inbox.stream not returning None properly and instead heading into an unbreakable loop

Describe the Bug

When running a inbox.stream(pause_after=-1) I’d expect it to:

1. Fetch previous unread.
2. 1. Return None after is done.
3. 1. When next polled, if no data is received, should return None, but it does not, it instead starts decreasing limit from 100 to 75, then jumps to 100 and repeats the cycle

Desired Result

Desired result: For it to return None, rather than going down to 99, 98 etc

Relevant Logs

The very first request it does are like the below, but it uses before="t4_ANID", unfortunately it scrolls off my screen too quick for me to grab it.

DEBUG:prawcore:Response: 200 (8580 bytes)
DEBUG:prawcore:Fetching: GET https://oauth.reddit.com/message/unread/
DEBUG:prawcore:Data: None
DEBUG:prawcore:Params: {'before': None, 'mark': False, 'limit': 96, 'raw_json': 1}
DEBUG:urllib3.connectionpool:https://oauth.reddit.com:443 "GET /message/unread/?mark=False&limit=96&raw_json=1 HTTP/1.1"
 200 8580
DEBUG:prawcore:Response: 200 (8580 bytes)
DEBUG:prawcore:Fetching: GET https://oauth.reddit.com/message/unread/
DEBUG:prawcore:Data: None
DEBUG:prawcore:Params: {'before': None, 'mark': False, 'limit': 95, 'raw_json': 1}
DEBUG:urllib3.connectionpool:https://oauth.reddit.com:443 "GET /message/unread/?mark=False&limit=95&raw_json=1 HTTP/1.1"
 200 8580
DEBUG:prawcore:Response: 200 (8580 bytes)
DEBUG:prawcore:Fetching: GET https://oauth.reddit.com/message/unread/
DEBUG:prawcore:Data: None
DEBUG:prawcore:Params: {'before': None, 'mark': False, 'limit': 94, 'raw_json': 1}
DEBUG:urllib3.connectionpool:https://oauth.reddit.com:443 "GET /message/unread/?mark=False&limit=94&raw_json=1 HTTP/1.1"
 200 8580
DEBUG:prawcore:Response: 200 (8580 bytes)
DEBUG:prawcore:Fetching: GET https://oauth.reddit.com/message/unread/
DEBUG:prawcore:Data: None
DEBUG:prawcore:Params: {'before': None, 'mark': False, 'limit': 93, 'raw_json': 1}
DEBUG:urllib3.connectionpool:https://oauth.reddit.com:443 "GET /message/unread/?mark=False&limit=93&raw_json=1 HTTP/1.1"
 200 8580
DEBUG:prawcore:Response: 200 (8580 bytes)
DEBUG:prawcore:Fetching: GET https://oauth.reddit.com/message/unread/
DEBUG:prawcore:Data: None
DEBUG:prawcore:Params: {'before': None, 'mark': False, 'limit': 92, 'raw_json': 1}

Code to reproduce the bug

def run_inbox(ist=0):
    if(ist==0):
     ist=g.reddit.inbox.stream(pause_after=-1)
    for y in ist:
     if(y==None):
      break
     # do stuff
    return ist

My code example does not include the Reddit() initialization to prevent credential leakage.

Yes

This code has previously worked as intended.

Yes

Operating System/Environment

Running on windows 10.

Python Version

3.8

PRAW Version

latest on pip

Prawcore Version

latest on pip

Anything else?

No response

[NathanTech7713]

The problem is probably the creation of the stream inside the function, this only works well if the same stream object is reused.

Yep this fixed it. Sorry for the bother.

[LilSpazJoekp]

The problem is probably the creation of the stream inside the function, this only works well if the same stream object is reused.

What’s the logic behind starting at 100, dropping to 75, and then looping back to 100 just out of interest? What causes that to happen?

Incrementing of the limit parameter is to help break caches.