Remove references to Disconnect category and disconnect_mapping.json

[boolean5]

This PR closes #118 by removing all references to the Disconnect category and to disconnect_mapping.json.

[boolean5]

@skim1102, @englehardt, should I also remove the disconnect_mapping.json file itself or is it still
used by other services?

Also, I noticed that the blocklists in https://github.com/mozilla-services/shavar-experiments are still using the Disconnect category. As a result, the corresponding created lists (fastblock1-track-digest256, fastblock2-track-digest256 and fastblock3-track-digest256) now contain fewer domains than before. Should we wait for the fastblock blocklists to be updated before merging this PR?

[say-yawn]

Also, I noticed that the blocklists in https://github.com/mozilla-services/shavar-experiments are still using the Disconnect category. As a result, the corresponding created lists (fastblock1-track-digest256, fastblock2-track-digest256 and fastblock3-track-digest256) now contain fewer domains than before. Should we wait for the fastblock blocklists to be updated before merging this PR?

@englehardt, as @boolean5 mentioned, the fastblock lists do indeed have Disconnect in its json files, see this fastblock1.json example. It seems the last time shavar-experiments was updated 2 years ago (1yr ago for CODE_OF_CONDUCT). Can we remove the fastblock from shavar-list-creation? Or should we update the shavar-experiments to no longer use the Disconnect category like we did on shavar-prod-lists?

[say-yawn]

@skim1102, @englehardt, should I also remove the disconnect_mapping.json file itself or is it still
used by other services?

@boolean5, we should keep the disconnect-mapping.json for now to ensure other components (iOS or Android) no longer uses/references the file. I will update the issue once I find out.

[englehardt]

@englehardt, as @boolean5 mentioned, the fastblock lists do indeed have Disconnect in its json files, see this fastblock1.json example. It seems the last time shavar-experiments was updated 2 years ago (1yr ago for CODE_OF_CONDUCT). Can we remove the fastblock from shavar-list-creation? Or should we update the shavar-experiments to no longer use the Disconnect category like we did on shavar-prod-lists?

Yes we can remove the Fastblock lists.

[boolean5]

Yes we can remove the Fastblock lists.

I added a commit that removes them. Thanks for clarifying this. Should we also remove the Fastblock lists from https://github.com/mozilla-services/shavar-list-creation-config/ before merging this?

[say-yawn]

@boolean5, yes. Along with that in config repo for shavar-list-creation we should also remove all refernces to Disconnect category in Stage and Prod .ini files.

[boolean5]

After discussing this with @skim1102, I'm pushing a commit that removes the disconnect_mapping.json file.

[boolean5]

@boolean5, yes. Along with that in config repo for shavar-list-creation we should also remove all refernces to Disconnect category in Stage and Prod .ini files.

I just submitted a PR for this: mozilla-services/shavar-list-creation-config#60

[boolean5]

While working on this I noticed that the deprecated Disconnect category is still present in https://github.com/mozilla-services/shavar-staging-lists/blob/master/disconnect-blacklist.json and https://github.com/mozilla-services/shavar-test-lists/blob/master/track.json. These lists are probably not actively used right now but I thought I'd mention it here in case they need to be updated.

[say-yawn]

While working on this I noticed that the deprecated Disconnect category is still present in https://github.com/mozilla-services/shavar-staging-lists/blob/master/disconnect-blacklist.json and https://github.com/mozilla-services/shavar-test-lists/blob/master/track.json. These lists are probably not actively used right now but I thought I'd mention it here in case they need to be updated.

@boolean5, thanks for bringing these up. Can you create an issue on shavar-list-creation to start documenting whether the repositories are being used or referenced anywhere?

[say-yawn]

Test on local dev and works as expected.

[say-yawn]

@boolean5, from my previous experiences I do not merge after 4:30PM my local time to save myself from the potential issues caused by the "5PM push". As a result, I will merge this first thing tomorrow!

[say-yawn]

Shucks, I cannot merge this unless all the commits are signed. Can you sign the commits?

[boolean5]

Sure, I signed the commits and rebased them on the current master.

[boolean5]

Can you create an issue on shavar-list-creation to start documenting whether the repositories are being used or referenced anywhere?

Done! I opened the issue here: #126

[boolean5]

I noticed that category_domains.py in mozilla-services/shavar-prod-lists#156 is planning to use disconnect_mapping.json to compare shavar-prod-lists and disconnect-tracking-protection lists. I see two options now:

1. Removing from this PR the commit that removes the disconnect_mapping.json file.
2. Keeping this PR as it is and moving disconnect_mapping.json to shavar-prod-lists, which is the only repo using it. This has been discussed before in Remove use of google-mapping.json mozilla-mobile/firefox-ios#5968 (comment) and seems like a good idea.

@skim1102, what do you think?

[say-yawn]

@boolean5 since we are going to be removing that file I do not think we should move the file. I left a comment to remove references to disconnect_mapping.json