Maniphest T397762

Write to afl_ip_hex in AbuseFilter
Closed, ResolvedPublic1 Estimated Story Points
Actions

Description

Summary

The abuse_filter_log table has an afl_ip_hex column. We should write to it to migrate from using afl_ip.

Background & Technical details

T391322: Special:AbuseLog: Support search by IP and IP range when temporary accounts are enabled requests that it be possible to search AbuseFilter log entries performed by temporary accounts by the IP address used to perform the action
To achieve this we need to be able to search by the IP address column in the abuse_filter_log table
- This involves writing to the new afl_ip_hex column for new entries, which will be the focus of this task. Existing entries will be migrated in another task
- Additionally, we will need to expire the value from afl_ip_hex in the same way as afl_ip to avoid storing the IPs for too long.

Acceptance criteria

The afl_ip_hex column is written to when creating new abuse_filter_log rows
The afl_ip_hex column is purged in the same way as afl_ip

Details

Related Changes in Gerrit:

	Subject	Repo	Branch	Lines +/-
	Write empty string to afl_ip_hex when wgAbuseFilterLogIP is false	mediawiki/extensions/AbuseFilter	master	+2 -3
	Write to afl_ip_hex	mediawiki/extensions/AbuseFilter	master	+38 -24

Customize query in gerrit

Related Objects
Search...

Status	Assigned	Task
Resolved	STran	T391322 Special:AbuseLog: Support search by IP and IP range when temporary accounts are enabled
Resolved	Dreamy_Jazz	T395612 AbuseFilter abuse_filter_log table: Store IP addresses as hex values
Resolved	Dreamy_Jazz	T400280 Drop `afl_ip` as the last step of the migration to `afl_ip_hex`
Resolved	STran	T400284 Stop writing to `afl_ip`
Resolved	mszwarc	T397941 Update code to read from afl_ip_hex
Resolved	Dreamy_Jazz	T397762 Write to afl_ip_hex in AbuseFilter
Resolved	Dreamy_Jazz	T401793 Add a default value for afl_ip

Event Timeline

Dreamy_Jazz created this task.Jun 24 2025, 7:28 PM

Dreamy_Jazz added a parent task: T397941: Update code to read from afl_ip_hex.Jun 26 2025, 2:43 PM

Dreamy_Jazz claimed this task.Jun 26 2025, 3:26 PM

Dreamy_Jazz set the point value for this task to 1.

Dreamy_Jazz moved this task from Priority Backlog to In Progress on the Trust and Safety Product Sprint (Sprint Baklava (June 16 - July 4)) board.

Dreamy_Jazz updated the task description. (Show Details)Jun 26 2025, 3:35 PM

Change #1164249 had a related patch set uploaded (by Dreamy Jazz; author: Dreamy Jazz):

[mediawiki/extensions/AbuseFilter@master] Write to afl_ip_hex

https://gerrit.wikimedia.org/r/1164249

gerritbot added a project: Patch-For-Review.Jun 26 2025, 3:47 PM

Dreamy_Jazz moved this task from In Progress to Needs Review on the Trust and Safety Product Sprint (Sprint Baklava (June 16 - July 4)) board.Jun 26 2025, 4:07 PM

Dreamy_Jazz moved this task from Inbox to Engineering on the Trust and Safety Product Team board.

Change #1164249 merged by jenkins-bot:

[mediawiki/extensions/AbuseFilter@master] Write to afl_ip_hex

https://gerrit.wikimedia.org/r/1164249

ReleaseTaggerBot added a project: MW-1.45-notes (1.45.0-wmf.8; 2025-07-01).Jun 27 2025, 2:00 PM

Maintenance_bot removed a project: Patch-For-Review.Jun 27 2025, 2:30 PM

kostajh moved this task from Needs Review to Needs QA on the Trust and Safety Product Sprint (Sprint Baklava (June 16 - July 4)) board.Jul 2 2025, 1:42 PM

kostajh edited projects, added Trust and Safety Product Sprint (Sprint Cannoli (July 7 - July 25)); removed Trust and Safety Product Sprint (Sprint Baklava (June 16 - July 4)).Jul 7 2025, 9:41 AM

kostajh moved this task from Priority Backlog to Needs QA on the Trust and Safety Product Sprint (Sprint Cannoli (July 7 - July 25)) board.

I see values being written to afl_ip_hex and I briefly checked that they matched the IP in afl_ip. When I run the PurgeOldLogData.php maintenance script it removes values from afl_ip_hex.

@Dreamy_Jazz One thing I notice is that when $wgAbuseFilterLogIP = false, we always write 0 to the afl_ip_hex. This will then get updated to an empty string when running PurgeOldLogData.php. Would it be more efficient to write an empty string in the first place, just to save having to do an update later?

Test environment: local docker Abuse Filter – (a8a00df) 14:26, 7 July 2025.

In T397762#10984252, @dom_walden wrote:

@Dreamy_Jazz One thing I notice is that when $wgAbuseFilterLogIP = false, we always write 0 to the afl_ip_hex. This will then get updated to an empty string when running PurgeOldLogData.php. Would it be more efficient to write an empty string in the first place, just to save having to do an update later?

We should be writing an empty string in this case. Let me look into this.

Change #1167246 had a related patch set uploaded (by Dreamy Jazz; author: Dreamy Jazz):

[mediawiki/extensions/AbuseFilter@master] Write empty string to afl_ip_hex when wgAbuseFilterLogIP is false

https://gerrit.wikimedia.org/r/1167246

gerritbot added a project: Patch-For-Review.Jul 8 2025, 3:40 PM

Dreamy_Jazz moved this task from Needs QA to Needs Review on the Trust and Safety Product Sprint (Sprint Cannoli (July 7 - July 25)) board.Jul 8 2025, 3:40 PM

Change #1167246 merged by jenkins-bot:

[mediawiki/extensions/AbuseFilter@master] Write empty string to afl_ip_hex when wgAbuseFilterLogIP is false

https://gerrit.wikimedia.org/r/1167246

Dreamy_Jazz moved this task from Needs Review to Needs QA on the Trust and Safety Product Sprint (Sprint Cannoli (July 7 - July 25)) board.Jul 8 2025, 4:38 PM

ReleaseTaggerBot edited projects, added MW-1.45-notes (1.45.0-wmf.10; 2025-07-15); removed MW-1.45-notes (1.45.0-wmf.8; 2025-07-01).Jul 8 2025, 5:00 PM

Maintenance_bot removed a project: Patch-For-Review.Jul 8 2025, 5:31 PM