Loosen assertion so it matches any production error #8907
Conversation
|
I would assume it means it is crashing for a different reason. |
The exact error code isn't important.
|
It was because the message was slightly different, which I fixed in 2afcf6e I figured loosening the invariant made sense, anyway. Although after rebasing, now I'm having a different problem where the error code transform isn't picking up that invariant, but works for other invariants in the same file. Weird. |
|
Ah, the message is the same as Stack but the template string passed to Also, we haven't run |
|
I guess it happens before release so it's not a big deal. (Or does it...?) |
|
Yea, it's in the release guide so shouldn't be necessary. |
The production error code system depends on the exact template string that is passed to `invariant`. This changes the Fiber templates to match the Stack ones.
acdlite
force-pushed
the
fiberproderrorcode
branch
from
0936484
to
ea2aa6e
Compare
|
@spicyj Going to assume this is good to merge now that we know why the codes were different (they're the same now), since you approved it earlier. |
|
I don't think we should loosen this unnecessarily – even if format changes are okay, problems building the URL or similar could surface and be caught by the old test. Can we revert this? (If I rescind my acceptance then please don't consider a PR accepted.) |
|
Ok I'll revert it and open a separate PR without the change to the test. |
|
And sorry, I didn't intentionally ignore your review... I thought it was ambiguous whether you had "rescinded" it or not, since there were two reviews, hence my comment :D |
|
Latest review wins. :) |
This test was failing because the error code didn't match. The specific error code isn't important.