* refactor: derive `AnymailInboundMessage` from `email.message.EmailMessage`
rather than legacy Python 2.7 `email.message.Message`
* feat(inbound): replace confusing `inline_attachments` with `content_id_map`
and `inlines`; rename `is_inline_attachment` to `is_inline`; deprecate old names
Closes#328
---------
Co-authored-by: Mike Edmunds <medmunds@gmail.com>
- Support Postmark's RawEmail option;
recommend it in docs
- Handle Bcc when provided by Postmark
- Obtain `envelope_sender` from Return-Path info
Postmark now adds, rather than parsing Received-SPF
Related:
- Add `AnymailInboundMessage.bcc` convenience prop
- Test against full Postmark "check" inbound payloads
(which don't match their docs or real inbound payloads)
- Don't warn about receiving "check" payload
Fix AnymailInboundMessage.to, .cc, .from_email when
message was built with AnymailInboundMessage.construct
using empty strings for those params. (Postmark inbound
relies on this.)
Fixes#307
Minimum supported versions are now Django 2.0, Python 3.5.
This touches a lot of code, to:
* Remove obsolete portability code and workarounds
(six, backports of email parsers, test utils, etc.)
* Use Python 3 syntax (class defs, raise ... from, etc.)
* Correct inheritance for mixin classes
* Fix outdated docs content and links
* Suppress Python 3 "unclosed SSLSocket" ResourceWarnings
that are beyond our control (in integration tests due to boto3,
python-sparkpost)
Work around Python 3 email parser change that can turn Unicode
characters into \u escape sequences when parsing a message (or
attachment) that uses "Content-Transfer-Encoding: 8bit".
Useful for cases where ESP could send raw 8bit message
(and its charset is something other than utf-8).
Also reworks earlier Python 2.7 workaround email.parser.Parser header
unfolding bugs to handle any text-like, file-like IO stream, without
trying to manipulate the entire message as a single string.
Make `AnymailInboundMessage.text`, `.html` and `.get_content_text()`
usually do the right thing for non-UTF-8 messages/attachments. Fixes
an incorrect UnicodeDecodeError when receiving an (e.g.,) ISO-8859-1
encoded message, and improves handling for inbound messages that were
not properly encoded by the sender.
* Decode using the message's (or attachments's) declared charset
by default (rather than always defaulting to 'utf-8'; you can
still override with `get_content_text(charset=...)`
* Add `errors` param to `get_content_text()`, defaulting to 'replace'.
Mis-encoded messages will now use the Unicode replacement character
rather than raising errors. (Use `get_content_text(errors='strict')`
for the previous behavior.)
And decide not to work around a Python 3.3 bug accessing MIME headers
that have non-ASCII characters in params. The bug is fixed in the
Python 3.4 email package (and didn't exist in Python 2.7). Python 3.3
was only supported with Django 1.8.
In AnymailInboundMessage, work around Python 2 email.parser.Parser's
lack of handling for RFC2047-encoded email headers. (The Python 3 email
package already decodes these automatically.)
Improves inbound handling on Python 2 for all ESPs that provide raw
MIME email or raw headers with inbound events. (Mailgun, Mandrill,
SendGrid, SparkPost.)
Work around Python 2 email.parser.Parser bug handling RFC5322 folded
headers. Fixes problems where long headers in inbound mail (e.g.,
Subject) get truncated or have unexpected spaces.
This change also updates AnymailInboundMessage.parse_raw_mime to use
the improved "default" email.policy on Python 3 (rather than the
default "compat32" policy). This likely fixes several other parsing
bugs that will still affect code running on Python 2.
Improves inbound parsing for all ESPs that provide raw MIME email.
(Mailgun, Mandrill, SendGrid, SparkPost)
