[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 184463 abandoned by Aaron Schulz: Introduced LazyDataUpdate class Reason: Wikibase no longer uses addSecondaryDataUpdates() https://gerrit.wikimedia.org/r/184463 TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to comment o

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

ori added a comment. @daniel, thanks for clarifying. Could you also explain (in practical terms) what are the updates being performed? TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to comment or attach files, or !close, !claim, !unsubscribe or !assign .

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

daniel added a comment. @ori: simply be cause there is currently no mechanism (as far as I know) for extensions to provide deferred updated. Only some specific hard-coded updates are deferred, and everything else is executed right away. DataUpdate itself could somehow indicate whether it wants

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

ori added a comment. @daniel Can you explain why this work needs to be done synchronously, blocking the user? TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to comment or attach files, or !close, !claim, !unsubscribe or !assign . EMAIL PREFERENCES http

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

aaron added a comment. @daniel: what do you think of https://gerrit.wikimedia.org/r/184463. Would be hard to build off? I'd really like to get stashing on again, since it saves 300+ms on average saves. TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to com

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 184463 had a related patch set uploaded (by Aaron Schulz): Introduced LazyDataUpdate class https://gerrit.wikimedia.org/r/184463 https://phabricator.wikimedia.org/tag/patch-for-review/ TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIO

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 183887 abandoned by Aaron Schulz: Made some DataUpdate classes serializable Reason: Doing this in a cleaner way https://gerrit.wikimedia.org/r/183887 TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to comment or attach fil

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 183887 had a related patch set uploaded (by Aaron Schulz): Made some DataUpdate classes serializable https://gerrit.wikimedia.org/r/183887 https://phabricator.wikimedia.org/tag/patch-for-review/ TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HAN

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 183890 merged by jenkins-bot: Fix ApiStashEdit wrt custom DataUpdates. https://gerrit.wikimedia.org/r/183890 TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to comment or attach files, or !close, !claim, !unsubscribe or !as

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

daniel added a comment. @anomie: adding the guarantee that DataUpdates must be serializable would make the mechanism a lot more complex (right now, it's an easy way to provide callbacks). I submitted a patch that uses a flag in ParserOutput - that could be extended to only trigger on non-serial

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 183890 had a related patch set uploaded (by Daniel Kinzler): Fix ApiStashEdit wrt custom DataUpdates. https://gerrit.wikimedia.org/r/183890 https://phabricator.wikimedia.org/tag/patch-for-review/ TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HA

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

Anomie added a comment. > Since DataUpdate objects are generally not serializable, ParserOutput skips > them during serialization (see ParserOutput::__sleep). I don't know about "are generally not serializable", but the reasoning given both in the comment and in Gerrit change 119342

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

gerritbot added a comment. Change 183861 merged by jenkins-bot: Skip ApiStashEdit if custom DataUpdates are present. https://gerrit.wikimedia.org/r/183861 TASK DETAIL https://phabricator.wikimedia.org/T86305 REPLY HANDLER ACTIONS Reply to comment or attach files, or !close, !claim, !unsubs

[Wikidata-bugs] [Maniphest] [Commented On] T86305: ApiStashEdit silently loses custom DataUpdates.

daniel added a comment. The only fix I can think of offhand would be to check if there are custom updates, and if there are, don't stash the edit. That would only happen after burning CPU cycles for parsing though. Note that this would effectively disable the use of ApiStashEdit for all wikis/