1757040 - Provide add_setup which interacts correctly with `.only` and `.skip` in xpcshell tests

Status: RESOLVED FIXED

(Testing :: XPCShell Harness, defect, P3)

Description

[:Gijs (he/him)]

+++ This bug was initially created as a clone of Bug #1720335 +++

Although we have registerCleanupFunction, we do not have an equivalent for test setup.

Right now .only() doesn't always work right, because it doesn't take into account setup tasks as, well, required setup.

If we had appropriate setup functions, this would work right.

---

bug 1720335 did this for browser mochitests; let's do the same for xpcshell.

Comment 1

[:Gijs (he/him)]

Attachment: Bug 1757040 - add add_setup() primitive to xpcshell test framework, r?standard8

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/64ef1f73d033
add add_setup() primitive to xpcshell test framework, r=Standard8

Comment 3

[Cristian Tuns]

https://hg.mozilla.org/mozilla-central/rev/64ef1f73d033

Comment 4

[Nick Alexander :nalexander [he/him]]

I just discovered that multiple add_setup invocations execute in reverse order. That is:

add_setup(a);
add_setup(b);

invokes as b(); a(). That surprised me. Should we either disallow multiple add_setup calls, or fix the ordering (and test that it stays consistent)?

Comment 5

[:Gijs (he/him)]

(In reply to Nick Alexander :nalexander [he/him] from comment #4)

I just discovered that multiple add_setup invocations execute in reverse order. That is:

add_setup(a);
add_setup(b);

invokes as b(); a(). That surprised me. Should we either disallow multiple add_setup calls, or fix the ordering (and test that it stays consistent)?

We should fix the ordering. Please file a follow-up.

Is this happening for xpcshell, for browser mochitests, or both? (The implementations are different 😭)

Comment 6

[Nick Alexander :nalexander [he/him]]

(In reply to :Gijs (he/him) from comment #5)

(In reply to Nick Alexander :nalexander [he/him] from comment #4)

I just discovered that multiple add_setup invocations execute in reverse order. That is:

add_setup(a);
add_setup(b);

invokes as b(); a(). That surprised me. Should we either disallow multiple add_setup calls, or fix the ordering (and test that it stays consistent)?

We should fix the ordering. Please file a follow-up.

Is this happening for xpcshell, for browser mochitests, or both? (The implementations are different 😭)

Filed and fixed by :gijs. Huzzah!