From 380520fed6b373071e64fcdee8679becdd381dd9 Mon Sep 17 00:00:00 2001
From: Martin Thomson <mt@lowentropy.net>
Date: Tue, 9 Jul 2024 11:48:51 +1000
Subject: [PATCH] Add a script to detect bad GitHub usernames

---
 get-email.sh | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 upload.mk    |  6 ++---
 2 files changed, 64 insertions(+), 4 deletions(-)
 create mode 100755 get-email.sh

diff --git a/get-email.sh b/get-email.sh
new file mode 100755
index 00000000..a96ebeb6
--- /dev/null
+++ b/get-email.sh
@@ -0,0 +1,62 @@
+#!/usr/bin/env bash
+# Given a tag and a draft, attempt to work out who is responsible for the submission.
+#
+# This tries several things, starting from a simple UPLOAD_EMAIL environment variable.
+# Then it looks at the email of the person that tagged the commit, if that is set.
+# Then it starts looking for who made the commit or who authored the commit.
+# For each of these, GitHub could use a <user>@users.noreply.github.com address.
+# In which case, this attempts to use the GitHub API to get their email.
+# Finally, this picks the first listed author email address from the draft.
+
+tried=()
+emailok() {
+    tried+=("$1")
+    if [ -n "$2" ]; then
+        echo "$2"
+        exit 0
+    fi
+}
+
+# Exit on errors,
+set -e
+# ... but don't print anything; we're using tokens.
+set +x
+
+emailok "\$UPLOAD_EMAIL environment variable" "$UPLOAD_EMAIL"
+
+tag="$1"
+draft="$2"
+
+for k in taggeremail committeremail authoremail; do
+    tagger="$(git tag --list --format '%('"$k"')' "$tag" | sed -e 's/^<//;s/>$//')"
+    ghuser="${tagger%@users.noreply.github.com}"
+    if [ "$ghuser" = "$email" ]; then
+        emailok "git $k" "$tagger"
+    elif [ -n "$GITHUB_API_TOKEN" ]; then
+        script="import json,sys
+e=json.load(sys.stdin).get('email')
+if isinstance(e, str): print(e)"
+        userjson="$(curl -SsLf \
+                    -H "Accept: application/vnd.github+json" \
+                    -H "Authorization: Bearer $GITHUB_API_TOKEN" \
+                    -H "X-GitHub-Api-Version: 2022-11-28" \
+                    "https://api.github.com/users/$ghuser" 2>/dev/null)"
+        apiok=$?
+        emailok "github API for $k ($ghuser)" \
+                "$([ "$apiok" -eq 0 ] && echo "$userjson" | python -c "$script" 2>/dev/null)"
+    fi
+done
+
+if [ -n "$draft" ]; then
+    emailok "draft author" "$(xmllint --xpath '/rfc/front/author[1]/address/email/text()' "$draft" 2>/dev/null)"
+fi
+
+# Give up
+{
+    echo "Unable to find email to use for submission."
+    echo "Tried:"
+    for t in "${tried[@]}"; do
+        echo "    $t"
+    done
+} 1>&2
+exit 1
diff --git a/upload.mk b/upload.mk
index 14621054..cfd79f8b 100644
--- a/upload.mk
+++ b/upload.mk
@@ -25,10 +25,8 @@ endif
 
 .%.upload: %.xml
 	set -ex; tag="$(notdir $(basename $<))"; \
-	email="$(UPLOAD_EMAIL)"; \
-	[ -z "$$email" ] && email="$$(git tag --list --format '%(taggeremail)' "$$tag" | sed -e 's/^<//;s/>$$//')"; \
-	[ -z "$$email" ] && email=$$(xmllint --xpath '/rfc/front/author[1]/address/email/text()' $< 2>/dev/null); \
-	[ -z "$$email" ] && ! echo "Unable to find email to use for submission." 1>&2; \
+	email="$$($(LIBDIR)/get-email.sh "$$tag" "$<")"; \
+	[ -z "$$email" ] && exit 1; \
 	replaces() { \
 	  [ "$${1##*-}" = "00" ] || return; \
 	  file="$$(git ls-files "$${1%-[0-9][0-9]}.*")"; \