Jean-Michaël Celerier

thanks for the report - we're kinda busy next week as we're having a week-long meeting but I'll do my best to look at it

Okay, I checked this and I'd be inclined to believe that this is a threading issue : the m_host_info could be written to by the network thread and isn't protected...

hmm I think so - AFAIK this is contrary to the spec and degrades performance, but on the other hand removing this may break existing patches...

> would that make sense ? yes, but it would change an *awful* lot of stuff in the implementation because now you have to have a scheduler and you can't...

> we're talking pd here, not i-score, right ? sure, but to make sense it would have to be at the libossia level, not the ossia-{pd,max,...} level because it makes...

Hello, not very much. I did a first experiment a long time ago as you can see, but due to the *extremely hard to use* Node API and exactly no...

hm, here it seems to load in 23 ms on windows, which seems fairly reasonable... 

you are really getting 22s of wait there ?

tried on mac too, it takes ~100ms which is slower but not shocking either

Thanks, that's really useful ! The dnssd thing really hints towards a network issue (it's the "technical" name of Bonjour). I'll maje you a build without it so that we...