jbrowneuk / iris-bot Goto Github PK

Using +hm guess <wrong word> shows the game embed as opposed to a message saying the guess was incorrect.

The package we use to query server statuses is deprecated and should be replaced as per the documentation at https://www.npmjs.com/package/minecraft-server-util

By moving to the new API we can reduce the complexity of the Minecraft server status plugin.

Each command states it needs a URL which is incorrect.

Response generator: if a phrase does not have a value for a specific mood, the bot will respond with the error response "Nothing to say..."

example:

1. Bot mood +100 - happy
2. Phrase: genericPhrase:

When this phrase is requested, according to the documentation the fallback none responses (Response1, Response2, Response3) should be chosen.

Actual behaviour: the bot will only respond with "Nothing to say..."
Expected behaviour: the bot will respond with Response1, Response2 or Response3

Request from server:

would also be fun if she (skye) could tell jokes. like we could ask for a joke and she would say one

If a long-running fetch task occurs (i.e. getting an image for a command) and more messages arrive before the API response is received, the cached "last message" is updated and the channel the response is sent to will change.

i.e.:
Channel 1 userA: +get-slow-api-response
*bot begins data request
Channel 2: userB: anything
*API returns results, bot processes
Channel 2: bot: api response

This means the bot can potentially post a response to a request in a different server.

The blog roll feature that polls for journal updates currently has an interval of 5 minutes. This timespan should be configurable to not DDOS the journal API.

It also has verbose logging which should be tweaked to prevent event log spam when running on production environments.

Make the streak, wins and losses fields inline:

Embed should show a visual representation of the game state.

Consider the images on Wikipedia.

The hangman game embeds read like so:

--A- (4 letters)
10 chancess left

The chancess text is double pluralised despite claiming to fix it in #90

Each personality plug-in persists its settings in a different way. The bot’s settings API should be extended to allow for consumers to persist their settings in a consistent manner to a central store.

1. Check out the latest tag (v0.4.5)
2. npm install, npm test - all pass
3. Run npm update
4. npm test - no specs found

If the call/response table has any addressed calls, it is possible to invoke that call without addressing the bot.

Consider the Call/Response pair:

The intent is that you can address the bot by sending the message botname prompt to which it will respond with My response.

However, it is also possible to get the same response by sending the text {£me} prompt. This is not intended behaviour and so should be fixed.