2062403 - Remove 29-character "subcategory" restriction in glean_parser

Status: RESOLVED FIXED

(Data Platform and Tools :: Glean: SDK, enhancement, P1)

Description

[Chris H-C :chutten]

In metrics-2.0 we have a max character limit on category of 40. We also have a regex which limits "subcategory" (e.g., subcategory_one.subcategory_two) to 29 characters. BigQuery doesn't require this. The Glean SDK doesn't require this. None of the Glean SDK's supported programming languages require this.

It may be that this pattern was introduced to discourage long subcategory names for some lost-to-time reason of naming quality. The implementing bug 1507788 references but does not summarize a conversation in a meeting. But we weren't being as careful about that in 2018 as we are now, so we can't really blame anyone (also, both involved persons are no longer contributing to Glean).

Nowadays, if we want to encourage or discourage a pattern via glean_parser, we add a lint for it. That way there's an unsafety valve of no_lint for folks to assert that they understand what we mean and know that what they're doing is despite that guidance. (I don't think the glinter was available in 2018, so another reason we can't blame folks for this.)

This bug is about either moving the subcategory restriction to a lint so it takes its more useful place as a recommendation, or removing it altogether. I'm leaning towards the latter.

Comment 1

[BMO Github Automation]

Attachment: [mozilla/glean_parser] bug 2062403 - Allow 'subcategories' longer than 29 characters (#858)

Comment 2

[Chris H-C :chutten]

PR merged. The next version of glean_parser will release with the restriction lifted.

Comment 3

[isabella]

Amazing, thank you for looking into this and lifting the restriction! 🙏