cplusplus / mattermost Goto Github PK

Sometimes it is useful to refresh the paper index immediately which is why we should add a command for this (for example to account for recently published papers).

Paperbot does not reply to paper references that are not enclosed in square brackets (if paperbot has not been explicitly mentioned in that message). This confuses some people. However replying to all those messages could result in lots of spam.

A solution could be to send a DM once to the author of the message in such cases to inform them.

Paperbot did not respond yet the process was up and running and the healthcheck was good.

paperbot-paperbot-1 | websocket re-established connection
paperbot-paperbot-1 | missed websocket event, act_seq=0 exp_seq=266
paperbot-paperbot-1 | missed websocket event, act_seq=1 exp_seq=266
paperbot-paperbot-1 | websocket closed
paperbot-paperbot-1 | websocket re-established connection
paperbot-paperbot-1 | missed websocket event, act_seq=0 exp_seq=266
paperbot-paperbot-1 | missed websocket event, act_seq=1 exp_seq=266
paperbot-paperbot-1 | websocket closed
paperbot-paperbot-1 | websocket re-established connection
paperbot-paperbot-1 | missed websocket event, act_seq=0 exp_seq=266
paperbot-paperbot-1 | missed websocket event, act_seq=1 exp_seq=266
paperbot-paperbot-1 | websocket closed
paperbot-paperbot-1 | websocket re-established connection
paperbot-paperbot-1 | missed websocket event, act_seq=0 exp_seq=266
paperbot-paperbot-1 | missed websocket event, act_seq=1 exp_seq=266
paperbot-paperbot-1 | websocket closed
paperbot-paperbot-1 | websocket re-established connection
paperbot-paperbot-1 | missed websocket event, act_seq=0 exp_seq=266
paperbot-paperbot-1 | missed websocket event, act_seq=1 exp_seq=266
paperbot-paperbot-1 | websocket closed
paperbot-paperbot-1 | websocket re-established connection
paperbot-paperbot-1 | missed websocket event, act_seq=0 exp_seq=266
paperbot-paperbot-1 | missed websocket event, act_seq=1 exp_seq=266
paperbot-paperbot-1 | websocket closed

During meetings papers get uploaded to the papersystem often shortly before the presentation happens. In that case it might not be in the wg21.link index yet.
In case we have a lookup miss in our cache we could try a fallback lookup directly in the paper system to catch more of these cases.

Add hosting instructions to Github and the wiki:

https://github.com/cplusplus/mattermost/blob/main/README.md, but the WG21 wiki is another option (https://wiki.edg.com/bin/view/Wg21PersistentInformation/RealTimeCommunicationTools).

The IRC and Slack chat services previously used by some WG21 participants feature a bot service that responds to chats that contain references to papers, CWG issues, or LWG issues by replying with additional information about the mentioned items. For example:

Carmen Chair: Discussion of [P0990R0] is starting now. [P0989R0] is up next in about 30 minutes.
npaperbot: [P0990R0]: [EWG] Rebuttal of Implementation Concerns for Bit Entanglement (by Tony Van Eerd) (2018-04-01)
npaperbot: [P0989R0]: [EWG] Standardizing Extended Integers (by Tony Van Eerd) (2018-04-01)

This is a useful service that would be worth replicating in Mattermost.

If multiple papers are requested in the same message paperbot does not always keep the order of papers requested in the response.

For example:

This should be changed I think.

Right now the bot does not respond with information if a papernumber is added as an edit to a message. This is not very intuitive if you consider someone fixing a papernumber in the message. For example if someone accidently wrote:

[2552]

instead of:

[P2552]

We should consider reacting on changes messages if it does not create too much noise.