minlsys: a fast, tiny and unix-like derivation engine for L-Systems

This program is a tiny implementation of D0L L-System’s grammar derivation.

Features

• 83 lines of commented C89 code
• Read a L-System description on stdin and output the derivation on stdout
• The only heap allocation are done by the only standard library functions used:
  • correct input: scanf and putc (on glibc 2.2.5, 4,096B and 1,024B respectively)
  • wrong input: scanf and fputs (on glibc 2.2.5, 4,096B and 0B respectively)
• Fast! 76 millions of characters derived in 486ms (on i7-7500U @ 2.70GHz)

Limitations

• Authorized “characters” are all 256 1-Byte values (so all of ASCII and more)
• Axiom has a maximum size of 256 characters
• 256 rules of 256 characters maximum

File format:

First line: number of iterations (int scanned by scanf). If wrong, will exit with error code -1 and a message on stderr.

Second line: axiom.

Rest: 1 rule per line: first character is the predecessor, second character is ignored, rest of the line is the successor. Empty line are ignored

Example:

12
Z
Z F[-Y][+Y]FZ
Y Y[+X][-X]F[Y]
X F[+X][-F-X]F[+X][-X]

(Un)Robustness

Robustness:

• In compilation, no warnings or errors with -Wall -Wextra -Werror with gcc or clang.
• No crashes found, even with /dev/urandom input file.
• No memory leaks found, tested with valgrind.
• Number of iteration is checked to be an integer with scanf.

Unrobustness:

• If duplicated rule, only the last one is conserved.
• If the maximum size of 256 is not respected for the axiom or rules, the result derivation will not make any sense.
• Recursive implementation: stack may overflow for high value of iterations.