From 80352f727f1226b4c7116d136ad5091678968242 Mon Sep 17 00:00:00 2001
From: "Peter J. Holzer" <hjp@hjp.at>
Date: Sat, 19 Jan 2019 21:18:14 +0100
Subject: [PATCH] Extract message ids from mbox files

Parses all mbox files on the command line and extracts message ids from
Message-Id, In-Reply-To and References headers.

Just a test script to see what message ids look like in practice.
---
 get_message_ids | 29 +++++++++++++++++++++++++++++
 1 file changed, 29 insertions(+)
 create mode 100755 get_message_ids

diff --git a/get_message_ids b/get_message_ids
new file mode 100755
index 0000000..eb5b235
--- /dev/null
+++ b/get_message_ids
@@ -0,0 +1,29 @@
+#!/usr/bin/python3
+
+import mailbox
+import re
+import sys
+
+for f in sys.argv[1:]:
+    print("F", f)
+    mb = mailbox.mbox(f)
+
+    for m in mb:
+        try:
+            for match in re.findall(r'<(.*?)>', m["Message-ID"]):
+                print('M', match)
+            if "In-Reply-To" in m:
+                h = str(m["In-Reply-To"]) # sometimes it's a string,
+                # sometimes an email.header.Header. But the latter's
+                # __str__ method returns something sensible, so let's just
+                # force it to be a string
+                for match in re.findall(r'<(.*?)>', h):
+                    print('I', match)
+            if "References" in m:
+                for match in re.findall(r'<(.*?)>', m["References"]):
+                    print('R', match)
+        except:
+            print("Error in message:", file=sys.stderr)
+            print(m.as_string(), file=sys.stderr)
+            sys.exit(1)
+