[PATCH] D155445: [analyzer][docs] Add CSA release notes

[Balázs Benics via Phabricator via cfe-commits]

steakhal added inline comments.


================
Comment at: clang/docs/ReleaseNotes.rst:922-923
+- The ``CStringChecker`` will invalidate less if the copy operation is
+  inferable to be bounded. For example, if the argument of ``strcpy`` is known
+  to be of certain length and that is in-bounds.
+
----------------
OikawaKirie wrote:
> The lengths of both src and dst buffers need to be known.
Applied!


================
Comment at: clang/docs/ReleaseNotes.rst:937
+
+  Similarly, functions like ``strsep`` now won't invalidate the source buffer,
+  because it can never overflow.
----------------
OikawaKirie wrote:
> I think this may be a typo here, as we do not invalidate the source buffer originally.
Exactly. Thanks!


================
Comment at: clang/docs/ReleaseNotes.rst:908
+  (`7cd1f3ad22e4 <https://github.com/llvm/llvm-project/commit/7cd1f3ad22e4>`_)
+- Fixed a null-pointer dereference crash inside the ``MoveChecker``.
+  (`d172b65ef001 <https://github.com/llvm/llvm-project/commit/d172b65ef001>`_)
----------------
xazax.hun wrote:
> steakhal wrote:
> > xazax.hun wrote:
> > > I think we usually do not mention crash fixes in the changelog. We have them in almost every release and sometimes there are quite a few of them.
> > I won't mention the explicit commit where it was fixed.
> > However, downstream users might wanna know about crashes and fixes that happened in this release.
> > And speaking about past practices about release notes, I think we can improve on that TBH.
> > We can move it down on the list if you want, but I'd rather keep it.
> Is this the only crash fix we had? Moving crash fixes to the bottom of the list sounds good to me. 
No, it wasn't. We also had one for init-expr global variable initializers. [[ https://github.com/llvm/llvm-project/commit/558b46fde2db | See ]] 
I swept that fix under the carpet of "Fixed some bugs around the handling of constant global arrays and their initializer expressions". I made it more explicit now.

However, at this point, I think it's okay to simply omit the mention of the null deref crash fix.
Second thoughts?


================
Comment at: clang/docs/ReleaseNotes.rst:920-922
+- The ``CStringChecker`` will invalidate less if the copy operation is bounded.
+  (`1bd2d335b649 <https://github.com/llvm/llvm-project/commit/1bd2d335b649>`_)
+  (`#55019 <https://github.com/llvm/llvm-project/issues/55019>`_)
----------------
OikawaKirie wrote:
> One tiny change to the abstraction.
> The ``CStringChecker`` will invalidate less if the copy operation is **inferable to be** bounded.
I decided to elaborate on this a bit. Let me know if it's too thorough now.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D155445/new/

https://reviews.llvm.org/D155445