test: https server close event #5106
Conversation
mscdex
added
https
|
The commit should target the |
braydonf
force-pushed
the
server-close
branch
from
d1d7854
to
094f507
Compare
braydonf
changed the title
|
|
||
| function shutdown() { | ||
| server.close(function() { | ||
| serverCloseEventCalled = true; |
There was a problem hiding this comment.
Nit: Instead of using the serverCloseEventCalled boolean, use common.mustCall() on the close callback.
braydonf
force-pushed
the
server-close
branch
from
094f507
to
f183608
Compare
| console.log('1..0 # Skipped: missing crypto'); | ||
| return; | ||
| } | ||
| var https = require('https'); |
|
LGTM if CI is green. Few nits tho. |
|
@jasnell This is a test that needs a bugfix. It is intended to fail. It should not land until there is a bug fix (assuming this is a bug and not expected behavior being misconstrued as a bug, but it looks like a genuine bug to me). Technically, I guess this shouldn't be a pull request and should instead be put in #5083. But the folks reporting this have done such a great job tracking down everything relevant to their bug report that I'm not inclined to complain. Anyway, with that context, I guess we can look at the all-red CI as a win. (Well, except the lint warning which says that |
|
For a bug fix, restoring the server property on socket fixes this particular test. This can be done by removing lines (381 and 382) in if (socket && socket.server === this.server)
socket.server = null;However these two tests will not pass:
|
|
Also, yep the goal with opening as a PR is to have CI run and having the failing tests (so we can fix). We can put a WIP on this if that's necessary. |
braydonf
force-pushed
the
server-close
branch
from
a124228
to
2d3b2c0
Compare
braydonf
changed the title
|
Bug fix found, all tests passing locally. |
braydonf
force-pushed
the
server-close
branch
from
2d3b2c0
to
88e8dd4
Compare
|
CI looks good. (Just one known flaky test that should be fixed when #5154 lands.) LGTM but sign-off from someone who has spent some quality time with |
mscdex
added
tls
|
Sorry, just found this after proposing #5262 . Please let me know if you still want to pursue this PR's approach. |
|
I ran the test included in this PR with the fix in #5262 and everything passed, perhaps we should include this test also. |
|
@braydonf sounds like a good idea, may I ask you to strip all other changes from this PR then? |
braydonf
force-pushed
the
server-close
branch
from
5f289e3
to
630eb1a
Compare
|
Okay rebased with only the test included. Made a backup of the branch with other changes, for reference. |
braydonf
changed the title
|
CI is green except one unrelated failure. |
|
LGTM |
|
Landing. |
|
LGTM |
|
Landed in 210e65a, thank you! |
PR-URL: #5106 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Rich Trott <[email protected]> Reviewed-By: Fedor Indutny <[email protected]>
PR-URL: #5106 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Rich Trott <[email protected]> Reviewed-By: Fedor Indutny <[email protected]>
|
This test is failing in v4.x-staging 😭 |
|
@thealphanerd I think this test depends on 7885b1d so if that didn't land on v4.x-staging, that might explain it. |
|
7885b1d just landed and this is now landing cleanly. Thank you as always for your swift and accurate response |
PR-URL: #5106 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Rich Trott <[email protected]> Reviewed-By: Fedor Indutny <[email protected]>
PR-URL: #5106 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Rich Trott <[email protected]> Reviewed-By: Fedor Indutny <[email protected]>
PR-URL: #5106 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Rich Trott <[email protected]> Reviewed-By: Fedor Indutny <[email protected]>
Closes: #5083