Fix nullptr dereference introduced in r268228
Created attachment 417983 [details] Patch
Comment on attachment 417983 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=417983&action=review > Source/WebCore/fileapi/Blob.cpp:274 > + if (bytesLoaded == m_loader->totalBytes()&& !m_bytesRead) nit: missing space before &&
Comment on attachment 417983 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=417983&action=review >> Source/WebCore/fileapi/Blob.cpp:274 >> + if (bytesLoaded == m_loader->totalBytes()&& !m_bytesRead) > > nit: missing space before && Also may want to use FileReaderLoader::isCompleted() for clarity.
Created attachment 417984 [details] Patch
Committed r271669: <https://trac.webkit.org/changeset/271669> All reviewed patches have been landed. Closing bug and clearing flags on attachment 417984 [details].
<rdar://problem/73417187>
This was rdar://problem/71627170
It is unfortunate that Safari 14.1 shipped with Blob.stream() but without this fix. Our web app, wormhole.app, uses feature detection to determine if Blob.stream() is available, so it broke when Safari 14.1 was released. We had to quickly add a patch to disable using Blob.stream() in Safari, since it would immediately crash the renderer process. If the feature was known to be buggy, I wish it hadn't shipped, or that the fix had been applied before shipping it. We did test in Safari Technology Preview, but it already had this fix applied so we didn't encounter the crash.
(In reply to John Hiesey from comment #8) > It is unfortunate that Safari 14.1 shipped with Blob.stream() but without > this fix. > > Our web app, wormhole.app, uses feature detection to determine if > Blob.stream() is available, so it broke when Safari 14.1 was released. We > had to quickly add a patch to disable using Blob.stream() in Safari, since > it would immediately crash the renderer process. > > If the feature was known to be buggy, I wish it hadn't shipped, or that the > fix had been applied before shipping it. > > We did test in Safari Technology Preview, but it already had this fix > applied so we didn't encounter the crash. Thanks for bringing this to our attention. We'll look into what happened.