[feature request] qlim doubt (this one should be easy!)


    Feb 25 2015 | 1:29 pm
    Hi. Sorry for the basic question.
    I have a high speed metro connected to a qlim. Imagine that the qlim is currently configured for 10000 ms If I change the value to 20 ms now, I have to wait for the next cycle of 10000 ms for qlim to update the timing peridiocity; only then will qlim start outputing at a 20 ms interval. I guess it was intentionally done this way, but I also think there should be an attribute to change it instantly, or a kind of reset message to force the update.
    Is there something I'm overlooking? Or an alternative?
    Thanks.

    • Feb 25 2015 | 2:11 pm
      you could brute force it and script a deletion/recreation each time you change the interval... other than that i don't see :s
    • Feb 25 2015 | 2:15 pm
      other than that, why not changing the qmetro interval rather than qlimiting its output ? because then you could send a restart each time you change qmetro interval output like this
    • Feb 25 2015 | 3:45 pm
      Thanks for the response, Vichug. There are "subtimings" for different parts of the patch. I had several qmetros and was trying to get away from that solution with qlim. It works well, with the mentioned exception which is annoying, particularly when using large values...
    • Feb 25 2015 | 5:22 pm
      i see... well without something like a "reset" message to qlim, this seems indeed difficult...
    • Feb 25 2015 | 6:42 pm
      hi, not elegant at all, but sometimes i use this trick to do the job :)
    • Feb 26 2015 | 9:28 am
      Hi, again. Thanks for your solution, Matmat. I think I've never used onebang before...
      Cycling: you might consider this a feature request for qlim: if it behaves this way intentionally, consider at least a "force timing update now" feature. It is particularly important when the qlim is set to a very high number...
      Thanks guys.
    • Feb 26 2015 | 10:49 am
      I agree that the new timing limit should take effect immediately in qlim, or that there should at least be a message to clear it. In the meantime, here's a workaround that should work for most messages.
    • Feb 26 2015 | 11:47 am
      Thanks, Cristopher. For the time being, as my different timings don't really have to be in exact sync, I think I'll use several qmetros instead.