Chris Sewell

@n-peugnet I think in some way this should be linked with https://github.com/sphinx-doc/sphinx/pull/12492#issuecomment-2197798927, i.e. that it's made possible for the parser to determine how it handles parsing. I think this could...

Hi @rffontenelle can you confirm this ready for review/merge, as you've made a few changes since opening 😅 (best to set as a Draft until it is ready)

> I reopened the issue because it's an issue for me yeh no problem; I guess my generic question would be, how come there is currently no failure of our...

> Are there any tests for this? None seems to break with the change. Hey @jakobandersen cheers, but yes I feel we should look to add a MWE test that...

heya yep agree that this should be in sphinx and is useful, in fact I'm just dealing with a case now of moving from `

Heya, It would be great if you could open a PR 😄

@jayaddison I'm -1 on 72b266d4ef102028b735ba89b2773f2fdb668483; you've taken a simple PR and suddenly increased its complexity 10-fold, for minimal gain 😅

If you want to include `inv_name` in the message, sure, but the path to the `objects.inv` is overkill, and really an implementation detail that users should not be particularly concerned

> I think it's important to report the specific source Yes and reporting the inventory key `inv_name` is sufficient for this, I don't think `intersphinx_mapping[inv_name][1][0]` is even handling it correctly,...

Thanks @picnixz, perhaps also there should be a try/except for the imports in sphinx.util.testing, that adds a message for users to use sphinx[dev]