Skip to content

Releases: nginadofficial/nginad

1.3.8 Release

12 Nov 22:31
Compare
Choose a tag to compare

Finally fixed the SQL caching bug with incremental writes that was messing up the stats with concurrent requests on high load. The Java sychronized keyword was pseudo implemented in PHP with an APC memory lock. pthreads could potentially have been used but would have made NginAd harder to install and hindered the adoption.

1.3.7 Release

10 Nov 18:54
Compare
Choose a tag to compare

NGIN-11: Added cookie matching and quality scoring pixels by default

1.3.6 Release

09 Nov 18:25
Compare
Choose a tag to compare

The delivery mechanism was sending out 2 Ad Tag requests back to the NginAd instance if a publisher or DSP's ad tag HTTP request was filled by a dashboard demand partner instead of a SSP. These 2 separate requests were completely unnecessary and a waste of bandwidth and resources. It was fixed by putting in flags to determine if the RTB request was filled by a remote SSP or by the demand dashboard partner's Loopback RTB request. If it was filled by a loopback RTB request, the demand customer's request is immediately processed instead of sending back a new ad tag.

1.3.5 Release

09 Nov 06:32
Compare
Choose a tag to compare

Added DateUpdated TimeStamp update on modification for various columns where it was missing. The most notable is the PublisherHourlyBids table

1.3.4 Release

09 Nov 05:59
Compare
Choose a tag to compare

Removed unwanted PublisherHourlyBidsID key in the cached writes when querying for an existing row.

1.3.3 Release

09 Nov 05:38
Compare
Choose a tag to compare

A unique key constraint was missing from the Publishers stats totals so it was added.
The default auth_Users would not insert because they were missing the MSA acceptance date column, that was added.
Percentage totals were calculated wrong in the reporting, that was fixed.

1.3.1 Release

07 Nov 23:41
Compare
Choose a tag to compare

This release adds views and tables to record the totals information for any one zone's sell side RTB request as opposed to only recording them for the SSP outgoing RTB bid. This way publishers and admins can see the fill rate and total requests for the impressions and take actions based on that data.
Also the MDYH columns were populated with 12 hour PHP date strings instead of by 24 hour PHP date strings which concatenated 2 hours of data into a 1 hour row. This has been fixed.

1.3 Release

06 Nov 01:56
Compare
Choose a tag to compare

This release was all about doing further bug fixes to the train-accident-like development of the reporting module by the Ukraine team. SQL views were fixed, totals were added to the ZF2 views, the AJAX responses, and more fixes were implemented. Also the AdCampaignBanner stats were fixed so that they are updated with the correct amounts as part of the crontab maintenance. Also the eCPMs from the SQL views were fixed so that they are calculated properly in the SQL queries for the views. Use the dbschema/releases/1.3/nginad-1.3.sql from now on when setting up an installation.

1.2.6 Release

19 Oct 23:00
Compare
Choose a tag to compare

All the macro replacements in the winning RTB tag invocation were not being replaced. So the RTB SSP tag invocation did not get the winning bid amount for their statistics. This was fixed. Second price decryption in Ad Delivery is only applied to DSP bids. The local demand winning bids do not get macro replaced in the invocation tag because we can get them from the AdCampaignBanner, so why expose that in the URL.
The bottom line is that putting the winning bid or second price auction bid in the invocation tag code is risky because if another buyer has good logic they will try to scoop it out of the URL and bid less than the SSP paid. The SSP bid price macro replacement in the winning auction tag should be implemented on a Ad Network level and I put a dummy routine which you can fill out in PingManager called encrypt_bid($unencrypted_bid_price)

1.1.3 Release

14 Oct 17:55
Compare
Choose a tag to compare

This release fixes various bugs with the stats reporting module authored by the Ukraine team which was quickly spot fixed before the 1.0 release. A lot of the fixes in this release are security fixes to the AJAX actions where the Ukraine team left holes which allowed anyone to view statistics on all users by calling the AJAX REST services directly. It also fixes a serious bug with the gross/net income for incoming RTB bids.