pcp
[Top] [All Lists]

Re: Global Timeout for Service Discovery

To: "Frank Ch. Eigler" <fche@xxxxxxxxxx>
Subject: Re: Global Timeout for Service Discovery
From: Dave Brolley <brolley@xxxxxxxxxx>
Date: Wed, 23 Jul 2014 11:49:55 -0400
Cc: PCP Mailing List <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <y0mzjg1upts.fsf@xxxxxxxx>
References: <53CEB9DB.8050708@xxxxxxxxxx> <y0mzjg1upts.fsf@xxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.5.0
On 07/22/2014 03:59 PM, Frank Ch. Eigler wrote:
That sounds good. The lack of fallback for non-threaded libpcp's
could be a problem; have you considered returning an outright error
instead of the "__pmNotifyErr() & return OK" in this case?
Done.
In __pmDiscoverServicesWithOptions(), during the thread cleanup,
it looks like the code cares about the sts of the pthread_cancel(),
but doesn't really.  -ESRCH is a valid non-error case there, FWIW.
Consider just a "(void) pthread_cancel(...);".
Done.

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

commit 1546ba50d1f0b305a446889565dfd4696fae8e48
Author: Dave Brolley <brolley@xxxxxxxxxx>
Date:   Wed Jul 23 11:45:23 2014 -0400

__pmDiscoverServicesWithOptions(): Return error codes for invalid unsupported options.

    Without this, the API user has no way of knowing that options were
    invalid or ignored.

    Also include minor cleanup based on testing and feedback.

<Prev in Thread] Current Thread [Next in Thread>