[PD-dev] [ pure-data-Bugs-1891178 ] netsend: messages sent just after connect are lost/corrupted

[SourceForge.net]

Bugs item #1891178, was opened at 2008-02-11 06:45
Message generated for change (Comment added) made by nobody
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1891178&group_id=55736

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: v0.40.1
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Matteo Sisti Sette (sistisette)
Assigned to: Nobody/Anonymous (nobody)
Summary: netsend: messages sent just after connect are lost/corrupted

Initial Comment:
If you send a great enough number of messages throught a netsend immediately after the connection is established (for example triggered by the 1 output of netsend), the first messages are lost (and no error message issued). Netreceive receives some corrupted data, and then the last N messages sent.

The attached patch illustrates the problem.


** TEST PROCEDURE:
Click on the [connect ...( message box

** EXPECTED BEHAVIOUR:
On PD output window the following should appear:

CONNECTED: 1
DATA: bla bla 1
DATA: bla bla 2
DATA: bla bla 3
...
DATA: bla bla 1000

** OBSERVED BEHAVIOUR
The following output appears:
CONNECTED: 1
DATA: 53
DATA: bla bla 954
DATA: bla bla 955
...
DATA: bla bla 1000

====

Note that if you later click the [bng], all works as expected, i.e. no message is lost: message loss only seems to happen when too many messages are sent immediately after connection.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2008-02-13 03:54

Message:
Logged In: NO 

This DOESN'T ONLY happen when the burst of messages is sent immediately
after connection.

I have experienced message loss even sending a big burst of messages to a
netsend that has been connected for a long time.

I may send a burst of M messages and loose messages K through N with
0<K<N<M.
In this case message N+1 is corrupted (its last part is received).

Also, it seems that the burst doesn't need to be sent in 0 logical time to
reproduce the bug. It seems it is sufficient to send enough messages in
enough little time.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1891178&group_id=55736