fix: skip out of order PR updates (#3324)

* fix: skip stale updates

* feat: reject out of order updates in main util as well

* fix: do not need this here

Author: beckermr

.github/workflows/bot-bot.yml

@@ -52,7 +52,7 @@ jobs:
         if: success() && ! env.CI_SKIP
         timeout-minutes: 210
         run: |
-          export TIMEOUT=7200
+          export TIMEOUT=2100
           export RUN_URL="https://github.com/regro/cf-scripts/actions/runs/${RUN_ID}"
 
           pushd cf-graph

conda_forge_tick/git_utils.py

@@ -10,7 +10,7 @@
 import threading
 import time
 from abc import ABC, abstractmethod
-from datetime import datetime
+from datetime import datetime, timezone
 from email import utils
 from functools import cached_property
 from pathlib import Path
@@ -1218,6 +1218,16 @@ def _munge_dict(dest, src, keys):
     return pr_json
 
 
+def parse_pr_json_last_updated(pr_data: Union[Dict, LazyJson]) -> Optional[datetime]:
+    """Parse the last updated time from a PR json blob. If it is not present, return None."""
+    last_updated = pr_data.get("updated_at", None)
+    if last_updated is not None:
+        last_updated = datetime.fromisoformat(last_updated)
+        if last_updated.tzinfo is None:
+            last_updated = last_updated.replace(tzinfo=timezone.utc)
+    return last_updated
+
+
 def lazy_update_pr_json(
     pr_json: Union[Dict, LazyJson], force: bool = False
 ) -> Union[Dict, LazyJson]:
@@ -1240,6 +1250,8 @@ def lazy_update_pr_json(
     pr_json : dict-like
         A dict-like object with the current PR information.
     """
+    last_updated = parse_pr_json_last_updated(pr_json)
+
     hdrs = {
         "Authorization": f"token {get_bot_token()}",
         "Accept": "application/vnd.github.v3+json",
@@ -1270,9 +1282,19 @@ def lazy_update_pr_json(
     )
 
     if r.status_code == 200:
-        pr_json = trim_pr_json_keys(pr_json, src_pr_json=r.json())
-        pr_json["ETag"] = r.headers["ETag"]
-        pr_json["Last-Modified"] = r.headers["Last-Modified"]
+        # I have seen things come in out of order for reasons I do not
+        # fully understand. We do not update in this case. - MRB
+        new_last_updated = parse_pr_json_last_updated(r.json())
+        if (
+            last_updated is None
+            or new_last_updated is None
+            or new_last_updated >= last_updated
+        ):
+            pr_json = trim_pr_json_keys(pr_json, src_pr_json=r.json())
+            pr_json["ETag"] = r.headers["ETag"]
+            pr_json["Last-Modified"] = r.headers["Last-Modified"]
+        else:
+            pr_json = trim_pr_json_keys(pr_json)
     else:
         pr_json = trim_pr_json_keys(pr_json)
 
@@ -1358,8 +1380,11 @@ def close_out_labels(
             pr_obj = get_pr_obj_from_pr_json(pr_json, gh)
             pr_obj.create_comment(
                 "Due to the `bot-rerun` label I'm closing "
-                "this PR. I will make another one as"
-                f" appropriate.\n\n<sub>This message was generated by {get_bot_run_url()} "
+                "this PR. I will make another PR as"
+                " appropriate. You should expect to "
+                "wait at least a few hours, or possibly "
+                "much longer, for a new PR."
+                f"\n\n<sub>This message was generated by {get_bot_run_url()} "
                 "- please use this URL for debugging.</sub>",
             )
             pr_obj.close()
@@ -1404,7 +1429,10 @@ def close_out_dirty_prs(
                 pr_obj.create_comment(
                     "I see that this PR has conflicts, and I'm the only committer. "
                     "I'm going to close this PR and will make another one as"
-                    f" appropriate.\n\n<sub>This was generated by {get_bot_run_url()} - "
+                    " appropriate. You should expect to "
+                    "wait at least a few hours, or possibly "
+                    "much longer, for a new PR."
+                    f"\n\n<sub>This was generated by {get_bot_run_url()} - "
                     "please use this URL for debugging.</sub>",
                 )
                 pr_obj.close()

conda_forge_tick/lazy_json_backends.py

@@ -1,3 +1,4 @@
+import base64
 import contextlib
 import functools
 import glob
@@ -942,7 +943,11 @@ def main_cache(ctx: CliContext):
 def _get_pth_blob_sha_and_content(pth, gh):
     try:
         cnt = gh.get_repo("regro/cf-graph-countyfair").get_contents(pth)
-        return cnt.sha, cnt.decoded_content.decode("utf-8")
+        # I was using the decoded_content attribute here, but it seems that
+        # every once and a while github does not send the encoding correctly
+        # so I switched to doing the decoding by hand.
+        data = base64.b64decode(cnt.content.encode("utf-8")).decode("utf-8")
+        return cnt.sha, data
     except github.UnknownObjectException:
         return None, None
 

conda_forge_tick/make_migrators.py

@@ -95,8 +95,8 @@
 
 logger = logging.getLogger(__name__)
 
-PR_LIMIT = 5
-MAX_PR_LIMIT = 50
+PR_LIMIT = 2
+MAX_PR_LIMIT = 20
 MAX_SOLVER_ATTEMPTS = 50
 CHECK_SOLVABLE_TIMEOUT = 90  # 90 days
 DEFAULT_MINI_MIGRATORS = [
@@ -370,10 +370,10 @@ def add_rebuild_migration_yaml(
     pr_limits = [
         min(2, nominal_pr_limit),
         nominal_pr_limit,
-        min(int(nominal_pr_limit * 4), MAX_PR_LIMIT),
-        min(int(nominal_pr_limit * 3), MAX_PR_LIMIT),
         min(int(nominal_pr_limit * 2), MAX_PR_LIMIT),
-        min(int(nominal_pr_limit * 1.5), MAX_PR_LIMIT),
+        min(int(nominal_pr_limit * 1.75), MAX_PR_LIMIT),
+        min(int(nominal_pr_limit * 1.50), MAX_PR_LIMIT),
+        min(int(nominal_pr_limit * 1.25), MAX_PR_LIMIT),
         min(nominal_pr_limit, MAX_PR_LIMIT),
     ]
 
@@ -709,6 +709,7 @@ def create_migration_yaml_creator(
                                 cfp_gx,
                                 pinnings=pinnings_together,
                                 full_graph=gx,
+                                pr_limit=1,
                             ),
                         )
             except Exception as e:
@@ -809,7 +810,7 @@ def initialize_migrators(
         version_migrator = Version(
             python_nodes=python_nodes,
             graph=gx,
-            pr_limit=PR_LIMIT * 4,
+            pr_limit=PR_LIMIT * 2,
             piggy_back_migrations=_make_mini_migrators_with_defaults(
                 extra_mini_migrators=[
                     PipWheelMigrator(),

conda_forge_tick/migrators/migration_yaml.py

@@ -127,7 +127,7 @@ def __init__(
         yaml_contents: str,
         name: str,
         graph: nx.DiGraph = None,
-        pr_limit: int = 50,
+        pr_limit: int = 0,
         top_level: Set["PackageName"] = None,
         cycles: Optional[Sequence["PackageName"]] = None,
         migration_number: Optional[int] = None,
@@ -407,7 +407,7 @@ def __init__(
         graph: nx.DiGraph,
         pin_impact: Optional[int] = None,
         full_graph: Optional[nx.DiGraph] = None,
-        pr_limit: int = 1,
+        pr_limit: int = 0,
         bump_number: int = 1,
         effective_graph: nx.DiGraph = None,
         pinnings: Optional[List[int]] = None,

conda_forge_tick/migrators/noarch_python_min.py

@@ -420,7 +420,7 @@ class NoarchPythonMinMigrator(Migrator):
     def __init__(
         self,
         *,
-        pr_limit: int = 10,
+        pr_limit: int = 0,
         graph: nx.DiGraph = None,
         effective_graph: nx.DiGraph = None,
         piggy_back_migrations: Sequence[MiniMigrator] | None = None,