# This tries to create a virtual machine, a virtual machine # which will be particular suited for the task of parsing # and compiling. # # The machine consists or a stack, a tape, and a work area. # The stack is manipulated with "push" and "pop", the tape with # "get" and "put", and the workarea is affected by the # "clear" and "add" commands. # # Also, the machine includes a pointer which can be manipulated # explicitly with the "+1;" (etc) commands, and which is # affected implicitly with the push and pop commands. In the following # text, the machine implements itself in the c language. # # get - places the contents of the current item of the tape into # the work area. The current item of the tape is the item # to which the pointer is currently pointing # put - writes the contents of the work-area to the current item # of the tape. # push # pushes the contents of the work-area onto the stack. Also # the tape pointer is incremented by one. # pop # pops the last item of the stack into the work-area. The # tape pointer is decremented by one. # clear # deletes the contents of the work-area. # add # adds the specified text to the text contained in the work-area. # +n # increments the tape pointer by n places. # -n # decrements the tape pointer by n places. # print # prints the contents of the work-area to the standard output. # # The tape of the virtual machine is conceived as suitable # for containing the "attributes" of the tokens as the grammar is being # parsed. For example, the attribute of the "add" command is # the text which follows the command and which will be added to the # work-area. The stack of the virtual machine is designed to hold the # tokens as the are encountered in the lexified input stream. # This virtual machine is somewhat inspired by the stream-editor program # sed, which also can be viewed as a simple virtual machine. The # machine is designed to allow the implementation of certain types # of grammars and uses right-reduction parsing. # # The machine assumes that the input-stream has already been tokenised # or lexed. +1; pop; pop; /|Word|Semi-Colon/ { clear; get; +1; /pop/ { clear; add "pointer--;"; add "sWorkArea ="; add "aaTokenStack(pointer);"; } put; push; } pop; /|Word|String|Semi-Colon/ { clear; get; /add/ { clear; add "sWorkArea = sWorkArea +"; +1; get; -1; add ";"; put; clear; put; } /print/ { clear; +1; } } pop; /|Pattern|Left-Brace|Command-List|Right-Brace/ { clear; }