Closed d-w-moore closed 2 months ago
does this explain it? even though the minimal example was encapsulated with a 'with' block?
Yes, I believe it does. If we're redirecting to "another host's" resource (in this case, cleverly renamed localhost), the stored session within the FileRaw object (f) is a cloned session with connections to the newly-targeted host. Which explains why the wrap-up of the with-block (ie the data object close operation) \didn't matter; the resource leak happened in the process of redirecting (during the data-object open operation) when I failed to release a still-running connection before grabbing a new one and stashing it into the same variable (conn
)
Excellent. Wrap it in a test and away we go. Consider putting some/all of your explanation in the commit message body as well.
Everything looks good to me, but there appears to be a conflict with one of the files. Please rebase and confirm things are working as expected, and I think we'll be good to go
Done.
Very good. # it
Done!
does this explain it? even though the minimal example was encapsulated with a 'with' block?