[PATCH] D85838: New TableGen Programmer's Reference document

[Paul C. Anagnostopoulos via Phabricator via llvm-commits]

Paul-C-Anagnostopoulos added inline comments.


================
Comment at: llvm/docs/TableGen/ProgRef.rst:197
+.. warning::
+  The ``field`` reserved word is deprecated.
+
----------------
kosarev wrote:
> How do we know that? There are some mentions of deprecated features in the code, but `field` doesn't seem to be amongst them. Also, the instruction decoder backend looks at the `field` mark to distinct encoding field definitions from ordinary TableGen class fields which is something we seem to rely on in our instruction definition .td's. If `field` is deprecated, then what's supposed to be a replacement for this use case?
We were hoping that no one would need to use it in the future, which is why it is flagged as deprecated. At some point I am supposed to try to figure out what to do about the current uses.


================
Comment at: llvm/docs/TableGen/ProgRef.rst:1153
+
+The ``foreach`` statement can also be used in a record :token:`Body`.
+
----------------
foad wrote:
> Is this true? This assertion did not appear in the previous documentation, and it doesn't seem to be borne out by the grammar productions: `BodyItem` only includes `let`, `defvar` and `assert` statements.
> 
> @Joe_Nash for awareness.
It is not true. Feel free to get rid of that sentence. Good catch.


================
Comment at: llvm/docs/TableGen/ProgRef.rst:602
+.. productionlist::
+   Def: "def" [`NameValue`] `RecordBody`
+   NameValue: `Value`
----------------
nhaehnle wrote:
> Missing a colon?
The colon is part of the RecordBody production.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D85838/new/

https://reviews.llvm.org/D85838