by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Alkor San
    @alkorsan
    yes
    easier to update too
    Vitaly Shukela
    @vi
    Implemented the --server and --binary options (after unrelated internal refactoring)
    Also now it defaults to --text, --line (this option has actually gone) and --linemode-retain-newlines (also gone).
    Alkor San
    @alkorsan
    wow I will test them all good thank you
    and come back with the results
    Vitaly Shukela
    @vi
    Now there is also --buffer-size and more lints (for advanced mode) and some fixes.
    Alkor San
    @alkorsan
    great options
    Vitaly Shukela
    @vi
    Now improving the documentation.
    Alkor San
    @alkorsan
    good a lot of things changed and strange things happening and 1000 question i have :/ the documentation will help
    Vitaly Shukela
    @vi
    It would be basically reformatted and fixed a bit existing documentation, not a Websocat Book.
    Alkor San
    @alkorsan
    ah ok
    please what does this error mean? or I open an issue for this?
    WARN 2018-07-03T19:20:09Z: websocat::broadcast_reuse_peer: A client's sink is NotReady for start_send
    I couldn t make things work with new updates
    Vitaly Shukela
    @vi
    It means client that is connected to a simple server is too slow for reading responses (or there is some other issue).
    Maybe just make simple server mode websocat -s 1234 accept only one simultaneous connection?
    Alkor San
    @alkorsan
    im not using websocket server but chrome server
    Vitaly Shukela
    @vi
    What is your use case for broadcast-reuse: overlay then?
    Alkor San
    @alkorsan
    it was working but now i removed -l and i got those errorq
    Vitaly Shukela
    @vi
    Anyway discussing the error without showing the full command line here is rather pointless.
    Alkor San
    @alkorsan
    I use broadcast-reuse: only in the exemple you gaves to me using nc
    ok
    I will update the same issue to continue there
    Vitaly Shukela
    @vi
    Pushed major README update. Please review.
    Alkor San
    @alkorsan
    ok thank you
    Vitaly Shukela
    @vi
    There are still remaining "specifier" term usages, but that ubiquitous as before.
    Your Chromium example is included as one of opener README examples.
    Alkor San
    @alkorsan
    good
    Vitaly Shukela
    @vi
    "Websocat v1.0.0 release" is near.
    (although there are known bugs and thigs to improve)
    Alkor San
    @alkorsan
    good
    I like the new doc :)
    Vitaly Shukela
    @vi
    There also new --queue-len option that helps to battle against A client's sink is NotReady.
    Alkor San
    @alkorsan
    oh good
    Vitaly Shukela
    @vi
    Also you can use reuse-raw: instead of reuse-broadcast: if you plan to have only one active incoming TCP connection at a time.
    Alkor San
    @alkorsan
    please now where is the -l ? is it integrated now ?
    Vitaly Shukela
    @vi
    -l is now the default in --text mode (which is also default now).
    Alkor San
    @alkorsan
    great , yes multiclient is an extreme case I think
    ah good to know
    Vitaly Shukela
    @vi
    There is --no-line and more radical --no-fixups instead.
    Note that in reuse-raw: mode it may stop reading from a WebSocket (which in turn makes Chromium stop writing to WebSocket) if there are no connection clients.
    In reuse-broadcast: mode messages simply got dropped when nobody is connected.
    Alkor San
    @alkorsan
    it seems safer to use then reuse-raw: , right?
    Vitaly Shukela
    @vi

    Also instead of this TCP-in-the-middle way maybe you can just use

    websocat cmd:./your_script.sh ws://...

    to drive the session?

    Alkor San
    @alkorsan
    ah yes this solution have a problem I couldn t solve:
    Vitaly Shukela
    @vi
    Depending on definition of "safer".
    Alkor San
    @alkorsan
    I will test both then and compare
    Vitaly Shukela
    @vi
    Even with reuse-raw: there expected to be lost messages, because of there is no good propagation of "can't write to this connection" status (which happens when somebody disconnects).
    Alkor San
    @alkorsan
    using sh-c: or cmd: I remember I could n t make for exemple the command timeout to work inside the subshell, that is why I prefered to use reuse-broadcast: , because I can use it in the same shell without limitation