Add supervisor:which_child/2
#8976
Conversation
CT Test Results 2 files 96 suites 1h 7m 17s ⏱️ Results for commit a720ea9. ♻️ This comment has been updated with latest results. To speed up review, make sure that you have read Contributing to Erlang/OTP and that all checks pass. See the TESTING and DEVELOPMENT HowTo guides for details about how to run test locally. Artifacts// Erlang/OTP Github Action Bot |
Maria-12648430
force-pushed
the
supervisor_which_child
branch
2 times, most recently
from
f76cb7b
to
a9a9a78
Compare
|
I found myself needing a similar functional several times. 👍 |
|
@Maria-12648430 I think it looks useful :) will bring it up on our PR-meeting today. |
|
@Maria-12648430 we decided this looks good would you please add test case too. |
Maria-12648430
force-pushed
the
supervisor_which_child
branch
from
a9a9a78
to
05b9b22
Compare
@IngelaAndin done 👍 |
IngelaAndin
added
the
testing
| which_children(Config) when is_list(Config) -> | ||
| {ok, SupPid} = start_link({ok, {#{}, []}}), | ||
|
|
||
| [] = supervisor:which_children(SupPid), |
There was a problem hiding this comment.
is this the 99% same test code in 2 tests? maybe consider having a helper function for easier maintenance in future - so that redundant code is reduced.
There was a problem hiding this comment.
tl;dr, yes it is 😅
I can do it as you suggest and will readily yield to any such decision.
However, personally, I would argue that those tests test behaviors which are merely very similar, but not the same, and that this fact justifies two distinct fixed tests, even if they are also very similar to each other.
But you decide 😉
There was a problem hiding this comment.
I think it is a tradeoff. Sometimes it could be harder to understand the test case if you sort or loose the context.
It could of course be compensated with good variable naming's. I am not sure what I would think in this case without an example of how the helper should look.
There was a problem hiding this comment.
I was tired when doing review. My bad. After re-reading the code, I don't think it is worth changing.
Sorry for confusion. Pls, leave it as it is.
There was a problem hiding this comment.
@u3s np, will do 🤗
There was a problem hiding this comment.
That said, I just augmented the which_children_simple_one_for_one test a little, to check that giving a non-pid to a simple_one_for_one supervisor results in {error, simple_one_for_one}.
Maria-12648430
force-pushed
the
supervisor_which_child
branch
from
05b9b22
to
a720ea9
Compare
|
Ok, just enabled this branch to be tested again, it gets automatically commented out when you make a new push. This unfortunately is not automatically mirrored by the label. |
Sometimes I have the need to obtain the Pid of a certain supervisor child. This often happens when the children form a work unit of cooperating processes, wherein one needs to talk to another.
To obtain such a Pid, currently I have to use
which_children/1to obtain a list of all children, then filter out the one I need, which is boilerplate code repeated over and over.This PR adds a function
which_child/2, which can be used to obtain the same information aswhich_children/1does, but for a particular single child.