[Exhentai] Html page is downloaded instead of the last image before hitting image limit

[Fukitsu]

I know it was fixed in #1437 but it started happening to me again, not sure since when but I suspect it started since I got the "More Pages" hath perks which raised the image limit up to 50000 instead of the usual 5000 though that doesn't make much sense. It happens with big galleries like this one for example:
https://exhentai.org/g/1966459/205507c4ed/

The image number 455 is downloaded fine but the number 456 is an html page. Does this happen to anyone else? Is there a way to make gallery-dl check if the last image is an html and give the link to the previous image (second to last), in this case number 456 instead of number 457?

[Twi-Hard]

My limit is also 50,000 but I'm not having that issue.
These are my settings:

"exhentai": {
      "domain": "auto",
      "limits": 50000,
      "metadata": true,
      "original": true,
      "wait-min": 6.0,
      "wait-max": 10.0,
      "sleep": 3,
      "sleep-request": 2,
      "sleep-extractor": 3,
      "username": null,
      "password": null,
      "skip": "abort",
      "cookies": {
        "ipb_member_id": REDACTED,
        "ipb_pass_hash": REDACTED,
        "hath_perks": REDACTED,
        "ipb_session_id": REDACTED,
        "sk": REDACTED,
        "star": REDACTED,
        "event": REDACTED
      },
      "base-directory": REDACTED,
      "directory": [
        "{eh_category}",
        "{gid} - {title}"
      ],
      "filename": "{num:>04}_{filename}.{extension!l}"
    },

If something's different in your config it might help identify the issue?

Edit: This is my version
1.18.4-dev

[Fukitsu]

I added "limits": 50000 and that seems to have fixed the issue, thank you

[mikf]

Do you still have that HTML file lying around? Does it contain the following text?

You have exceeded your image viewing limits. Note that you can reset these limits by going <a href="https://exhentai.org/home.php">here</a>.

When limits is not a number, gallery-dl checked for the image limit by comparing the Content-Length header to a specific value, which would be different when the error message is not the same for everyone. I've now changed it to only check for a text/html Content-Type 5785462, but I'd still like to know why the previous method didn't work for you.

[Fukitsu]

Yes, it's that same message. It's weird since the previous method had been working fine until I raised my image limits