Index error when parsing message from some user.

[eivl]

During testing, I found this error.

Here are a few details.

I was able to get the message string from a previous error from the same type as message as the one in the screenshot above.

"@badge-info=founder/102;badges=moderator/1,founder/0;color=#8A2BE2;display-name=SheepFarmer;emotes=;first-msg=0;flags=;id=f9912e9d-474a-4cba-8897-0eb7bb9391e7;mod=1;reply-parent-display-name=hughmyron17;reply-parent-msg-body=!giveaway;reply-parent-msg-id=500ea2a8-f58f-407c-97a4-4c0e3274f374;reply-parent-user-id=133248066;reply-parent-user-login=hughmyron17;reply-thread-parent-msg-id=500ea2a8-f58f-407c-97a4-4c0e3274f374;reply-thread-parent-user-login=hughmyron17;returning-chatter=0;room-id=56649026;subscriber=1;tmi-sent-ts=1696780419533;turbo=0;user-id=69861108;user-type=mod :sheepfarmer!sheepfarmer@sheepfarmer.tmi.twitch.tv PRIVMSG #quin69 :@hughmyron17 For Quins $1,000 giveaway & more, enter here: http://bit.ly/quin69giveaway"

testable example:

>>> message = "@badge-info=founder/102;badges=moderator/1,founder/0;color=#8A2BE2;display-name=SheepFarmer;emotes=;firs
... t-msg=0;flags=;id=f9912e9d-474a-4cba-8897-0eb7bb9391e7;mod=1;reply-parent-display-name=hughmyron17;reply-parent-msg
... -body=!giveaway;reply-parent-msg-id=500ea2a8-f58f-407c-97a4-4c0e3274f374;reply-parent-user-id=133248066;reply-paren
... t-user-login=hughmyron17;reply-thread-parent-msg-id=500ea2a8-f58f-407c-97a4-4c0e3274f374;reply-thread-parent-user-l
... ogin=hughmyron17;returning-chatter=0;room-id=56649026;subscriber=1;tmi-sent-ts=1696780419533;turbo=0;user-id=698611
... 08;user-type=mod :sheepfarmer!sheepfarmer@sheepfarmer.tmi.twitch.tv PRIVMSG #quin69 :@hughmyron17 For Quins $1,000
... giveaway & more, enter here: http://bit.ly/quin69giveaway"
>>>
>>> name = message.split("PRIVMSG")[0].split("!")[0].split(":")[1]
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
IndexError: list index out of range

>>>

I don't know enough to talk about a solution, maybe indexing and splitting is not the right tool when the data is not predictable. and many selecting negative indexes would work. name = message.split("PRIVMSG")[0].split("!")[-2].split(":")[1] as an example for discussion.

Since the results are unclear, regex may be the right tool for this.

eivl

[eloonstra]

Thanks for reporting the error and contributing to the project, much appreciated. I'll release a fix as soon as possible.

[eivl]

My pleasure.

[eivl]

    badges = {badge.rstrip("/1") for badge in message.split("badges=")[1].split(";")[0].split(",")}
                                              ~~~~~~~~~~~~~~~~~~~~~~~~^^^
IndexError: list index out of range

another split issue.

i dont have the message that broke this. Adding it here.

I could try to rewrite the parser based on the RFC 1459 spec.

[eloonstra]

I'll be adjusting this as soon as possible. Apologies for the issues you're encountering.

[eivl]

no worries, im just glad i did not have to write my own socket and data parser from scratch :D

here is a message that failed in the parser.

m-id=56649026;subscriber=0;tmi-sent-ts=1696789151006;turbo=0;user-id=400074398;user-type= :faintsmlle!faintsmlle@faintsmlle.tmi.twitch.tv PRIVMSG #quin69 :guys stop saying shit about this being quins audio, this is much better

[eivl]

Im doing a full test to see if it breaks more places.

 @classmethod
    def parse(cls, message: str) -> "TwitchPrivateMessage":
        """
        Parse a Twitch private message.
        :param message: The message to parse.
        """
        try:
            badges = {badge.rstrip("/1") for badge in message.split("badges=")[1].split(";")[0].split(",")}
            color = message.split("color=")[1].split(";")[0]
            display_name = message.split("display-name=")[1].split(";")[0]
            first_message = message.split("emotes=")[1].split(";")[0] == "-1"
            message_id = message.split("id=")[1].split(";")[0]
            mod = message.split("mod=")[1].split(";")[0] == "1"
            broadcaster = "broadcaster" in badges
            returning_chatter = message.split("room-id=")[1].split(";")[0] == "1"
            room_id = int(message.split("room-id=")[1].split(";")[0])
            subscriber = message.split("subscriber=")[1].split(";")[0] == "1"
            turbo = message.split("turbo=")[1].split(";")[0] == "1"
            user_id = int(message.split("user-id=")[1].split(";")[0])
            name = message.split("PRIVMSG")[0].split("!")[-2].split(":")[1]
            channel = message.split("PRIVMSG")[1].split(":")[0].strip().lstrip("#")
            message = ":".join(message.split("PRIVMSG")[1].split(":")[1:]).strip()
        except IndexError:
            print()
            print(message)
            print()
            return cls(
                set(),
                "",
                "",
                False,
                "",
                False,
                False,
                False,
                0,
                False,
                False,
                0,
                "",
                "",
                "",
            )
        return cls(
            badges,
            color,
            display_name,
            first_message,
            message_id,
            mod,
            broadcaster,
            returning_chatter,
            room_id,
            subscriber,
            turbo,
            user_id,
            name,
            channel,
            message,
        )

I just wrap it in a try except and return an empty object back.

And here is another one that failed.

82a1-4d44-afc7-7acd1cbe1baa;mod=0;returning-chatter=0;room-id=56649026;subscriber=0;tmi-sent-ts=1696789889991;turbo=0;user-id=30278702;user-type= :anduinl!anduinl@anduinl.tmi.twitch.tv PRIVMSG #quin69 :Aware 2015

I ran this messages through a RFC 1459 tokenizer, and these lines still fails as it does not have the source. makes me think that twitch does not follow the IRC spec like I expect it to.

[eivl]

a97eba093e39423f9c4c2;color=#B22222;display-name=Khazerox;emotes=;first-msg=0;flags=0-3:P.3;id=9e48a829-3f4a-4bae-b9e2-7b30e9fc5acb;mod=0;returning-chatter=0;room-id=56649026;subscriber=1;tmi-sent-ts=1696790043418;turbo=0;user-id=58797167;user-type= :khazerox!khazerox@khazerox.tmi.twitch.tv PRIVMSG #quin69 :C***

and another one.

Should give you a few ideas how to troubleshoot I hope.