Looking at my previous comment, instead of just services we instead would want to note that it is the ObjectFactory spec, not just services.

@daniel I may be completely missing something. but doesn't a handler have access to its full path via Handler::$request? Could we use something like $this->getRequest()->getUri()->getPath(); in the PageHistoryHandler?

@eprodromou question: I'm still a little confused by the "segment" param, the syntax is unclear to me by how it is written in the description. (Using timestamp below as an example)

Actually! Just remembered @tstarling has this change in patch already out (but not yet merged).

Unsure if this was intentionally omitted to prevent people from using it until the new Hooks system is fully released, but I think we need to just add HookHandlers to extension.schema.v2.json.

It looks really good to me!! Couple of ideas:

@Aklapper Appreciate it !

Renaming this would break some of the logic that generates the hook function names. A proposed alternative by @tstarling was to introduce hook aliases instead of actually renaming the hook itself. This would also allow us to alias some of the hooks that have double colons in the name (e.g. ApiMain::moduleManager) which as of now is converting the colons into underscores (e.g. ApiMain__moduleManager) which looks slightly awkward.

Checked to see if this was still occurring, and last logged instance was 6 occurrences between 2020-04-03T10:35:35 and 2020-04-03T10:37:36

Talked with @Anomie and he brought up a good point that it may be worth it to implement PoolCounter for all requests, regardless of whether or not it bypasses ParserCache. That would lead to less complicated/case-by-case code and even if the request does hit ParserCache it may still be a cache miss and it would have to do the full slower parse anyway.

Untagging Core Platform. Tracking work for this in T226857.

@Anomie Yea sounds good!

@Peter.ovchyn ok :) I spent half of yesterday trying to re-produce locally so i could test some fixes and was unable to, even on the correct branches everything was passing. The only thing I did notice was that was another extension is reporting the same error (https://integration.wikimedia.org/ci/job/quibble-vendor-mysql-php72-docker/50811/console)

Nice, ok lets see re-check these tests once that gets a +2 ?

Closing this ticket as "declined" as we have removed the need to add spdx ID for now.