Keep CFDataRefs in SharedBuffer instead of merging them
Created attachment 232919 [details] Patch
Comment on attachment 232919 [details] Patch There is a case for (count == 1) above the modified code, should it be updated in the same way too? r=me assuming that there is a measurable improvement somewhere.
> There is a case for (count == 1) above the modified code, should it be updated in the same way too? No, I misread that code.
(In reply to comment #2) > (From update of attachment 232919 [details]) > r=me assuming that there is a measurable improvement somewhere. Thanks for the review. By itself this will not improve performance but next up I would like to pass the SharedBuffer all the way down to IPC message send so that we can avoid merging the CFData and directly write to the IPC message.
Comment on attachment 232919 [details] Patch Clearing flags on attachment: 232919 Committed r169868: <http://trac.webkit.org/changeset/169868>
All reviewed patches have been landed. Closing bug.