RESOLVED FIXED 231000
Add weakThis check in addition to null check added in r282881
https://bugs.webkit.org/show_bug.cgi?id=231000
Summary Add weakThis check in addition to null check added in r282881
Alex Christensen
Reported 2021-09-29 17:18:45 PDT
Remove log added in r282881
Attachments
Patch (2.11 KB, patch)
2021-09-29 17:27 PDT, Alex Christensen
no flags
Patch (2.15 KB, patch)
2021-09-29 17:32 PDT, Alex Christensen
no flags
Alex Christensen
Comment 1 2021-09-29 17:27:37 PDT
Alex Christensen
Comment 2 2021-09-29 17:32:35 PDT
Chris Dumez
Comment 3 2021-09-29 17:34:37 PDT
Comment on attachment 439686 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=439686&action=review > Source/WebKit/NetworkProcess/NetworkResourceLoader.cpp:353 > LOADER_RELEASE_LOG("startNetworkLoad: Going to the network (description=%" PUBLIC_LOG_STRING ")", m_networkLoad->description().utf8().data()); What does m_networkLoad->description() return? Hopefully it doesn't contain anything sensitive like URL? Also, why can't we move this LOG line *before* calling start()?
Brady Eidson
Comment 4 2021-09-29 17:49:55 PDT
Comment on attachment 439686 [details] Patch This patch isn't about verifying the string returns the right thing. Putting the log line before start() makes it categorically false, as start() may not actually *start*
Chris Dumez
Comment 5 2021-09-29 18:37:30 PDT
(In reply to Brady Eidson from comment #4) > Comment on attachment 439686 [details] > Patch > > This patch isn't about verifying the string returns the right thing. And yet still a good thing to verify and you didn't answer. > Putting the log line before start() makes it categorically false, as start() > may not actually *start* I feel like a safer fix would have been to update the actual code that introduced the regression and have NetworkLoad::start() call didCompleteWithError: asynchronously when it fails due to FTP. There are other clients to NetworkLoad than NetworkResourceLoader and having start() call that may fail synchronously and delete |this| might be an issue for other call sites as well.
Chris Dumez
Comment 6 2021-09-29 18:41:37 PDT
(In reply to Chris Dumez from comment #5) > (In reply to Brady Eidson from comment #4) > > Comment on attachment 439686 [details] > > Patch > > > > This patch isn't about verifying the string returns the right thing. > > And yet still a good thing to verify and you didn't answer. 0x105d74000 - [pageProxyID=21, webPageID=22, frameID=3, resourceID=32, isMainResource=1, destination=3, isSynchronous=0] NetworkResourceLoader::startNetworkLoad: Going to the network (description=LocalDataTask <25187A1B-3324-4E2C-982A-AB468520286C>.<1>) Now I have my answer.
EWS
Comment 7 2021-09-29 19:01:44 PDT
Committed r283294 (242319@main): <https://commits.webkit.org/242319@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 439686 [details].
Radar WebKit Bug Importer
Comment 8 2021-09-29 19:02:16 PDT
Brent Fulgham
Comment 9 2022-02-12 20:22:16 PST
*** Bug 230767 has been marked as a duplicate of this bug. ***
Note You need to log in before you can comment on or make changes to this bug.