net: return if not _connecting#2251
Conversation
evanlucas
added
confirmed-bug
There was a problem hiding this comment.
Can you use common.PORT and common.localhostIPv4 for the port and host here.
|
Two small nits. Can you also come up with a slightly more descriptive first line of the commit message? Other than that, LGTM. |
evanlucas
force-pushed
the
fixconnectingdestroy
branch
from
bfc3ad1 to
0ef648e
Compare
|
Ok, updated. I tried to start a CI job on it, but it seems to not be working. |
|
Weird...looks like the current build has been running for 2 days now? ping @Fishrock123 |
|
^ ping @nodejs/build :) |
|
LGTM if the tests are happy. Tagged the original issue with |
Looking into it... one ubuntu slave is stuck |
lib/net.js
Outdated
There was a problem hiding this comment.
Just an observation. We can actually avoid the empty statement here if we flip the condition
There was a problem hiding this comment.
Yea. We could go the route @cjihrig went just use if (self._connecting) ...
There was a problem hiding this comment.
Oh yeah, just checked the closed PR :D
evanlucas
force-pushed
the
fixconnectingdestroy
branch
from
0ef648e to
b516a61
Compare
|
Will land after CI finishes provided it is happy. |
|
LGTMT 👍 |
|
@evanlucas CI looks good. |
Fixes regression introduced in af249fa. With connect being deferred to the next tick, Socket.destroy could be called before connect. Socket.destroy sets _connecting to false which would cause an assertion error. Fixes: nodejs#2250 PR-URL: nodejs#2251 Reviewed-By: Colin Ihrig <cjihrig@gmail.com> Reviewed-By: Jeremiah Senkpiel <fishrock123@rocketmail.com> Reviewed-By: Sakthipriyan Vairamani <thechargingvolcano@gmail.com>
evanlucas
force-pushed
the
fixconnectingdestroy
branch
from
b516a61 to
503b089
Compare
5 participants
Fixes regression introduced in af249fa.
With
connectbeing deferred to the next tick,Socket.destroycould becalled before
connect.Socket.destroysets _connecting to false whichwould cause an assertion error.
Fixes: #2250