pcp
[Top] [All Lists]

pcp updates: marko merge, ds389log fix, qa

To: PCP <pcp@xxxxxxxxxxx>
Subject: pcp updates: marko merge, ds389log fix, qa
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Thu, 21 Jul 2016 02:39:17 -0400 (EDT)
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <1925358948.7583046.1469083096548.JavaMail.zimbra@xxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: VYBh08o2LoY46OX6pcwDwnMEK298vA==
Thread-topic: pcp updates: marko merge, ds389log fix, qa
Changes committed to git://git.pcp.io/pcp.git master

Marko Myllynen (5):
      pmrep: revert to ascii
      pmrep: sync qa filters
      python api: ignore fetchgroup context target for local contexts
      python api: destroy fetchgroup only if it really exists
      pmrep: use pmfg instead of own fetch/result decoding

Nathan Scott (4):
      pmda ds389log: use dirsrv as default user account
      python api: fix syntax error in recent context changes
      pmrep: add back a dropped line, from overlapping change merge botch
      qa: fix test 168 to ensure wait for pmlogger on completion


 qa/1069                            |    1 
 qa/1069.out                        |  100 +++++++-------
 qa/1070                            |    6 
 qa/1071                            |    6 
 qa/1071.out                        |   12 +
 qa/1072                            |    6 
 qa/168                             |   13 +
 qa/960                             |    1 
 src/pmdas/ds389log/pmdads389log.pl |    2 
 src/pmrep/pmrep.1                  |   16 ++
 src/pmrep/pmrep.conf               |    7 -
 src/pmrep/pmrep.py                 |  250 ++++++++++---------------------------
 src/python/pcp/pmapi.py            |    7 -
 13 files changed, 175 insertions(+), 252 deletions(-)


commit 5a9748bfda3288b81c044cfb2517afc4bff496a6
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Thu Jul 21 16:36:10 2016 +1000

    qa: fix test 168 to ensure wait for pmlogger on completion

commit ee769fc918484779e4c94aaf5793b667c4bb3373
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Thu Jul 21 16:34:36 2016 +1000

    pmrep: add back a dropped line, from overlapping change merge botch

commit f05032bc4c3269b2407dad5c9354efb1444aad28
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Thu Jul 21 10:15:11 2016 +1000

    python api: fix syntax error in recent context changes

commit ad12db98e3791cddd26b4587cc559614203bf11f
Author: Marko Myllynen <myllynen@xxxxxxxxxx>
Date:   Thu Jul 21 10:09:12 2016 +1000

    pmrep: use pmfg instead of own fetch/result decoding

commit 2451b740e2b3490fc295a828d64aa5bd368a9aa2
Author: Marko Myllynen <myllynen@xxxxxxxxxx>
Date:   Thu Jul 21 10:03:10 2016 +1000

    python api: destroy fetchgroup only if it really exists

commit c4619b3b12be1e6f2e990205f74619c30c9a07dc
Author: Marko Myllynen <myllynen@xxxxxxxxxx>
Date:   Thu Jul 21 09:50:40 2016 +1000

    python api: ignore fetchgroup context target for local contexts
    
    When calling pmNewContext(3) with type PM_CONTEXT_LOCAL, name is
    ignored.  Allow clients using pmapi.py fetchgroups to pass None.

commit 138c117b4375df6a6facaa39153bc420b917509b
Author: Marko Myllynen <myllynen@xxxxxxxxxx>
Date:   Thu Jul 21 09:18:17 2016 +1000

    pmrep: sync qa filters

commit 18d662737355564a651196ec5165cfbb5649f2fb
Author: Marko Myllynen <myllynen@xxxxxxxxxx>
Date:   Thu Jul 21 09:13:54 2016 +1000

    pmrep: revert to ascii
    
    In 06b8c19 we compressed the pmrep unit strings in headers and for
    microseconds we used non-ascii Âs form. This works with Python 3
    but we failed to remember that with Python 2 strings are not
    Unicode by default and later when formatting the output, it's the
    format string's encoding that determines the encoding used.
    
    So this is what we should use:
    
    #!/usr/bin/python2
    # -*- coding: utf-8 -*-
    
    a = [":", "u/s", ":", "us/s", ":"]
    u = [":", u"Âs/s", ":", u"Âs/s", ":"]
    fmt = u"{0}{1:>15.15}{2}{3:>15.15}{4}"
    print(fmt.format(*tuple(a)))
    print(fmt.format(*tuple(u)))
    
    Without using unicode formatting will be off as multi-byte
    characters take the same screen estate as single-byte chars.
    
    But: even after the corresponding changes to pmrep I still get QA
    failures with Python 2 here. Given the recent Python QA issue Ken
    has hit elsewhere it sounds like this might be a recipe for lots
    of additional QA headaches so I'm leaning towards reverting back
    to ASCII. And "us" instead "Âs" is already used elsewhere in PCP.
    Later, if wanted, once the QA is on green on again we can see if
    we want introduce this (not-so-critical) change again.

commit 1f0ee8705d108b628e59746e851f51fd1599190c
Author: Nathan Scott <nathans@xxxxxxxxxx>
Date:   Wed Jul 20 11:47:30 2016 +1000

    pmda ds389log: use dirsrv as default user account
    
    In all current Fedora versions and since RHEL7.3 the default
    user for 389-ds is dirsrv instead of nobody.
    
    Update defaults for pcp-pmda-ds389log to reflect this change.
    The configuration file can still be used to explicitly set an
    alternative user if needed.
    
    Resolves Red Hat BZ #1357607.


<Prev in Thread] Current Thread [Next in Thread>
  • pcp updates: marko merge, ds389log fix, qa, Nathan Scott <=