Re: [htdig3-dev] SIGALRM handler fixed


Subject: Re: [htdig3-dev] SIGALRM handler fixed
From: Geoff Hutchison (ghutchis@wso.williams.edu)
Date: Tue Feb 29 2000 - 06:04:10 PST


At 2:23 PM +0100 2/29/00, Gabriele Bartolini wrote:
>1) the retries management (now it retries only once): should we
>issue a new config attribute or just use the max_retries in
>parallel? Please focus your attention here.

As I said earlier, I think we should at least have some sort of
low-level retry because it's going to get through a dead connection
faster. I don't think it matters whether we have more than one
attribute or not. Also, if someone can demonstrate that the time
difference between retries is going to be insignificant with the
high-level one, then it clearly doesn't matter.

>2) Connect() sleeps for 5 seconds after receiving the signal. Can we
>modify it?

I agree with Loic. This seems a reasonable default--I also wouldn't
mess with it too much because if you set it too short, even with a
few retries, it may just race the server.

>3) It's not very clear the error EALREADY. Can you give me an
>explanation, please? I read man page but it's not very explanaitory.

I don't think there's much documentation *besides* the man pages. I'm
not much of a socket programmer though.

-Geoff

------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
htdig3-dev-unsubscribe@htdig.org
You will receive a message to confirm this.



This archive was generated by hypermail 2b28 : Tue Feb 29 2000 - 06:12:53 PST