On Fri, Jul 09, 2010 at 01:03:24PM +0200, Arne Babenhauserheide wrote: > On Friday 09 July 2010 11:51:11 Sergiu Ivanov wrote: > > Sadly, since I am very-very short on time now, nsmux is currently > > stuck at almost the same capabilities as about a year ago: it works > > for the simplest file,,a,,b . It does not shutdown the translators a > > and b after the client that has requested the translation closes the > > port. > > Putting it differently: It has a memory leak the size of the started > translators :)
True :-) As far as I can remember, nsmux keeps the control ports of the started translators, which, as I understand it, won't let them go away because of absence of clients. > Does it reuse the translators if I access the same file,,a,,b? Nope. We didn't fully consider implementing this because reusing a translator would require the translator to be aware of the fact that it might be reused by different clients. Having nsmux always provide a brand new translator stack prevents clients from seeing stale data. We did discuss whether reusing translators would be relevant in the case of larger translators, but we decided to leave it for some other day, as far as I remember. > And can I kill the translators manually? Yes, you can kill them manually, but nsmux does nothing about dead control ports; this memory leak you cannot avoid easily :-( Nevertheless, I'd expect that killing translators won't have other bad effect than leaving nsmux with invalid control ports: nsmux only uses control ports for closing the translators when it itself closes. Best regards, Sergiu
