Fix mixed log using string concat#200
Conversation
a4d2e8b to
28320c3
Compare
28320c3 to
7e174c4
Compare
|
Thanks @aureplop ! Are you able to build the plugin and test it? I'll do a separate PR to clear up the CI issues. |
I will give it a try on Friday. Stay tuned 🥁 |
|
And so I succeed to test it locally. And this is working as expected 🎉 With current master, using a config with On the branch on this PR, with the same setup, I did not get any issues even with a hundred of reloads. I believe an (integration) test could be added, but it may be a bit boilerplate due to randomness (we could run the reload and check 100 times to mitigate) and it needs a change to the test config (have multiple worker_processes, currently it uses the default value of Do you need something more from me to merge this PR @cgilmour? Thanks! |
|
Right, it's hard to be certain if this resolves the problem or just makes it less likely to occur. Thanks for performing the test to confirm it resolved the issue in your case! |
cgilmour
left a comment
There was a problem hiding this comment.
Looks good and seems to resolve the reported problem!
|
Merged, thanks @aureplop |
|
Thanks! No emergency to release, we are using a env var to disable the startup logs causing parsing issues and this is good enough for now ;) |
(Should) fix #195.
I did not test it yet this version where we had the issues but it would be nice to check that before merging it, unless reviewers of this are already confident it will do the trick to fix the issue mentioned above.
Concatenation with "+" may not be the most performant thing but anyway this default logging function is very rarely used. I only saw the log in #195 using this format.
Also, are we ok that the log message will still be flushed to
stderrcorrectly, even if it's not usingstd::endl?Finally, I'm not sure to understand why the "build" step of CI is failing, but it also fails on other recent PR 🤔