Change Details

As described in T320434#8741708, we should create a new table for storing tracking tools: ```lang=mysql CREATE TABLE ce_tracking_tools ( cett_id BIGINT UNSIGNED AUTO_INCREMENT NOT NULL, -- Surrogate primary key cett_event BIGINT UNSIGNED NOT NULL, -- Foreign key to campaign_events.event_id cett_tool_id INT NOT NULL, -- Identifier of the tool, same as the existing campaign_events.event_tracking_tool_id cett_tool_event_id BLOB NOT NULL, -- Identifier of the event within this tool, same as the existing campaign_events.event_tracking_tool_event_id cett_sync_status INT NOT NULL, -- Status of the sync, the meaning of each value is defined in the application cett_last_sync BINARY(14) NOT NULL -- Timestamp of the last synchronization ); ``` Also, `campaign_events.event_tracking_tool_id` and `campaign_events.event_tracking_tool_event_id` need to be dropped. Note that these fields can only be NULL for now. ==Acceptance criteria== [] Create the new table and drop the old columns in the code -- We don't need a backwards-compatible schema change for this: both production and the beta cluster use a database other than the wiki's local DB, meaning schema changes cannot be applied automatically anyway (related: T314908) [] Adjust the code so that it reads from and writes to the new schema -- This **needs** to be backwards-compatible, or things will break horribly [] Create the new schema in beta (task TBC) [] Create the new schema in production (task TBC) [] Remove back-compat code

As described in T320434#8741708, we should create a new table for storing tracking tools: ```lang=mysql CREATE TABLE ce_tracking_tools ( cett_id BIGINT UNSIGNED AUTO_INCREMENT NOT NULL, -- Surrogate primary key cett_event BIGINT UNSIGNED NOT NULL, -- Foreign key to campaign_events.event_id cett_tool_id INT NOT NULL, -- Identifier of the tool, same as the existing campaign_events.event_tracking_tool_id cett_tool_event_id BLOB NOT NULL, -- Identifier of the event within this tool, same as the existing campaign_events.event_tracking_tool_event_id cett_sync_status INT NOT NULL, -- Status of the sync, the meaning of each value is defined in the application cett_last_sync BINARY(14) NOT NULL -- Timestamp of the last synchronization ); ``` Also, `campaign_events.event_tracking_tool_id` and `campaign_events.event_tracking_tool_event_id` need to be dropped. Note that these fields can only be NULL for now. **Note**: this only has to be backwards-compatible in the sense that the code won't crash: the schema is still unused, so we don't really care what is being read/written. Also, we don't need all the files for the schema change: both production and the beta cluster use a database other than the wiki's local DB, meaning schema changes cannot be applied automatically anyway (related: T314908). On top of that, the extension is still marked as experimental and no-one else should be using it. We will use a simple boolean variable to control the migration, whose values correspond to the write_old_read_old and write_new_read_new stages. ==Acceptance criteria== [] Create the new table in the code, make the old columns DEFAULT NULL [] Add the migration logic to the code, so that it can read/write either schema. By default, it will read and write the old schema. [] Create the new schema in beta (task TBC) [] Configure beta to read/write the new schema [] Create the new schema in production (task TBC) [] Configure prod to read/write the new schema [] Drop the old columns in the code [] Remove the migration logic [] Drop the old columns in beta [] Remove the migration variable from the beta config [] Drop the old columns in prod [] Remove the migration variable from the prod config

As described in T320434#8741708, we should create a new table for storing tracking tools: ```lang=mysql CREATE TABLE ce_tracking_tools ( cett_id BIGINT UNSIGNED AUTO_INCREMENT NOT NULL, -- Surrogate primary key cett_event BIGINT UNSIGNED NOT NULL, -- Foreign key to campaign_events.event_id cett_tool_id INT NOT NULL, -- Identifier of the tool, same as the existing campaign_events.event_tracking_tool_id cett_tool_event_id BLOB NOT NULL, -- Identifier of the event within this tool, same as the existing campaign_events.event_tracking_tool_event_id cett_sync_status INT NOT NULL, -- Status of the sync, the meaning of each value is defined in the application cett_last_sync BINARY(14) NOT NULL -- Timestamp of the last synchronization ); ``` Also, `campaign_events.event_tracking_tool_id` and `campaign_events.event_tracking_tool_event_id` need to be dropped. Note that these fields can only be NULL for now. **Note**: this only has to be backwards-compatible in the sense that the code won't crash: the schema is still unused, so we don't really care what is being read/written. Also, we don't need all the files for the schema change: both production and the beta cluster use a database other than the wiki's local DB, meaning schema changes cannot be applied automatically anyway (related: T314908). On top of that, the extension is still marked as experimental and no-one else should be using it. We will use a simple boolean variable to control the migration, whose values correspond to the write_old_read_old and write_new_read_new stages. ==Acceptance criteria== [] Create the new table and drop the old columns in the code -- We don't need a backwards-compatible schema change for this: both production and the beta cluster use a database other than the wiki's local DBin the code, meaning schema changes cannot be applied automatically anyway (related: T314908) [] Adjust the code so that it reads from and writes to the new schemaake the old columns DEFAULT NULL -- This **needs** to be backwards-compatible[] Add the migration logic to the code, so that it can read/write either schema. By default, or things will break horriblyit will read and write the old schema. [] Create the new schema in beta (task TBC) [] Configure beta to read/write the new schema [] Create the new schema in production (task TBC) [] Remove back-compat codeConfigure prod to read/write the new schema [] Drop the old columns in the code [] Remove the migration logic [] Drop the old columns in beta [] Remove the migration variable from the beta config [] Drop the old columns in prod [] Remove the migration variable from the prod config