Test which unicode characters survive in spidermonkey

test_js.py

# Run me from the root of mozilla/app-validator (requires updates from pull #27).

import json
import re
import subprocess
import sys
from tempfile import NamedTemporaryFile

from appvalidator import unicodehelper

# Flip me to see if it works.
ESCAPE_JS = False


def make_char(i):
    return eval("u'\\u{code:0>4}'".format(code=i))


def encode_char(char):
    return json.dumps(unicodehelper.decode(char, js_safe=ESCAPE_JS))

START = 0
END = 10000

with NamedTemporaryFile() as f:
    for i in xrange(START, END):
        char = make_char(i)
        f.write(char.encode('utf-8'))
        f.write('\n')
    f.flush()
    chars = open(f.name).read().splitlines()

print "There are {l} chars".format(l=len(chars))

js_script = """
var line = readline();
while (line !== null) {
    print(JSON.stringify(JSON.parse(line)));
    line = readline();
}"""
encoded_chars = [encode_char(char) for char in chars]
failed = []

cmd = ['/usr/local/bin/js', "-e", js_script]
shell_obj = subprocess.Popen(
    cmd, shell=False, stdin=subprocess.PIPE, stderr=subprocess.PIPE,
    stdout=subprocess.PIPE)
data, stderr = shell_obj.communicate('\n'.join(encoded_chars))

for i, out_char in enumerate(data.splitlines()):
    char = chars[i]
    try:
        json.loads(out_char)
        passed = True
    except:
        passed = False
        failed.append(char)
    if not passed:
        print(u'char {i} {passed} ({char!r})'.format(
            i=i, char=char, passed='passed' if passed else 'failed'))

if failed:
    print(failed)
    print('{l} failed'.format(l=len(failed)))
else:
    print('all passed')