fix(find_git_conflicts): actually label conflicting PRs and stop skipping past 500 (#15182)

* fix(find_git_conflicts): actually label conflicting PRs and stop skipping past 500

* @priya-sundaram-dev #15181
* [x] Fix a script

* Workaround for mapfile

Co-authored-by: priya-sundaram-dev <oc-409d01@agentmail.to>

* Update scripts/find_git_conflicts.sh

Co-authored-by: priya-sundaram-dev <oc-409d01@agentmail.to>

* Update scripts/find_git_conflicts.sh

Co-authored-by: priya-sundaram-dev <oc-409d01@agentmail.to>

---------

Co-authored-by: priya-sundaram-dev <oc-409d01@agentmail.to>

Author: cclauss

scripts/find_git_conflicts.sh

@@ -1,16 +1,79 @@
 #!/bin/bash
+#
+# Find every open pull request that has git merge conflicts with the base
+# branch and label it "git merge conflict".
+#
+# Why this is not a one-liner: GitHub computes PR mergeability *asynchronously*,
+# so `gh pr list --json mergeable` frequently reports UNKNOWN for PRs it has not
+# recomputed yet. Viewing a PR individually nudges GitHub to compute the value,
+# so we re-query only the UNKNOWN PRs a few times before giving up. We also
+# avoid `--limit 500`, which silently skips any PR past the 500th (this repo has
+# well over 500 open PRs).
+#
+# Usage:
+#   scripts/find_git_conflicts.sh            # label conflicting PRs
+#   DRY_RUN=1 scripts/find_git_conflicts.sh  # list only, add no labels
+#
+# Environment overrides: REPO, LABEL, SLEEP (seconds between UNKNOWN retries).
 
-# Replace with your repository (format: owner/repo)
-REPO="TheAlgorithms/Python"
+set -euo pipefail
+
+REPO="${REPO:-TheAlgorithms/Python}"
+LABEL="${LABEL:-git merge conflict}"
+DRY_RUN="${DRY_RUN:-0}"
+SLEEP="${SLEEP:-2}"
 
-# Fetch open pull requests with conflicts into a variable
 echo "Checking for pull requests with conflicts in $REPO..."
 
-prs=$(gh pr list --repo "$REPO" --state open --json number,title,mergeable --jq '.[] | select(.mergeable == "CONFLICTING") | {number, title}' --limit 500)
+# Make sure the label exists (idempotent; ignore "already exists").
+if [[ "$DRY_RUN" != "1" ]]; then
+    gh label create "$LABEL" --repo "$REPO" \
+        --color "d93f0b" \
+        --description "This pull request has git merge conflicts with the base branch" \
+        2>/dev/null || true
+fi
+
+# First pass: one bulk call. Fast, but mergeable is often UNKNOWN.
+# First pass: one bulk call. Fast, but mergeable is often UNKNOWN.
+rows=()
+while IFS= read -r row; do
+    rows+=("$row")
+done < <(
+    gh pr list --repo "$REPO" --state open --limit 5000 \
+        --json number,mergeable --jq '.[] | "\(.number)\t\(.mergeable)"'
+)
+echo "Found ${#rows[@]} open pull requests to inspect."
+
+conflicting=()
+unknown=()
+for row in "${rows[@]}"; do
+    number="${row%%$'\t'*}"
+    mergeable="${row##*$'\t'}"
+    case "$mergeable" in
+        CONFLICTING) conflicting+=("$number") ;;
+        UNKNOWN) unknown+=("$number") ;;
+    esac
+done
 
-# Process each conflicting PR
-echo "$prs" | jq -c '.[]' | while read -r pr; do
-    PR_NUMBER=$(echo "$pr" | jq -r '.number')
-    PR_TITLE=$(echo "$pr" | jq -r '.title')
-    echo "PR #$PR_NUMBER - $PR_TITLE has conflicts."
+# Second pass: re-query only the UNKNOWN PRs until GitHub finishes computing.
+for pr in ${unknown[@]+"${unknown[@]}"}; do
+    mergeable="UNKNOWN"
+    for _ in 1 2 3; do
+        mergeable=$(gh pr view "$pr" --repo "$REPO" --json mergeable --jq '.mergeable')
+        [[ "$mergeable" != "UNKNOWN" ]] && break
+        sleep "$SLEEP"
+    done
+    [[ "$mergeable" == "CONFLICTING" ]] && conflicting+=("$pr")
 done
+
+# Label the conflicting PRs.
+for pr in ${conflicting[@]+"${conflicting[@]}"}; do
+    echo "PR #$pr has conflicts."
+    if [[ "$DRY_RUN" != "1" ]]; then
+        gh pr edit "$pr" --repo "$REPO" --add-label "$LABEL"
+    fi
+done
+
+# Machine-readable summary (mirrors the close_pull_requests_with_*.sh scripts).
+printf 'CONFLICTING_COUNT=%d CONFLICTING_PRS=%s\n' \
+    "${#conflicting[@]}" "$(IFS=,; echo "${conflicting[*]}")"