Page MenuHomePhabricator

[SPIKE] Investigate feasibility of revising Flow migration script
Closed, ResolvedPublic

Description

Volunteers at Gomwiki (Konkani) proposed two changes to the Flow migration script scheduled to run this week, via T380911:

  1. Archive-destination pagename: They want "/Archive [number]" instead of "/Flow"
  2. Deletion instead: 668 of the 833 Flow boards on gomwiki are completely empty - can we please delete those instead? - See listing in P71749

This task involves assessing the technical feasibility of making the above changes to the script that will run at Gomwiki.

Decision(s) to be made

  • What – if any – changes will we make to the Flow migration script prior to running it at Phase 2b wikis?

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

First change should be trivial. Second change is probably best done by passing that list to something like deleteBatch.php before running the FlowMoveBoardsToSubpages script.

Looks like it's a parameter to the script already: add --subpage Archive to name things Archive. If we want numbering, that might be another matter. Perhaps we could get some dry-run output to see what the current script would do?

Numbering could be done with a tweak -- the findValidSubpage method in the script already appends numbers to get to a not-yet-existing page. It could watch out for cases where the subpage already ends in a number and increment that rather than appending a new number. Then we could run the script with --subpage "Archive 1" as the input and it'd notice and increment.

Main limitation would be that it currently just tries three times and then gives up, so if there's already an Archive 3 page out there we'd be skipping that page-move. We could either increment the number of tries -- 3 was entirely arbitrary -- or manually handle whatever tiny number of moves that represents.

Change #1128932 had a related patch set uploaded (by Zoe; author: Zoe):

[mediawiki/extensions/Flow@master] Added an option to always number archived pages

https://gerrit.wikimedia.org/r/1128932

What – if any – changes will we make to the Flow migration script prior to running it at Phase 2b wikis?

Per today's offline discussion, we're going to exclude gomwiki from tomorrow's running of the Flow migration script.

Reason: this way, we can test the adjustments we've made to the script to meet gomwiki needs before running the script in production.

To put the above into effect, I've:

  • Created T389247 where the work of running the Flow migration script at gomwiki will happen
  • Updated the T380911

Main limitation would be that it currently just tries three times and then gives up, so if there's already an Archive 3 page out there we'd be skipping that page-move. We could either increment the number of tries -- 3 was entirely arbitrary -- or manually handle whatever tiny number of moves that represents.

It seems that there are just two 'Archive 2' pages on gomwiki, and no 'Archive 3' pages. So we should not face a problem due to this.

Change #1130569 had a related patch set uploaded (by Zoe; author: Zoe):

[mediawiki/extensions/Flow@master] Use builtin getTalkpageManager to move Flow boards

https://gerrit.wikimedia.org/r/1130569

zoe added a project: Patch-For-Review.

oops – still had one patch that was here. the getTalkpageManager patch I've moved to another ticket

Change #1128932 merged by jenkins-bot:

[mediawiki/extensions/Flow@master] Added an option to always number archived pages

https://gerrit.wikimedia.org/r/1128932