[1] made them the same.
[2] made them different again, to fix a regression introduced by [1] ( Issue 623616 ).
It is preferred to make them the same again.
It seems like that would require splitting willSendRequest up into
a prepareRequest step and a sendingTheRequestForRealNow step. (a comment by japhet@ from [2])
[1] https://codereview.chromium.org/1889973002/
[2] https://codereview.chromium.org/2184823006/
Cc: -hirosh...@chromium.org japhet@chromium.org Owner: hirosh...@chromium.org Status: Started (was: Assigned)
Restarted.
I'll also try to clarify what kind of modifications are made to ResourceRequest, after FetchRequest is created, by e.g. removing FetchRequest::mutableResourceRequest() callers.
To clarify what kind of modifications are made to ResourceRequest after FetchRequest is created, I'll reduce non-const usage of FetchRequest::m_resourceRequest, including mutableResourceRequest().
I expect most of them can be avoided.
After I remove unnecessary modifications via FetchRequest::m_resourceRequest or mutableResourceRequest(), all mutableResourceRequest() will represent necessary modifications to ResourceRequest after FetchRequest is created.
Comment 1 by bugdroid1@chromium.org
, Jul 29 2016