Floating point exceptions (floating point number too large/small, divide by zero, etc.) may occasionally be generated by user defined functions. Some of the demos in particular may cause numbers to exceed the floating point range. Whether the system ignores such exceptions (in which case gnuplot labels the corresponding point as undefined) or aborts gnuplot depends on the compiler/runtime environment.
The bessel functions do not work for complex arguments.
The gamma function does not work for complex arguments.
As of gnuplot version 3.7, all development has been done using ANSI C. With current operating system, compiler, and library releases, the OS specific bugs documented in release 3.5, now relegated to old_bugs, may no longer be relevant.
Bugs reported since the current release may be located via the official distribution site:
ftp://ftp.dartmouth.edu/pub/gnuplot
http://www.cs.dartmouth.edu/gnuplot_info.html
Please e-mail any bugs to bug-gnuplot@dartmouth.edu.