} else if(match(input, /^(\\[^"]|[^\\"])*[^\\]"/)) { # stuff"

            } else if(match(input, /^(\\.|[^\\"])*"/)) { # stuff"
                # sense of regex: zero or more of either (1) a
                # backslash-escaped character, don't care what, could
                # be a double quote; or (2) a character which is
                # neither a backslash nor a doublequote; followed by a
                # doublequote. if that sounds obvious, good; but in
                # previous revisions, this regex failed to recognize
                # the end of a string under certain conditions.
                

# input: double-quote, backslash, newline, backslash, letter n,
# double-quote. expected behavior: the double-quote begins a string.
# the backslash escapes the following real newline. we continue
# parsing, still inside the string, on the next input line. there is a
# backslash and letter n, denoting a newline character; and a
# double-quote, denoting the end of the string. the result of parsing
# is a string containing a single newline character. actual behavior:
# the double-quote after the backslash and letter n seems to somehow
# get escaped. the tokenizer thinks it is still inside a string
# literal.
lisp_eval_should_be_tricky_tricky 'IlwKXG4iCg==' 'IlxuIgo=' 'backslash-n-endquote at beginning of string continuation line'

lisp_eval_should_be_tricky_tricky 'IlwKXCIiCg==' 'IlwiIgo=' 'backslash-doublequote in string continuation line'

** Bugs fixed
*** Backslash escapes on a continuation line ending a string
A bug in a regular expression in the tokenizer made it fail to
recognize the end of a string literal if there were backslash-escaped
characters on that line, particularly right before the ending
double-quote character. This is now tested for and fixed, and the code
and tests are commented.