than that, we depend on the arious support libraries and the operating system to find and render something reasonable in response to a requested font. In the case of the png terminal, the library involved is libgd, which looks for fonts in the search path defined by environmental variable GDFONTPATH. "Times-Roman" is one of the core PostScript fonts guaranteed to be present when PostScript-based Adobe products are installed. That doesn't mean there is any standard directory to look in, however. Adobe fonts use a non-standard "Adobe Custom" encoding.
I am closing this as "out of date" since everything mentioned now has newer versions. If the problem is still present, please re-open or start a new report.