Sander
Sander
http://127.0.0.1:8080/config/switches/#propagation_delay Posts will be paused until they are at least this age. Setting job priority to Force will skip the delay.
I cannot reproduce. See below. HappyEyeballs AFAIR works like this: simultananously it starts all connections. The first achieved connection is the winner. No need to wait nor print the later,...
> User on discord talking about getting slower than expected speeds with frugal unless he hardcodes in a specific IP. Interesting. So: slower connect, but faster download? Do you have...
> HappyEyeBalls uses a staggered start. Ah, yes: https://github.com/sabnzbd/sabnzbd/blob/c6920e09bf95a30006223686eb94d515ca8c67b4/sabnzbd/happyeyeballs.py#L42-L43 So, @thezoggy ... you could try with a CONNECTION_ATTEMPT_DELAY of 0.0 (and thus not be RFC8305 compliant)
OK, poll: what do we want? * current staggered approach, like SAB does now, compliant with RC8305? * non-staggered approach, (like SAB did in the past?), compliant with older RFC6222...
> But setting it to 1ms is basically the same as just setting it to 0, and thus removing the whole idea of HE. > Should we just do that?...
prototype code: it does not choose the first, but the fastest: And safe with time-out on non-responding server: ``` sander@zwarte:~/happy$ python3 happy5.py news.usenetprime.com 563 20:26:19 [INFO] [SUCCESS] news.usenetprime.com -> 185.151.12.153:563...
Rainy days working too: FQDN does resolve, but no reaction on port 563: ``` sander@zwarte:~/happy$ time python3 happy8.py www.google.nl 563 21:02:06 [WARNING] [FAILURE] www.google.nl -> 142.250.179.163:563 - 888.82 ms (timed...
> Again, I am fine with just picking the fastest. Nice > But then we just don't call it HappyEyeBalls anymore. Not nice. See my earlier statement. Plus: seems like...
This has been discussed before, with conclusion "No", but as a user I still would like to have this. So that you can use SAB to remove a Linux iso...