pcp
[Top] [All Lists]

Re: JSON PMDA with indom cache changes

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: JSON PMDA with indom cache changes
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Wed, 3 Jun 2015 20:09:25 -0400 (EDT)
Cc: David Smith <dsmith@xxxxxxxxxx>, pcp <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <556F6F20.9010003@xxxxxxxxxxxxxxxx>
References: <555DEF05.7030108@xxxxxxxxxx> <461406362.3814890.1432280874168.JavaMail.zimbra@xxxxxxxxxx> <55677BB7.3060805@xxxxxxxxxx> <1432466455.9615965.1433146915880.JavaMail.zimbra@xxxxxxxxxx> <2020113955.10852155.1433298140646.JavaMail.zimbra@xxxxxxxxxx> <556F6F20.9010003@xxxxxxxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: lNvBZj9rj1vpFYygaCZn96zUKdOeHg==
Thread-topic: JSON PMDA with indom cache changes

----- Original Message -----
> On 03/06/15 12:22, Nathan Scott wrote:
> > ...
> > 1052 is also still in need of some sorting love too, I haven't got to
> > that part yet (from my later mail).  Ken, maybe dbpmda should give us
> > a sort-the-names-pdu-contents option, for QA?  (see attachment to the
> > earlier mail I sent David on this for sample .bad)
> 
> I'm not sure what you're looking for in the
> "sort-the-names-pdu-contents" option ... I save the 1052.out.bad in the
> email, but don't have a baseline 1052.out (in my tree) to compare it to
> so perhaps you could send that to me by email, or I'll wait till all of
> this makes it into the official tree.

(see attached)

> But if I was to guess, I would think it is more appropriate to chop,
> sort and paste output in a QA test (as we've done elsewhere) rather than
> ask dbpmda to rewrite a PDU.

Well, it'd be reordering the printfs rather than the PDU itself ... but,
yeah I understand the reluctance.  Dealing with it in the QA test will be
a bit of a pain because we send through a pipeline of commands and would
ideally just control the output at a high level.

Yes, could split/cut/sed/awk etc but it will just make the test complex
(and other tests wanting similar dbpmda output).

Unfortunately, its not quite as simple a change (adding a qsort) as I'd
originally hoped cos __pmDecodeNameList passes back two related arrays
(namelist & statuslist) not just the one ... bleurgh.

cheers.

--
Nathan

Attachment: 1052
Description: application/shellscript

Attachment: 1052.out
Description: Binary data

Attachment: 1052.out.bad
Description: Binary data

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