Permalink
Commits on Mar 11, 2010
  1. fixed logging err msg typo

    steveyen committed Mar 11, 2010
  2. bug 375, assigning erl listen port range 21100-21199

    Signed-off-by: Sean Lynch <seanl@literati.org>
    steveyen committed with seanlynch Mar 11, 2010
Commits on Mar 10, 2010
  1. Do a config swap every time dist erl membership changes for bug 809

    This means that in the cases where the local node doesn't see the
    remote node immediately, it will as soon as the disco event fires.
    
    Signed-off-by: Steve Yen <steve.yen@gmail.com>
    dustin committed with steveyen Mar 10, 2010
  2. Reduce memory usage (and work in general) in stats_aggregator for bug…

    … 806
    
    ringdicts are used a lot.  This change prevents us from creating the
    default value for storage that we almost never use.
    
    Signed-off-by: Sean Lynch <seanl@literati.org>
    dustin committed with seanlynch Mar 10, 2010
  3. Make ns_cluster an explicit gen_fsm on the path to bug 763

    This is not strictly required, but I need it to be clear how one might
    fall back from a failed join as the fsm expands.
    
    It already *was* an fsm, but done a bit more manually.
    dustin committed with seanlynch Mar 9, 2010
  4. Bug 801: Revert "Turn auto_create back on"

    This reverts commit a5368d9.
    seanlynch committed with dustin Mar 9, 2010
  5. clear bucket creation errors when showing creation dialog. Bug 792.

    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 10, 2010
  6. [cells] invoke callbacks in 'quiescent' state. References bug 706.

    previously due to cell value propagation delay it was possible to be
    running callback for cell which value if calculated now would be
    undefined.
    
    It was possible to trigger (harmless, it seems) exceptions due to
    this. For example, queued update of analytics graph would run even if
    user already switched UI to another section.
    
    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 10, 2010
  7. detect aborted requests and don't panic. References bug 706

    The cause of bug 706 seems to be our own aborted requests,
    mis-detected as XHR failures.
    
    But there are other bad things that are still possible. User can abort
    request by pressing Escape. And it seems there's no portable way to
    distinguish that from not responding server.
    
    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 10, 2010
  8. Added favicon. Bug 749

    Signed-off-by: Aliaksey Kandratsenka <alk@tut.by>
    ingenthr committed with Aliaksey Kandratsenka Mar 10, 2010
  9. Bug 806: move global singleton to last in startup sequence

    This will hopefully reduce the number of cases in which a node's coming
    up will cause the global singleton to crash. It won't, however, prevent
    the global singleton from crashing if a node comes up without network
    connectivity and then gains it.
    
    Signed-off-by: Dustin Sallings <dustin@spy.net>
    seanlynch committed with dustin Mar 10, 2010
Commits on Mar 9, 2010
  1. Empty out local config before joining remote cluster. bug 748

    Signed-off-by: Sean Lynch <seanl@literati.org>
    dustin committed with seanlynch Mar 9, 2010
  2. Added emoxi testing for MS Windows (TM) platoforms

    Patrick Galbraith committed Mar 9, 2010
  3. Fix for login screen coming up instead of app. Bug 796.

    Reviewed by Eric Lambert on github, so committing.  Quote:
      "my name is eric lambert and i approve this commit Eric Lambert @ 1:17"
    ingenthr committed Mar 9, 2010
  4. Implemented very basic diagnostics tool.

    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    
    Note, this is related to enhancement in bug 694.  Normally this wouldn't
    be cherry-picked over until completion, but the value of this in assisting
    our debugging efforts makes it worthwhile.  The remainder of the
    implementation will come over in a different commit and  then the bug
    number will show in the summary.
      Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 9, 2010
  5. Made 'About' work even on login screen. Bug 689.

    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    
    This also lead to bug 794 being opened to documenting the new interface.
    Aliaksey Kandratsenka committed with ingenthr Mar 9, 2010
  6. Reload pool details for free memory value post bucket creation. Bug 762.

    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 9, 2010
  7. Reload app with reset auth cookie on 401 XHR status. Bug 750.

    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 9, 2010
  8. Fixed alerts refresh. Bug 637.

    Signed-off-by: Matt Ingenthron <ingenthr@cep.net>
    Aliaksey Kandratsenka committed with ingenthr Mar 9, 2010
  9. Terminate and delete existing mc_accept children for bug 679

    This is buried deep within pool code allowing multiple acceptors.  We
    have no way to support this feature.
    
    This great scenario from the c2 wiki on YAGNI explains what's going on
    here (http://c2.com/xp/YouArentGonnaNeedIt.html):
    
    You're working on some class. You have just added some functionality
    that you need. You realize that you are going to need some other bit
    of functionality. If you don't need it now, don't add it now. Why not?
    
    "OK, Sam, why do you want to add it now?"
    
    "Well, Ron, it will save time later."
    
    But unless your universe is very different from mine, you can't 'save'
    time by doing the work now, unless it will take more time to do it
    later than it will to do now. So you are saying:
    
    "We will be able to do less work overall, at the cost of doing more
    work now."
    
    But unless your project is very different from mine, you already have
    too much to do right now. Doing more now is a very bad thing when you
    already have too much to do.
    
    And unless your mind is very different from mine, there is a high
    chance that you won't need it after all, or that you'll need to
    rewrite or fix it once you do need it. If either of these happens, not
    only will you waste time overall, you will prevent yourself from
    adding things that you do need right now.
    
    "But Ron, I know how to do it right now, and later I might not."
    
    "So, Sam, you're telling me that this class you're writing is so
    complex that even you won't be able to maintain it?"
    
    Keep it simple. If you need it, you can put it in later. If you don't
    need it, you won't have to do the work at all. Take that day off.
    
    Signed-off-by: Steve Yen <steve.yen@gmail.com>
    dustin committed with steveyen Mar 9, 2010
  10. Logging changes in support of bug 679

    Signed-off-by: Steve Yen <steve.yen@gmail.com>
    dustin committed with steveyen Mar 9, 2010
  11. added TAGS rule for whole ns_server

    Aliaksey Kandratsenka committed Mar 9, 2010
  12. removed no more actual TODOs

    Aliaksey Kandratsenka committed Mar 9, 2010
  13. commented-out $m debugging aids

    Aliaksey Kandratsenka committed Mar 9, 2010
  14. removed template debugging code

    Aliaksey Kandratsenka committed Mar 9, 2010
  15. really ignore errors from XHR posts of client-side error

    Aliaksey Kandratsenka committed Mar 9, 2010
  16. Excise the NaNs from Web UI. Bug 593.

    While there's still lack of clarity for memory status for a down
    node (for instance, it may show as memory available with a down
    notification, if memcached crashed), the decision was made to just
    excise the NaNs from the response.
    
    In testing I found one other case which would show Inifinity.  I've
    normalized both to 0.
    
    This was tested with FireFox 3.5 and IE 6.
    ingenthr committed Mar 8, 2010
  17. Make ns_port_sup return pids of the cushion's child.

    Previously, it was returning the cushion's pid, which was being used by
    ns_port_init:reconfig to determine memcached's parameters. This would
    have always returned ok, so we'd have either crashed or always restarted
    memcached on reconfig.
    
    Since we were returning the cushion's pid even when the cushion's child
    was crashed, ns_heart would think memcached was running even when the
    cushion was just sleeping before crashing.
    
    Fixes bug 194.
    Reviewed by dustin.
    seanlynch committed Mar 9, 2010
Commits on Mar 8, 2010
  1. copyright headers

    steveyen committed Mar 8, 2010