Fail decoding an invalid SessionID
Created attachment 376152 [details] Patch
<rdar://problem/54251062>
I kept a release assert for now when encoding a bad sessionID. Plan is to let it seat for a while and move it back to a regular ASSERT. I can change it back to a regular ASSERT() if that is thought as too crash risky though.
Comment on attachment 376152 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=376152&action=review > Source/WebCore/PAL/pal/SessionID.h:86 > + // FIXME: Change to a regular ASSERT. > + RELEASE_ASSERT(isValid()); I think we should just leave it a RELEASE_ASSERT
Comment on attachment 376152 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=376152&action=review > Source/WebCore/PAL/pal/SessionID.h:-107 > - // FIXME: We should fail to decode an invalid sessionID. I think we ought to return nullopt if isValid() is false.
Comment on attachment 376152 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=376152&action=review >> Source/WebCore/PAL/pal/SessionID.h:86 >> + RELEASE_ASSERT(isValid()); > > I think we should just leave it a RELEASE_ASSERT OK. >> Source/WebCore/PAL/pal/SessionID.h:-107 >> - // FIXME: We should fail to decode an invalid sessionID. > > I think we ought to return nullopt if isValid() is false. It is returning nullopt, see line above.
Comment on attachment 376152 [details] Patch Clearing flags on attachment: 376152 Committed r248663: <https://trac.webkit.org/changeset/248663>
All reviewed patches have been landed. Closing bug.