[erlang-questions] gen_tcp send_timeout

Valentin Micic <>
Wed Jun 6 23:19:07 CEST 2007


As far as I can remember, the only gotcha is that sending buffer must become full first. So, if your sending buffer is *big*, and traffic volumes are low, it may take a while before the timeout kicks in.
Actually I'm surprised to hear that this feature is not supported, as we've been using it since R9 -- I've never used R10, but jumped straight to R11, so not sure if the problem was introduced in R10. Come to think of it -- I don't remember testing in on R11. Hmmm... should I be worried?

V.
  ----- Original Message ----- 
  From: Heinrich Venter 
  To:  
  Sent: Wednesday, June 06, 2007 10:01 AM
  Subject: [erlang-questions] gen_tcp send_timeout


  Hi all

   

  I am running into the infamous gen_tcp:send/2 that blocks indefinately issue on R10.  Upgrading to R11 is in the works, but it is not feasible right now.

  The undocumented feature of send_timeout in setopts might be the way out.  Can any one that uses this feature tell me what the gotchas might be with it?

  Why is it not a supported feature yet?  Under what circumstances migtht it break?

   

  Thanks

   

  -]-[



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


  _______________________________________________
  erlang-questions mailing list
  
  http://www.erlang.org/mailman/listinfo/erlang-questions
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20070606/0e084823/attachment.html>


More information about the erlang-questions mailing list