Jump to content

Matteo Beccati

Administrators
  • Posts

    269
  • Joined

  • Last visited

Posts posted by Matteo Beccati

  1. I had a look at the site and I'm sorry to say that you might not using the right tool for the job. A custom application or wordpress/drupal/etc module would probably serve your purpose better than a proper adserver can do.

     

    That said, try to disable the cookie based fatures (e.g. delivery capping/blocking) and see if that helps.

  2. Thanks for the report. It is a known issue, since many years actually. The problem is that the oadest parameter is required for certain functionalities to work, and to this day no solution has been found that allows it while making it safer. I have spent myself numerous hours on it but I couldn't come up with anything good enough.

     

    If you have a working solution, we'd be happy to review it and apply it.

  3. Upgrade from phpAdsNew should be working from 2.0.11. If you're running 2.0.7, you might first need to upgrade to 2.0.11-pr1. I would suggest doing that with an old enough version of PHP (on a staging server, possibly). From there, it should be possible to upgrade directly to the latest Revive, but to be honest I haven't tested that in a while.

  4. @al_bullit Did you apply both the patch files? DId you properly recompile/upgrade your packages and restart the webserver/php-fpm? If so, I'm afraid I don't know what else to suggest.

     

    @msellers Whatever works for you. If you're able to fix the adserver and influence garbage collection so that it doesn't trigger the GC bug, I'll be very happy to apply your changes and lift the ban for the PHP versions affected by such bug.

  5. Hi Mark,

     

    thanks for the report. To be honest I wasn't aware of it as my knowledge of RPM-based distros is quite limited. I made a search a while ago and I thought that RHEL/Centos 7 had PHP 5.5 out of the box, which sadly isn't actually true.

     

    The fact here is that, even if we wanted to, we can't make Revive Adserver compatible with PHP 5.4.16 as it is affected by a bug that causes segmentation faults when running it. Such bug has been fixed in PHP 5.5.2 and PHP 5.4.20.

     

    Debian had the fix backported at some point to their Wheezy (going from memory) php package, but RedHat didn't yet. It would probably take some RH customer to submit a ticket for them to think about backporting the fix.

     

    The only suggestion I have (thanks Remi) would be to use PHP 5.5 from RHSCL:

     

    https://www.softwarecollections.org/en/scls/rhscl/php55/

     

    I hope this helps.

  6. You are correct in that the cookies related to delivery cappin are not set properly. I'm not sure how the Flash video player is involved in the process. I'm using JW Player 6.10 with ads.

     

    (A) When I observed cookies in Firefox/Firebug it loads these:

    _OACAP[1]

    _OASCAP[1]

    _OABLOCK[1]

    _OACCAP[1]

    _OACBLOCK[1]

    _OAZCAP[1]

     

    ( B) After first quartile those cookies just disappear without a trace to these cookies:

     

    OACAP

    OASCAP

    OABLOCK

    OACCAP

    OACBLOCK

    OAZCAP

    OAZBLOCK

     

    Is this the intended behaviour? (Rhetoric question)

    Yes, in theory. Cookies beginning with an underscore are "temporary" cookies that get merged to the permanent ones.

    When I added this line

    if(@$_REQUEST["vast_event])MAX_cookieFlush();//quick'n'dirty bug fix for frequency capping with video ad delivery 
     

    to www/delivery/fc.php after line 31 it happens so that cookie set (A) starts to "interact" with cookie set ( B).

     

    Any insight on this?

    Capping should be set when the impression is recorded, not when the request is made, so it is techincally wrong.

×
×
  • Create New...