Regex start of string replace seems to be repeatedly applied

[Neradoc]

In MicroPython on ESP32 (v1.18 release and v1.18-202-gade2720e5).
Start-of-string ^a regex in sub acts like it's applied repeatedly in some way.

MicroPython v1.18 on 2022-03-09; ESP32 module with ESP32
Type "help()" for more information.
>>> import re
>>> re.sub('^ab', "0", "abababcabab")
'000cabab'

In C Python we get the expected result:

>>> import re
>>> re.sub('^ab', "0", "abababcabab")
'0ababcabab'

[dpgeorge]

Thanks for the report, I can confirm the issue. The problem is that the sub() function keeps resetting (after a substitute) the start of the string that its scanning, so the ^ keeps matching.

A slightly more complicated example of the issue, showing that it's not simple to solve:

>>> re.sub('(^ab|cab)', '0', 'abababcabab')
'00000'

[dpgeorge]

Also similar issue for split():

print(re.compile('^ab').split('abababcabab'))
['', '', '', 'cabab']

[dpgeorge]

I found a relatively simple fix; see #8403.

[dpgeorge]

Fixed by 63f0e70