[libcxx-commits] [libcxx] [RFC][libc++] Testing feature test macro script. (PR #90889)

[Louis Dionne via libcxx-commits]

================
@@ -1857,6 +1858,38 @@ def produce_docs():
         f.write(doc_str)
 
 
+def get_table(data):
----------------
ldionne wrote:

Can you document the format of the data taken by this function? Also, the script assumes a lot of global state right now (like `get_std_dialects()`), it would be nice to improve on this as we're basically rewriting this. Can't we actually infer the standard dialects from the json data we're reading?

We could potentially have a Python class that represents this JSON data and provides a few useful attributes like `.dialects` and more, and then we could query the things we implement with something like

```python
for std in table.dialects:
  for ftm in table.entries:
    if ftm.implemented:
      # print information for this FTM in this dialect
```

https://github.com/llvm/llvm-project/pull/90889