From 5c189bc13dc702ba69fb23f6dcaf85a210114ca0 Mon Sep 17 00:00:00 2001
From: Nithin <sarvanithin@gmail.com>
Date: Sun, 30 Aug 2026 05:15:46 -0400
Subject: [PATCH] fix(workday): accept canonical public job links

---
 src/fortune_intel/connectors/workday.py | 23 +++++++++++++++--
 tests/test_workday_connector.py         | 34 +++++++++++++++++++++++++
 2 files changed, 55 insertions(+), 2 deletions(-)

diff --git a/src/fortune_intel/connectors/workday.py b/src/fortune_intel/connectors/workday.py
index c0370c9..685310b 100644
--- a/src/fortune_intel/connectors/workday.py
+++ b/src/fortune_intel/connectors/workday.py
@@ -389,13 +389,32 @@ def _public_job_url(self, value: object) -> str:
             segment in {".", ".."} or "/" in segment or "\\" in segment
             for segment in decoded_segments
         )
+        external_host = (parsed.hostname or "").casefold().rstrip(".")
+        # Workday boards can serve their anonymous CXS API from a tenant
+        # ``*.myworkdayjobs.com`` host while returning the applicant-facing URL
+        # from Workday's canonical ``wdN.myworkdaysite.com`` host. It is still
+        # the same board only when the URL repeats the exact tenant and site.
+        canonical_public_alias = _MYWORKDAYSITE_HOST_PATTERN.fullmatch(
+            external_host
+        ) is not None and parsed.path.casefold().startswith(
+            (
+                f"/recruiting/{quote(self.workday.tenant, safe='')}/"
+                f"{quote(self.workday.site, safe='')}/job/"
+            ).casefold()
+        )
+        allowed_host = external_host == self.workday.host or canonical_public_alias
         if (
             parsed.scheme != "https"
-            or (parsed.hostname or "").casefold().rstrip(".") != self.workday.host
+            or not allowed_host
             or parsed.username is not None
             or parsed.password is not None
             or parsed.port not in {None, 443}
-            or not parsed.path.startswith(self.workday.public_job_path_prefix)
+            or (
+                not canonical_public_alias
+                and not parsed.path.casefold().startswith(
+                    self.workday.public_job_path_prefix.casefold()
+                )
+            )
             or unsafe_path
             or (
                 self.workday.uses_recruiting_path
diff --git a/tests/test_workday_connector.py b/tests/test_workday_connector.py
index fd0853a..09762e0 100644
--- a/tests/test_workday_connector.py
+++ b/tests/test_workday_connector.py
@@ -110,6 +110,40 @@ def test_recruiting_path_source_fetches_manifest_and_detail_from_same_fixed_host
     )
 
 
+def test_accepts_workday_canonical_public_alias_for_same_tenant_and_site():
+    """A tenant API host can return the same board's canonical apply URL."""
+
+    detail = workday_detail("JR-1")
+    detail["jobPostingInfo"]["externalUrl"] = (
+        "https://wd5.myworkdaysite.com/recruiting/acme/External/job/US-NY/Role-JR-1_JR-1"
+    )
+    client = StubClient([{"total": 1, "jobPostings": [workday_summary("JR-1")]}, detail])
+    key = workday_source("acme.wd5.myworkdayjobs.com", "acme", "External").key
+
+    result = WorkdayConnector(key, client=client).fetch()
+
+    assert result.complete is True
+    assert result.jobs[0].url.startswith(
+        "https://wd5.myworkdaysite.com/recruiting/acme/External/job/"
+    )
+
+
+def test_accepts_case_variant_of_same_workday_site_path():
+    """Workday can capitalize a site path differently from its board URL."""
+
+    detail = workday_detail("JR-1")
+    detail["jobPostingInfo"]["externalUrl"] = (
+        "https://acme.wd5.myworkdayjobs.com/EXTERNAL/job/US-NY/Role-JR-1_JR-1"
+    )
+    client = StubClient([{"total": 1, "jobPostings": [workday_summary("JR-1")]}, detail])
+    key = workday_source("acme.wd5.myworkdayjobs.com", "acme", "External").key
+
+    result = WorkdayConnector(key, client=client).fetch()
+
+    assert result.complete is True
+    assert result.jobs[0].url.endswith("/EXTERNAL/job/US-NY/Role-JR-1_JR-1")
+
+
 def test_fixture_preserves_exact_posting_date_and_native_id():
     client = StubClient(
         [