[asterisk-dev] asterisk curl function and socket keep alive

[Terry Wilson]

On Mar 1, 2011, at 1:53 PM, Paul Albrecht wrote:

> On Tue, 2011-03-01 at 12:36 -0600, Terry Wilson wrote:
>>> I asked this question yesterday and didn't get much of a response so
>>> I'll try again ... Why is the keep alive socket option used for curl
>>> connection requests? The obvious disadvantage of keep alive is that the
>>> socket file descriptor is retained until the channel thread terminates.
>> 
>> Don't start a new thread just because you didn't get the answer you want within a day. It isn't going to help. You also were talking with the author of func_curl. He gave you the reason "because it is the default" and suggested that the only reason it hasn't been changed is because no one had submitted a suitable patch to change the functionality.
>> 
> 
> I'm simply reporting a problem and haven't received an appropriate
> response. An appropriate response would be either: it's a feature or
> it's a bug. If it's the former than I expect a proper explanation, if
> it's the latter than I expect a fix.

No. You are re-reporting it in a way that makes it harder to follow a discussion thread. If you wanted to continue the discussion, you should do it in the same thread. Creating a new thread is not the right answer. The response you got was that the reason that it is that way is because it is the default of the library--which means that it wasn't an intentional choice. You also have no right to expect a fix from someone who is working on something in their free time. Calm down and be polite. If you think it is a bug, submit a bug report.

> What's a proper explanation? It's not "that's the way I wrote the code
> so that's the way it works." The author of the function should be able
> to give me an application scenario where the particular implementation
> makes sense from either an efficiency or usability perspective.

If that is the only reason that exists ("I didn't override the default"), then that is the reason. He isn't saying that it is the right way to do it. That is why "patches welcome" was part of the response. People don't request patches for things they think are already optimal.

> As for submitting a patch, that seems rather pointless since the author
> is unable to state a reason for writing the code one way or another.

See above.

> And finally, simply complaining when someone points out a problem in
> code you're maintaining is awfully lame. If you don't have anything
> "technical" to contribute to the discussion why do you feel the need to
> respond?


I didn't complain about you pointing out a problem. I complained about you pointing it out twice in two different threads when someone was actually responding to you in the other thread when it had been less than a day from the last response. Be patient. Be polite. Follow posting guidelines that make it easy to track a thread. Don't imply that people responding to you should just shut up. These are things that would make it more likely that you will get the kind of responses you are looking for.

Terry