Summary: | Network process crash when decoding SecItemResponseData | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Chris Dumez <cdumez> | ||||
Component: | WebKit2 | Assignee: | Chris Dumez <cdumez> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | achristensen, beidson, commit-queue, ggaren, mitz, rniwa, timothy, webkit-bug-importer | ||||
Priority: | P2 | Keywords: | InRadar | ||||
Version: | WebKit Nightly Build | ||||||
Hardware: | Unspecified | ||||||
OS: | Unspecified | ||||||
Attachments: |
|
Description
Chris Dumez
2019-05-30 13:24:39 PDT
Created attachment 370975 [details]
Patch
Comment on attachment 370975 [details]
Patch
may as well
Comment on attachment 370975 [details] Patch Clearing flags on attachment: 370975 Committed r245911: <https://trac.webkit.org/changeset/245911> All reviewed patches have been landed. Closing bug. Does it really make sense for the length of the decoded array to be different from the length of the array that was encoded? (In reply to mitz from comment #6) > Does it really make sense for the length of the decoded array to be > different from the length of the array that was encoded? The issue is that we're not always able to encode elements, or decode them. E.g. The SecIdentityRef encoder will not encode a key if copyPersistentRef() return null. The SecIdentityRef decoder, will not construct a SecIdentityRef if not key was encoded, or if the decoding process does not have ProcessPrivilege::CanAccessCredentials privilege. Yet decode() will return true in those cases. In theory, the SecCertificateRef decoder can also return true while not constructing a SecCertificateRef, if SecCertificateCreateWithData() returns null. |