optimize all_name_chars

[benjaminp]

Remove redundant PyUnicode_Check call. Use a static table for checking chars.

[tiran]

LGTM

[benjaminp]

For posterity, here's how I generated the table:

for i in range(128):
    if chr(i) in "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz":
        print "1,",
    else:
        print "0,",
    if i % 16 == 15:
        print

[tiran]

I take back my LGTM ... you seriously used Python 2 to generate the table? Shame on you! :)

[tiran]

for i in range(128):
    print('{:d}, '.format(chr(i) in "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz"), end='')
    if i % 16 == 15:
        print()

[benjaminp]

It's just nostalgia.

[serhiy-storchaka]

LGTM too. I have found a bug in all_name_chars(), but this is different issue.

[serhiy-storchaka]

Only one nit: please update the comment before all_name_chars().

[serhiy-storchaka]

You could use !Py_ISALNUM(*s) && *s != '_' and get rid of ok_name_char.