pcp
[Top] [All Lists]

Re: PCP trees for web and middleware development

To: pcp@xxxxxxxxxxx
Subject: Re: PCP trees for web and middleware development
From: fche@xxxxxxxxxx (Frank Ch. Eigler)
Date: Fri, 26 Sep 2014 12:58:39 -0400
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <2143674986.55503362.1411640081377.JavaMail.zimbra@xxxxxxxxxx> (Nathan Scott's message of "Thu, 25 Sep 2014 06:14:41 -0400 (EDT)")
References: <1970205420.36245669.1408665579915.JavaMail.zimbra@xxxxxxxxxx> <050301cfc01c$04889170$0d99b450$@internode.on.net> <1078845537.37533359.1408945446955.JavaMail.zimbra@xxxxxxxxxx> <y0m4mwqq5o9.fsf@xxxxxxxx> <638874581.42408155.1409659444170.JavaMail.zimbra@xxxxxxxxxx> <20140902134012.GE4825@xxxxxxxxxx> <1859208890.52189214.1411109762944.JavaMail.zimbra@xxxxxxxxxx> <y0ma95sy9s3.fsf@xxxxxxxx> <2143674986.55503362.1411640081377.JavaMail.zimbra@xxxxxxxxxx>
User-agent: Gnus/5.1008 (Gnus v5.10.8) Emacs/21.4 (gnu/linux)
Hi -

For those who have trouble following what has been happening, I will
briefly list my perception of the claims & evidence at play here.


claim:     no one else cares about whereabouts of pmwebd
evidence:  hardly anyone spoke up
counter:   topic made excessively complicated; people not specifically
           polled with clear question


claim:     pmwebd bundled graphite/grafana webapps pose license risks
evidence:  FUD
counter:   graphite webapps (and other parts) are already packaged in fed/deb
           grafana license is same ASL2/MIT, easily audited


claim:     pmwebd c++ core depends on bundled graphite/grafana webapps
evidence:  none
counter:   neither build- nor runtime requirements, just user
           installation-convenience data files; all pmwebd c++ core code
           (including graphite-api-emulation) is executable without
           bundled webapps (and done with pcpqa)


claim:     pmwebd bundled webapps too large
evidence:  opinion
counter:   absence of objective file-size limits, usefulness and convenience
           outweighs a few megabytes


claim:     pmwebd c++ core testing does not fit pcpqa
evidence:  none
counter:   pmwebd c++ core is already well-tested in qa in usual way


claim:     pmwebd bundled webapp testing does not fit pcpqa
evidence:  none
counter:   unclear; mechanical javascript/webapp testing is a future problem,
           it could conceivably be driven from the qa suite (with optional
           auxiliary tools)


claim:     pmwebd build system needs to change to cmake
evidence:  none
counter:   ain't broken


claim:     mere presence of pmwebd makes pcpqa more difficult
evidence:  none
counter:   individual pcpqa tests are / can be made optional, if that
           is desirable


claim:     pmmgr is relevant to this discussion
evidence:  surprise summary action just taken (git-rm)
counter:   zero technical interaction or interdependency between pmwebd + pmmgr;
           sole technical commonalities: initial author and language


claim:     (?) pmmgr is untested
evidence:  (?) no test case in pcpqa
counter:   test cases have existed for months, refusal to merge/review
           due (?) to their long runtime; considerable hand-testing &
           deployment on busy machines


claim:     it is appropriate to cut off discussion and git-rm pmmgr and pmwebd
evidence:  actions just taken
counter:   community judgement, I hope


I believe the community would benefit from a phone conversation, to
discuss what happened here and related maintainership role questions.


- FChE

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