Skip to content

Commit cb14bc0

Browse files
Update changelogs from fern repo (#5217)
Co-authored-by: dsinghvi <10870189+dsinghvi@users.noreply.github.com>
1 parent c31fc4d commit cb14bc0

2 files changed

Lines changed: 17 additions & 0 deletions

File tree

fern/products/cli-api-reference/cli-changelog/2026-04-28.mdx

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,13 @@
1+
## 4.107.0
2+
**`(feat):`** Add `infer-discriminated-union-base-properties` OpenAPI parser setting.
3+
When enabled, the parser intersects the resolved (allOf-flattened) property
4+
maps of every variant of a discriminated `oneOf` and lifts properties that
5+
appear in all variants with structurally-equal schemas into the union's
6+
common properties. SDKs (e.g. C#, Java, Go) that already support union base
7+
properties can then expose those shared fields directly on the union type
8+
without forcing callers to cast to a concrete variant. Defaults to false.
9+
10+
111
## 4.106.3
212
**`(fix):`** Send `gitattributesEntries` to Fiddle on `fern replay init` so the
313
generated PR marks `.fern/replay.lock` as `linguist-generated=true`.

fern/products/sdks/generators/python/changelog/2026-04-28.mdx

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,10 @@
1+
## 5.8.3
2+
**`(fix):`** Fix async SSE parsing to only split on SSE-spec line terminators (\n, \r, \r\n).
3+
Previously, aiter_sse() used httpx's aiter_lines() which splits on \u2028 and
4+
\u2029 via Python's str.splitlines(), corrupting data fields containing those
5+
characters (common in LLM JSON output).
6+
7+
18
## 5.8.2
29
**`(chore):`** Bump @fern-api/generator-cli to 0.9.16, which bundles the autoversion
310
pipeline's private workspace deps (@fern-api/logging-execa,

0 commit comments

Comments
 (0)