-
-
Notifications
You must be signed in to change notification settings - Fork 34
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Close Multi timers atexit. Add 1.6 CI #234
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There should only be one function registered, since you cannot deregister them, this PR otherwise would create a memory leak
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## master #234 +/- ##
==========================================
- Coverage 90.62% 89.70% -0.92%
==========================================
Files 5 5
Lines 576 583 +7
==========================================
+ Hits 522 523 +1
- Misses 54 60 +6 ☔ View full report in Codecov by Sentry. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Isn't there already a global list of these elsewhere
Not that I'm aware of, no. |
How does it reuse connections then if there is no list of the |
Should |
The Multi constructor already calls Downloads.jl/src/Curl/Multi.jl Line 10 in 82b8da5
I guess the finalizer isn't being called before the trailing task warning? |
Connections are only reused within a |
Ah, it sounds like there may need to be a global weakref list of active Downloader objects then? |
I see no warning with this locally now. |
Yes, I think that Ian was trying to do this with a global (non-weak) list of timers to close. You think that doing a global weakref list of downloaders is a better approach? That always feels like such a failure of the programming language... |
This reverts commit 79fbe38.
fbaaa9b
to
ed67afe
Compare
@vtjnash I just noticed that on 1.10.0 this doesn't avoid the IO warnings, but it does on master. Has when |
Ah JuliaLang/julia#51849 so this isn't a fix for 1.10 Indeed, CI shows the warning on 1.10 |
@StefanKarpinski I guess it is actually worth dropping 1.3 because of these windows CI issues? |
Sure, go ahead and drop it. Not worth the effort anymore. |
Should be ready now |
Fixes #223