Skip to content

Message input displaying twice despite only being sent once in desktop app  #466

Closed

Description

@laurenreidexpensify

Problem -

Message input displaying twice despite only being sent once in desktop app

My view on Version 1.0.1-46 (1.0.1-46) of DM between @laurenreidexpensify and @FlaviaFaria -

2020-09-11_15-02-55

My view on Version 1.0.1-46 (1.0.1-46) of Deprecate-slack of @abekkala's message:

image

@april's view deprecate-slack -

image

Activity

  1. abekkala commented on Sep 11, 2020

    @abekkala
    Contributor

    I'll note here as well - I'm not using Desktop I'm using Web. This is noted in my original post in this GH.

  2. mallenexpensify commented on Sep 11, 2020

    @mallenexpensify
    Contributor

    Thanks @laurenreidexpensify and @abekkala , adding this to milestone! Labeling HIGH because it's happening to multiple people (I've seen it too).

  3. tgolen commented on Sep 11, 2020

    @tgolen
    Contributor

    I assume that refreshing makes this go away?

  4. mallenexpensify commented on Sep 11, 2020

    @mallenexpensify
    Contributor

    For me, in the past, it has. And I confirmed earlier this week that the person on the other side doesn't see duplicate messages, just the single one

  5. sketchydroide commented on Sep 15, 2020

    @sketchydroide
    Contributor

    I'll take a look into this one as well, as I might have some insight into it, unless you want to take it @tgolen

  6. tgolen commented on Sep 15, 2020

    @tgolen
    Contributor

    I'm willing to be it's a sequenceNumber bug, which we know won't be perfect until we are using something like ROWID.

  7. tgolen commented on Sep 15, 2020

    @tgolen
    Contributor

    Thanks for looking into it!

  8. sketchydroide commented on Sep 16, 2020

    @sketchydroide
    Contributor

    I've been going trough the code both on Web and the Chat app, and I don't think we can actually fix this without some sort of unique ascending key for each action. And some changes to both the Api and the app. Like Tim mentioned some sort of ROWID.
    The problem is that we can't guarantee that the sequenceNumber we create optimistically is the same the API will assign the comment. Most of the times it will be. But when it's not we can't check if it's a repeated message or the user just entering the same message several times (I don't expect that to happen often but it can).

    Using the timestamp of the comment when it's created on the app, and then using to check if the comment is repeated might be a solution. By passing it along to the api and then getting it back when the API finishes adding the comment. We could then use it to filter out any comment that has the same timestamp.

    What are your thoughts @tgolen?

  9. tgolen commented on Sep 16, 2020

    @tgolen
    Contributor

    @cead22 was working on a solution to utilize ROWID instead of relying on sequence number, so he is probably the best to chime in on this.

    I personally don't think we should do any more hacks for this until we have the real solution in place, and it will just be known that sometimes, there might be small bugs like this.

  10. cead22 commented on Sep 16, 2020

    @cead22
    Contributor
  11. mallenexpensify commented on Sep 16, 2020

    @mallenexpensify
    Contributor

    I have a feeling this is related, let me know if not and I'll create a GH.
    I was doing some testing with Rafe and he asked that I send multiple chats, a second or two apart. I sent them in numbered order but in the report they show as
    image

  12. tgolen commented on Sep 17, 2020

    @tgolen
    Contributor

    @mallenexpensify there is something a little strange about that screenshot... Why are there relative timestamps? They should look like this:

    image

    That would indicate to me that you've got a super old version of the chat app that you're using.

  13. tgolen commented on Sep 17, 2020

    @tgolen
    Contributor

    oh oh, sorry... I see. That's a screenshot from the report page on expensify.com.

  14. mallenexpensify commented on Sep 17, 2020

    @mallenexpensify
    Contributor

    ah, yeah, I was double checking to see if it showed differently (ie. in the correct numbered order) on the reports page and it didn't, it matches exactly

    image

  15. sketchydroide commented on Sep 18, 2020

    @sketchydroide
    Contributor

    Ok I'm unasigning myself from this, as I don't think there is actually anything actionably at the moment and we are using this more to track the issue, and related issues, and that when https://github.com/Expensify/Expensify/issues/141208 is done we can then do something about this.

  16. removed their assignment
    on Sep 18, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions