fruby

#!/usr/bin/ruby

# (c) 2008 B. Crowell, GPL v 2

# This is designed to be more or less a drop-in replacement for eruby. It supports a subset
# of eruby, and has better error handling.
# Usage:
#   fruby infile >outfile
# How it's different:
#   - Better error handling.
#   - Parses and evaluates the input file on the fly, rather than
#     translating it into a .rb file.
#   - Supports <% %>, <%= %>, and <%# %>, but not % at beginning of line. This makes
#     it compatible with TeX.
#   - Doesn't support this kind of thing:
#       <% 3.times do %> rah <% end %>
#       <% if foo %> foo <% end %>
#     Every block has to be a syntactically self-contained piece of ruby code.
#     Variables assigned to in one piece of code *are* available in later pieces of code.
#     This type of thing has to be accomplished by using a single block of ruby code,
#     with print statements inside it.
# Bugs:
#   - Doesn't check if <% and %> are properly matched.
# Security:
#   Will happily run any code in the file.

def die(message)
  $stderr.print "fruby: " + message + "\n"
  exit(-1)
end

# The following dummy subroutine is the environment in which we evaluate the code.
# see pickaxe book, pp. 296, 419
# This accomplishes two things for us:
# - If we change a variable in one <% %> block, the change persists in later blocks.
# - Local variables in <% %> blocks are separate from local variables in this program.
def binding_for_eval()
  return binding()
end

if ARGV.empty? then die("Usage: fruby infile >outfile") end
infile = ARGV[0]
if ! File.exist?(infile) then die("File #{infile} does not exist.") end
t = nil
File.open(infile,'r') { |f|
  t = f.gets(nil) # nil means read whole file
}

b = binding_for_eval()
inside = false # even if there is a ruby expression at the very beginning of the string, split() gives us a null string as our first string
line = 1
t.split(/(?:\<\%|\%\>)/).each { |x|
  if inside then
    what = 'execute'
    if x=~/\A=/ then what='evaluate' end
    if x=~/\A#/ then what='comment' end
    x.gsub!(/\A[=#]/,'')
    if what != 'comment' then
      y = eval(x,b,infile,line)
      if what == 'evaluate' then print y end
    end
  else
    print x
  end
  x.scan(/\n/) {line += 1}
  inside = !inside
}