Adding retain/release method to the client protocol and use them to protect the client from disposal.
Created attachment 331779 [details] patch
Comment on attachment 331779 [details] patch View in context: https://bugs.webkit.org/attachment.cgi?id=331779&action=review > Source/WebCore/platform/network/curl/CurlRequestClient.h:39 > + virtual void retain() = 0; > + virtual void release() = 0; Can these be called ref and deref? Then we can just make RefPtr's instead of manually calling retain and release. We do something similar in several places, like IDBConnectionToServerDelegate, IDBConnectionToClientDelegate, CSSFontFace::Client, CSSRuleList, CSSStyleDeclaration, and ServiceWorkerJobClient
(In reply to Alex Christensen from comment #2) > Comment on attachment 331779 [details] > patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=331779&action=review > > > Source/WebCore/platform/network/curl/CurlRequestClient.h:39 > > + virtual void retain() = 0; > > + virtual void release() = 0; > > Can these be called ref and deref? Then we can just make RefPtr's instead > of manually calling retain and release. > We do something similar in several places, like > IDBConnectionToServerDelegate, IDBConnectionToClientDelegate, > CSSFontFace::Client, CSSRuleList, CSSStyleDeclaration, and > ServiceWorkerJobClient Okay, thanks. We'll take a look into those.
Created attachment 332065 [details] use ref/deref
The commit-queue encountered the following flaky tests while processing attachment 332065 [details]: js/arity-mismatch-at-vmentry.html bug 182014 (authors: cdumez@apple.com and msaboff@apple.com) The commit-queue is continuing to process your patch.
Comment on attachment 332065 [details] use ref/deref Clearing flags on attachment: 332065 Committed r227449: <https://trac.webkit.org/changeset/227449>
All reviewed patches have been landed. Closing bug.
<rdar://problem/36794134>