#2882 6.0.5 segmentation fault on UBUNTU

[Anonymous]

just for the record:

I tried to install (according to the rules in INSTALL>gnu)
the new version 6.0.5
on UBUNTU 20.04.6 LTS

The installation works fine, but the
executable src/gnuplot
crashes with a segmentation fault right after start (first output)

gnuplot 6.0.4 works (and installs) fine
so Il keep that for the time being

(I tried to compile with -g, but could not make any sense of the gdb output)

[Ethan Merritt]

That is distressing.
Could you please first double-check that there were no error messages from the configure and make steps For example:

export CFLAGS="-Wall -g -Og"
./configure
make all >& make.log
grep WARN config.log
grep warn make.log
grep -i error make.log

Then use gdb to run the executable
Could you please attach a screen-shot or copy/paste the output of gdb after typing where?
I.e.

cd src
gdb ./gnuplot
(gdb) run
... As I understand it you get a crash immediately here.
... When that happens, type "where" and report everything that follows.
... Here is a sample; yours won't look exactly like this
(gdb) where
#0  0x00007ffff491c79d in read () at /lib64/libc.so.6
#1  0x00007ffff48aa5dc in __GI__IO_file_underflow () at /lib64/libc.so.6
#2  0x00007ffff48ab952 in _IO_default_uflow () at /lib64/libc.so.6
#3  0x00000000004a0818 in getc_wrapper (fp=0x7ffff49f1a80 <_IO_2_1_stdin_>) at readline.c:86
#4  0x00007ffff7f7814b in rl_read_key () at /lib64/libreadline.so.8
#5  0x00007ffff7f5b97f in readline_internal_char () at /lib64/libreadline.so.8
#6  0x00007ffff7f5c305 in readline () at /lib64/libreadline.so.8
#7  0x000000000042dca1 in rlgets (s=0x6926d0 "", n=2048, prompt=<optimized out>)
    at command.c:3683
#8  0x000000000042dd74 in gp_get_string
    (buffer=<optimized out>, len=<optimized out>, prompt=<optimized out>) at command.c:3854
#9  0x000000000042e8d8 in read_line
    (prompt=prompt@entry=0x546ab4 "gnuplot> ", start=start@entry=0) at command.c:3893
#10 0x00000000004325f7 in com_line () at command.c:421
#11 0x000000000048be94 in main (argc_orig=<optimized out>, argv=0x7fffffffdd28) at plot.c:654

[PuQ]

I got a segfault when running src/gnuplot -d <demo/mask_pm3d.dem, perhaps related to this bug.

Reduced repro:

gnuplot> set table $HULL
gnuplot> unset table
gnuplot> plot $HULL
AddressSanitizer:DEADLYSIGNAL
=================================================================
ERROR: AddressSanitizer: SEGV on unknown address 0x000000000000
The signal is caused by a READ memory access.
Hint: address points to the zero page.
    #0 0x000001103adc in df_gets src/datafile.c:633
    #1 0x0000010fa82e in df_readascii src/datafile.c:1936
    #2 0x0000010f59f6 in df_readline src/datafile.c:1891
    #3 0x0000012e5213 in get_data src/plot2d.c:723
    #4 0x0000012d6e79 in eval_plots src/plot2d.c:3349
    #5 0x0000012cb789 in plotrequest src/plot2d.c:345
    #6 0x0000010d3720 in plot_command src/command.c:2262
    #7 0x0000010ce3b1 in command src/command.c:863
    #8 0x0000010cd951 in step_through_line src/command.c:546
    #9 0x0000010ccf1f in do_line src/command.c:443
    #10 0x0000010cca38 in com_line src/command.c:431
    #11 0x0000010bbb29 in main src/plot.c:683
SUMMARY: AddressSanitizer: SEGV src/datafile.c:633 in df_gets

Broken by [7bbf83].

[Ethan Merritt]

I am puzzled. Your minimal reproducer is a real bug (trying to plot from an empty data block). I pushed a fix to both 6.0 and 6.1

But that pattern does not occur in mask_pm3d.dem. I cannot reproduce failure with that demo either before or after the fix.

Edit: OK, I see that the problem could occur with that demo if the loadpath, or GNUPLOT_LIB, is set incorrectly. Then it would try to create a data block from a file not in the search path, resulting in an empty data block.