hi. for transmitting OSC over TCP/IP i thought about using slip (rather than [(un)packOSCstream] [1]).
now martin has already done this, but unfortunately i'm not very satisfied yet. probably i misunderstand something, but some things just don't work for me (even in the help patch):
encoding [1 2 3 4( results in [192 1 2 3 4 192( (fine), and decoding this results in [1 2 3 4( (great), but why do i get an error: "slipdec_list: dropped 1 bytes after packet" ???
<snip> in: 192 1 2 3 4 192 valid: 1 decoded: 1 2 3 4 slipdec_list: dropped 1 bytes after packet </snip>
the example has [192, 1, 219, 220, 192( which 'should give 1 192', but really is only decoded to [192(.
<snip> in: 192 in: 1 valid: 0 in: 219 in: 220 valid: 1 decoded: 192 in: 192 </snip> sending the data as a single list, get's me the correct [1 192(, but again with the "dropped 1 bytes after packet" error.
other single serialized lists won't work at all, e.g. sending the first example as [192, 1, 2, 3, 4, 192(, will give me no result at all
<snip> in: 192 in: 1 valid: 0 in: 2 valid: 0 in: 3 valid: 0 in: 4 valid: 0 in: 192 </snip>
obviously this defeats the purpose of the object a bit (after all, it's meant to reassemble broken up packages).
i thought, it should be able to decode [1, 192, 2 3, 4, 5 192 6, 192( into [2 3 4 5, 6( (dropping the leading 1, as the line is unintialized)
fgmasdr IOhannes
[1] i really find the slip approach more convincing when it comes to lost packages; with un/packOSCstream you must never miss the beginning of the stream, else you are lost forever...
On 2010-08-23 17:17, IOhannes m zmoelnig wrote:
now martin has already done this, but unfortunately i'm not very satisfied yet. probably i misunderstand something, but some things just don't work for
unfortunately the [verbosity 1( doesn't seem to change much (i remember that it once worked, but i cannot reproduce that now...)
fgmasdr IOhannes
On 2010-08-23 17:17, IOhannes m zmoelnig wrote:
now martin has already done this, but unfortunately i'm not very satisfied yet. probably i misunderstand something, but some things just don't work for me (even in the help patch):
and i already have one wishlist item: the 2 slip objects currently have a maximum slip package length of 1006 bytes. i know that rfc1055 mentions this maximum size, however:
implementation...
long OSC-messages (e.g. OSC bundles)
for one thing, i don't think that [slipdec] should have a size restriction. usually such implementations should be conservative at the sender side, but liberal at the receiver side. (it probably would be nice if the user could clear the current decoding buffer without output - just in case...)
i would also like, if [slipenc] would have a user-settable maximum size (of course, defaulting to 1006), so one could produce larger packages on demand. this is not high priority, as it is easy to repack a number of slip packages into a single one, but still...
fgmasd IOhannes
Is there such an animal We need it for a project
pp
No one has built it, but it shouldn't be so hard. A good place to
start would be to try the library template on FreeBSD, then we can
incorporate the changes in about 25 libraries:
http://puredata.info/docs/developer/LibraryTemplate
Mostly, you'll want to look at the Makefile.
.hc
On Aug 23, 2010, at 1:51 PM, Pagano, Patrick wrote:
Is there such an animal We need it for a project
pp
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
The arc of history bends towards justice. - Dr. Martin Luther
King, Jr.