CI Fix w/o tm helper

[WillAyd]

Alternate to #31791

[jbrockmendel]

LGTM

[ShaharNaveh]

TBH I am -1 on that, as we are trying to get rid of all the bare pytest raises.
maybe the name of external_error_raised is not good, and needs a different (and better) name.

As far as I know we are planning to have a CI check for no bare pytest raises, obviously there are some exceptions such as an external error message, we need to agree on how to mark bare pytest raises that are bare on purpose.

A conversion on that topic has been started at #30999 (comment)

[WillAyd]

@momisbestfriend the problem is that helper isn’t backported so can’t be used to fix CI

[jreback]

@MomIsBestFriend the problem is that helper isn’t backported so can’t be used to fix CI

we can simply backport the helper, no?

[WillAyd]

Sure that would be an alternative

[WillAyd]

Looks like the CI failures may have resolved themselves on the 1.0.x branch, so I guess no need to do this. However, should probably revert original PR then - @jbrockmendel do you agree? Something you have time to tackle?

[jbrockmendel]

Something you have time to tackle?

I'm stretched pretty thin ATM. Won't hitting the green button here get the job done?

[WillAyd]

Closing in favor of #31931