This is an automated email from the ASF dual-hosted git repository.
hello-stephen pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/master by this push:
new 38a38dda22b [fix](ci) Use GPT-6 Sol for review capacity fallback
(#69988)
38a38dda22b is described below
commit 38a38dda22b44692cb3ed317b1f510a9b9b01493
Author: shuke <[email protected]>
AuthorDate: Mon Nov 2 17:54:33 2026 +0800
[fix](ci) Use GPT-6 Sol for review capacity fallback (#69988)
### What problem does this PR solve?
Related PR: #69953
When GPT-6.1 Sol encounters five capacity failures, review recovery
currently switches to GPT-6 Astra, which has continued to encounter
capacity failures in observed runs. Change the capacity fallback to
`gpt-6-sol` and update the existing test fixtures and assertions to
exercise that target.
The production change is one workflow argument. The primary model
remains `gpt-6.1-sol` with `xhigh`; the five-failure threshold, seven
total capacity attempts, 300-second backoff, same-session recovery,
shared budget, ledger and round preservation, head-change checks, and
submission verification are unchanged.
---
.github/scripts/test_run_review_with_resume.py | 14 +++++++-------
.github/workflows/code-review-runner.yml | 2 +-
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/.github/scripts/test_run_review_with_resume.py
b/.github/scripts/test_run_review_with_resume.py
index 97018bff25c..5d71f6adc19 100644
--- a/.github/scripts/test_run_review_with_resume.py
+++ b/.github/scripts/test_run_review_with_resume.py
@@ -86,7 +86,7 @@ class ResumeReviewTest(unittest.TestCase):
base_sha="b" * 40,
model="gpt-5.6-sol",
fallback_model=None,
- capacity_fallback_model="gpt-6-astra",
+ capacity_fallback_model="gpt-6-sol",
effort="xhigh",
budget_seconds=1000,
)
@@ -435,7 +435,7 @@ class ResumeReviewTest(unittest.TestCase):
+ [{"events": [thread_event(), completed()], "status": 0}]
))
self.assertEqual(
- ["gpt-6.1-sol"] * 5 + ["gpt-6-astra"],
+ ["gpt-6.1-sol"] * 5 + ["gpt-6-sol"],
[c[c.index("--model") + 1] for c in self.commands],
)
self.assertEqual([3600, 3295, 2990, 2685, 2380, 2075], self.timeouts)
@@ -444,11 +444,11 @@ class ResumeReviewTest(unittest.TestCase):
self.assertEqual(["resume", THREAD], command[-3:-1])
self.assertIn("model_reasoning_effort=xhigh", command)
self.assertNotIn("Capacity recovery switched", self.commands[4][-1])
- self.assertIn("Capacity recovery switched to gpt-6-astra",
self.commands[5][-1])
+ self.assertIn("Capacity recovery switched to gpt-6-sol",
self.commands[5][-1])
self.assertIn("including child agents", self.commands[5][-1])
self.assertIn("cumulative round count", self.commands[5][-1])
self.assertEqual(before, self.ledger.read_text())
- self.assertEqual("gpt-6-astra\n", (self.context /
"codex-review-model.txt").read_text())
+ self.assertEqual("gpt-6-sol\n", (self.context /
"codex-review-model.txt").read_text())
self.verification.assert_called_once()
def test_capacity_fallback_is_optional(self):
@@ -502,7 +502,7 @@ class ResumeReviewTest(unittest.TestCase):
def
test_unsupported_capacity_fallback_does_not_restart_or_switch_back(self):
self.args.budget_seconds = 3600
- rejection = "The 'gpt-6-astra' " + runner.UNSUPPORTED_CHATGPT_MODEL
+ rejection = "The 'gpt-6-sol' " + runner.UNSUPPORTED_CHATGPT_MODEL
self.assertEqual(1, self.execute(
[{"events": [thread_event(), failed()]}] * 5
+ [{"events": [thread_event(), failed(rejection)]}]
@@ -666,7 +666,7 @@ class ResumeReviewTest(unittest.TestCase):
self.assertEqual([300] * 6, self.sleeps)
self.assertEqual(7, len(self.commands))
self.assertEqual(
- [self.args.model] * 5 + ["gpt-6-astra"] * 2,
+ [self.args.model] * 5 + ["gpt-6-sol"] * 2,
[c[c.index("--model") + 1] for c in self.commands],
)
self.assertEqual(runner.CAPACITY_MESSAGE, self.last_error())
@@ -1146,7 +1146,7 @@ HELPER
check=True,
)
args = json.loads(result.stdout)
- self.assertEqual("gpt-6-astra",
args[args.index("--capacity-fallback-model") + 1])
+ self.assertEqual("gpt-6-sol",
args[args.index("--capacity-fallback-model") + 1])
actual = int(args[args.index("--budget-seconds") + 1])
self.assertLessEqual(actual, expected)
self.assertGreaterEqual(actual, expected - 2)
diff --git a/.github/workflows/code-review-runner.yml
b/.github/workflows/code-review-runner.yml
index 802567c1beb..719f460466c 100644
--- a/.github/workflows/code-review-runner.yml
+++ b/.github/workflows/code-review-runner.yml
@@ -919,7 +919,7 @@ jobs:
--base-sha "$BASE_SHA" \
--model "gpt-6.1-sol" \
--fallback-model "gpt-6-sol" \
- --capacity-fallback-model "gpt-6-astra" \
+ --capacity-fallback-model "gpt-6-sol" \
--effort xhigh \
--budget-seconds "$budget_seconds"
status=$?
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
