Add Clang ReleaseNotes that --autocomplete breaks backward compatibily

Summary:
--autocomplete flag now handles all the flags passed to shell, and this
implementation breaks backward compatibily before Clang 6.0.

Reviewers: teemperor, v.g.vassilev

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D44191

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@326889 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/docs/ReleaseNotes.rst b/docs/ReleaseNotes.rst
index b29cd0627d05c85bd3563bbf8ffba5aea7b514e3..bb645e14fdd5d1a1334a19002fca3c4310e4a9b4 100644 (file)
--- a/docs/ReleaseNotes.rst
+++ b/docs/ReleaseNotes.rst
@@ -71,6 +71,15 @@ future versions of Clang.
 
 - ...
 
+Modified Compiler Flags
+-----------------------
+
+- Before Clang 7.0, we prepended the "#" character to the --autocomplete argument to
+enable cc1 flags. For example, when the -cc1 or -Xclang flag is in the clang invocation,
+the shell executed clang --autocomplete=#-<flag to be completed>. Clang 7.0 now
+requires the whole invocation including all flags to be passed to the --autocomplete
+like this: clang --autocomplete=-cc1,-xc++,-fsyn.
+
 New Pragmas in Clang
 -----------------------