New flag --save-all-html to record all html content.

[stevemac007]

Add logic to trap all GET and POST calls for debug purposes.

[max-rocket-internet]

Great idea 👍

[vide]

When trying this branch to debug my own problem related I guess to #179 I get

ERROR:root:write() argument must be str, not bytes
Traceback (most recent call last):
  File "/home/vide/tmp/aws-google-auth/aws_google_auth/__init__.py", line 79, in cli
    process_auth(args, config)
  File "/home/vide/tmp/aws-google-auth/aws_google_auth/__init__.py", line 243, in process_auth
    google_client.do_login()
  File "/home/vide/tmp/aws-google-auth/aws_google_auth/google.py", line 188, in do_login
    sess = self.get(self.login_url)
  File "/home/vide/tmp/aws-google-auth/aws_google_auth/google.py", line 123, in get
    out.write(response.text.encode('utf-8'))
TypeError: write() argument must be str, not bytes

but if I print before on stdout response.text I do see the HTML

[jack1902]

@vide if you drop .encode('utf-8') from these lines:

aws-google-auth/aws_google_auth/google.py

Line 100 in d6dfdb6

out.write(response.text.encode('utf-8'))

and

aws-google-auth/aws_google_auth/google.py

Line 122 in d6dfdb6

out.write(response.text.encode('utf-8'))

Does the error persist?

[vide]

@jack1902 yeah removing the encode('utf-8') works, I tried it before reporting but only on L122 and then I didn't see it was complaining about L100. Removing it in both places works as expected, I get the files saved to disk

[jack1902]

wondering if using the -l/--log level would be better to reduce the number of flags the tool has?

I tend to use --log debug --save-failure-html together when debugging. Wondering what others think?