[Webkit-unassigned] [Bug 164631] REGRESSION: API test _WKDownload.ConvertResponseToDownload is a flaky timeout

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=164631

--- Comment #6 from Carlos Garcia Campos <cgarcia at igalia.com> ---
But r208521 shouldn't change the behavior for non NetworkSession code path. Maybe the change in NetworkResourceLoader. It could be that the NetworkLoad is deleted earlier now for the non NetworkSession case. When the load is converted to download, we now transfer the NetworkLoad to the download manager that creates the Download. In the case of !NetworkSession the download manager doesn't adopt the network load, so it's deleted after convertNetworkLoadToDownload. After the load is converted to download the web process deletes the resource loader and sends RemoveLoadIdentifier that calls abort on the NetworkResourceLoader. Since the load was converted to download, the abort simply does the cleanup, that used to delete the network load in that case. I don't see how this change would affect the API, though. If the download is cancelled after the network load is deleted, the download manager already has a Download object to cancel it. Maybe the download finishes before the NetworkProcess:

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20161112/1c408294/attachment.html>