My interest has been piqued by netpd and a bunch of us here have been playing with it. It's pretty fantastic --- there's only one caveat, which is that the netpd server crashes and pretty easily at that. I've seen the netpd.org server crash a bunch of times now. I downloaded and installed the server patch on an ubuntu Gutsy machine and that crashes similarly. I looked at the patch --- nothing too complex. I suspect there is nothing wrong with the patch itself.
I'm suspecting the problem is with netserver, written by Olaf Matthes . I suspect this because netsend~ and netreceive~ crash in my use in similar ways to what I am seeing now with netpd --- Olaf Matthes wrote those too. I was wondering if anybody could confirm or refute that netserver is what is crashing netpd. It would be great if we fixed this. In general I think we have a problem with some externals involving network communication and if we fixed this in several of the externals maybe we would solve a bunch of problems at once. I might be up for a shot at it myself.
-John
P.S. Some of us tuned in for the netpd jam that we thought happens every Thursday at 22:00 CEST. Nobody was there except us. Is this just old outdated info? http://www.netpd.org/Schedule
On Fri, 2008-08-08 at 08:25 -0500, John Harrison wrote:
My interest has been piqued by netpd and a bunch of us here have been playing with it. It's pretty fantastic --- there's only one caveat, which is that the netpd server crashes and pretty easily at that. I've seen the netpd.org server crash a bunch of times now. I downloaded and installed the server patch on an ubuntu Gutsy machine and that crashes similarly. I looked at the patch --- nothing too complex. I suspect there is nothing wrong with the patch itself.
I'm suspecting the problem is with netserver, written by Olaf Matthes . I suspect this because netsend~ and netreceive~ crash in my use in similar ways to what I am seeing now with netpd --- Olaf Matthes wrote those too. I was wondering if anybody could confirm or refute that netserver is what is crashing netpd. It would be great if we fixed this. In general I think we have a problem with some externals involving network communication and if we fixed this in several of the externals maybe we would solve a bunch of problems at once. I might be up for a shot at it myself.
hi john
what you describe is known to us, but unfortunately i didn't find a way to solve it by myself. also the netpd-server patch running on netpd.org crashes from time to time. i am not absolutely sure about the reason and i also tried to debug it by running pd within gdb. therefor i compiled millers pd, zexy and maxlib (the dependencies of the server patch) with debug symbols, but unfortunately it still didn't show me the symbols of the point where it crashed. i am not too familiar with these tools, so probably i made something wrong. from my experience i can at least confirm, that most likely [netserver] from maxlib is the trouble maker. from what i can tell, crashes happen more likely, if several clients disconnect without saying 'adieu'. (http://en.wikipedia.org/wiki/Transmission_Control_Protocol#Connection_termin... ) this happens, when internet connection of a client suddenly is lost or if pd of a client crashes. sometimes it doesn't crash for months, when all clients have good connection (an they don't crash too often). i would be interested to know, what conditions are needed to crash pd/server-patch on your box. does it behave the same? if so, i wonder if [netserver] really complies with tcp. i also wonder what host is supposed to do if only one end point finishes the connection (and this what happens, if a client crashes, i suppose).
P.S. Some of us tuned in for the netpd jam that we thought happens every Thursday at 22:00 CEST. Nobody was there except us. Is this just old outdated info? http://www.netpd.org/Schedule
yo.. i am very sorry. we are (i should rather say: i am) sometimes quite lazy and undisciplined people. actually it is supposed to still be actual, but during the last weeks there wasn't much traffic and we often dropped the sessions (and i was also more busy than usual). but yeah, lets have jam sessions. probably inform us by mail, if you know beforehand that you are going to have a session. i would like to join. there is also this other happening every second tuesday of the month, where we have our session broadcasted on an internet radio (http://audioasyl.net). usually we are three or more people there.
roman
Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de
On Aug 8, 2008, at 3:12 PM, Roman Haefeli wrote:
On Fri, 2008-08-08 at 08:25 -0500, John Harrison wrote:
My interest has been piqued by netpd and a bunch of us here have been playing with it. It's pretty fantastic --- there's only one caveat, which is that the netpd server crashes and pretty easily at that.
I've seen the netpd.org server crash a bunch of times now. I downloaded
and installed the server patch on an ubuntu Gutsy machine and that
crashes similarly. I looked at the patch --- nothing too complex. I suspect there is nothing wrong with the patch itself.I'm suspecting the problem is with netserver, written by Olaf
Matthes . I suspect this because netsend~ and netreceive~ crash in my use in similar ways to what I am seeing now with netpd --- Olaf Matthes
wrote those too. I was wondering if anybody could confirm or refute that netserver is what is crashing netpd. It would be great if we fixed
this. In general I think we have a problem with some externals involving network communication and if we fixed this in several of the
externals maybe we would solve a bunch of problems at once. I might be up for a shot at it myself.hi john
what you describe is known to us, but unfortunately i didn't find a
way to solve it by myself. also the netpd-server patch running on
netpd.org crashes from time to time. i am not absolutely sure about the
reason and i also tried to debug it by running pd within gdb. therefor i compiled millers pd, zexy and maxlib (the dependencies of the server patch)
with debug symbols, but unfortunately it still didn't show me the
symbols of the point where it crashed. i am not too familiar with these tools, so probably i made something wrong. from my experience i can at least confirm, that most likely
[netserver] from maxlib is the trouble maker. from what i can tell, crashes happen more likely, if several clients disconnect without saying 'adieu'. (http://en.wikipedia.org/wiki/ Transmission_Control_Protocol#Connection_termination ) this happens, when internet connection of a client suddenly is lost or if pd of a client crashes. sometimes it doesn't crash for months, when all clients have good connection (an they don't crash too often). i would be interested to know, what conditions are needed to crash pd/server-patch on your box. does it behave the same? if so, i
wonder if [netserver] really complies with tcp. i also wonder what host is supposed to do if only one end point finishes the connection (and this what happens, if a client crashes, i suppose).
Sounds like you guys should try Martin Peach's [tcpserver].
.hc
P.S. Some of us tuned in for the netpd jam that we thought happens
every Thursday at 22:00 CEST. Nobody was there except us. Is this just old outdated info? http://www.netpd.org/Scheduleyo.. i am very sorry. we are (i should rather say: i am) sometimes
quite lazy and undisciplined people. actually it is supposed to still be actual, but during the last weeks there wasn't much traffic and we
often dropped the sessions (and i was also more busy than usual). but yeah, lets have jam sessions. probably inform us by mail, if you know beforehand that you are going to have a session. i would like to join. there is also this other happening every second tuesday of the month, where we have our session broadcasted on an internet radio (http://audioasyl.net). usually we are three or more people there.roman
___________________________________________________________ Telefonate ohne weitere Kosten vom PC zum PC: http:// messenger.yahoo.de
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
kill your television
On Fri, 2008-08-08 at 17:49 -0400, Hans-Christoph Steiner wrote:
Sounds like you guys should try Martin Peach's [tcpserver].
yeah, that is actually the plan on the mid-term/long run. since basically only [netserver] and [netclient] are really needed from maxlib, i would like to switch to mrpeach's classes.
roman
___________________________________________________________ Der frühe Vogel fängt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de
Roman Haefeli wrote:
i would be interested to know, what conditions are needed to crash pd/server-patch on your box. does it behave the same?
it appears to behave the same i.e. crashes about the same amount of time and I don't detect a consistent pattern to the crashes.
I saw the same behavior (crashing) with netsend~ and netreceive~ so a part of me just wants to see what it would take to fix Olaf's code so that all of these things work. I might be looking into this further...
On Tue, 2008-08-12 at 07:13 -0500, John Harrison wrote:
Roman Haefeli wrote:
i would be interested to know, what conditions are needed to crash pd/server-patch on your box. does it behave the same?
it appears to behave the same i.e. crashes about the same amount of time and I don't detect a consistent pattern to the crashes.
I saw the same behavior (crashing) with netsend~ and netreceive~ so a part of me just wants to see what it would take to fix Olaf's code so that all of these things work. I might be looking into this further...
i think, i will follow hans' suggestion, since it doesn't change any dependencies on the client side (actually i thought about writing a server in a different language than pd, probably lua, but first i have to learn it a bit more).
i think i go for [tcpserver] first. i rather like the mrpeach net classes anyway, because they are plain TCP, while [netserver]/[netclient] follows the FUDI protocol (so does [netsend]/[netreceive]).
i let you know when i have a replacement.
roman
Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de
-
Hans-Christoph Steiner -
John Harrison -
Roman Haefeli