Mailgun has two different template mechanisms and two different ways
of providing substitution variables to them. Update Anymail's
normalized merge_data handling to work with either (while preserving
existing batch send and metadata capabilities that also use Mailgun's
custom data and recipient variables parameters).
Completes work started by @anstosa in #156.
Closes#155.
Track Sendinblue API updates:
* Multiple tags are now supported
* When using a template, display name is now supported on 'to', 'bcc', 'cc' and 'replyTo'
* Templates now support overriding 'from_email' and 'subject'
* Templates no longer require separate API endpoint
* 'merge_global_data' can be used without templates
* Rework and simplify personalizations code (that had grown convoluted
through several feature additions).
* Stop putting merge_global_data in legacy template "sections"; instead
just merge it into individual personalization substitutions like we
do for dynamic templates. (The "sections" version didn't add any
functionality, had the potential for conflicts with the user's own
template section tags, and was needlessly complex.)
Support merge_metadata in Mailgun, Mailjet, Mandrill, Postmark,
SparkPost, and Test backends. (SendGrid covered in earlier PR.)
Also:
* Add `merge_metadata` to AnymailMessage, AnymailMessageMixin
* Add `is_batch()` logic to BasePayload, for consistent handling
* Docs
Note: Mailjet implementation switches *all* batch sending from their
"Recipients" field to to the "Messages" array bulk sending option.
This allows an independent payload for each batch recipient.
In addition to supporting merge_metadata, this also removes the
prior limitation on mixing Cc/Bcc with merge_data.
Closes#141.
Postmark docs notwithstanding, Postmark allows sending mail without a
To field, as long as there is some recipient in Cc or Bcc. The API
response has a slightly different shape in this case, and Anymail now
handles that.
Also updates related recipient status parsing. Previously, Anymail's
Postmark backend converted all recipient emails to lowercase for status
reporting, and omitted Cc or Bcc recipients from
`message.anymail_status.recipients[email]`. Now, the backend preserves
the case of each recipient email as originally sent, and includes Cc
and Bcc status.
Because client code may have been relying on lowercasing recipient
emails to check status, this is a potentially breaking change.
Fixes#135
Mailgun's API silently drops attachments without filenames (and inline
attachments without Content-IDs). Raise an AnymailUnsupportedFeature
error on attempts to send these attachments.
Fixes#128
Workaround requests/requests#4652 (urllib3/urllib3#303), where
uploaded files in multipart/form-data are improperly given RFC 2231
encoded filenames. That format is not accepted by Mailgun's API (and is
prohibited by RFC 7578), resulting in the attachments being silently
dropped.
Fix is to patch up the multipart/form-data before posting to remove
the RFC 2231 encoding.
Fixes#125
Optionally dump API requests and responses to stdout, to simplify
debugging of the raw API communications. Currently implemented only
for Requests-based backends.
This (undocumented) setting can log things like API keys, so is not
appropriate for use in production.
Use Postmark /email/batch or /email/batchWithTemplates APIs when
merge_data provided.
Parse Postmark batch-send API responses, and improve accuracy of
parsing individual recipient status from all responses.
Closes#122
Postmark issues an error if Django's default empty strings are used
with template sends.
Include template send in Postmark integration tests. (Requires real
Postmark API token -- templates aren't testable with Postmark's
sandbox token.)
Fixes#121
Python 3.3 moved various collections abstract base classes from
`collections` to `collections.abc`, but also kept them available in
`collections` for compatibility with Python 2. Python 3.8 will allow
importing only from `collections.abc`.
(`collections.abc` hasn't yet been added to six.moves; see
https://github.com/benjaminp/six/issues/155.)
If you are using an SES ConfigurationSet with open or click tracking
enabled, SES replaces non-ASCII characters with question marks as it
rewrites the message to add tracking, if the bodies are sent with
`Content-Transfer-Encoding: 8bit` (which is Django's default for utf8
body parts).
Force potentially problematic parts to use CTE: quoted-printable
as a workaround.
Fixes#115.
SendGrid does not always correctly provide the sent Message-ID header value
to a tracking webhook's smtp-id field, making it unreliable to use for Anymail's
`message_id`.
Instead, generate a UUID `message_id` for Anymail tracking, and pass it from
send to webhooks in SendGrid custom args as anymail_id.
Webhooks will fall back to smtp-id for compatibility with previously-sent
messages that didn't have an anymail_id custom arg.
Fixes#108
Include the name of the field with the the unparsable email address
in AnymailInvalidAddress error messages.
Should help tracking down problems like in #98.
Django's SMTP EmailBackend allows spoofing the To header by setting
`message.extra_headers["To"]`` different from `message.to`.
No current Anymail ESP supports this. Treat extra_headers["To"] as
an unsupported ESP feature, to flag attempts to use it.
Also document Anymail's special header handling that replicates
Django's SMTP EmailBackend behavior.
New EmailMessage attribute `envelope_sender` controls ESP's sender,
sending domain, or return path where supported:
* Mailgun: overrides SENDER_DOMAIN on individual message
(domain portion only)
* Mailjet: becomes `Sender` API param
* Mandrill: becomes `return_path_domain` API param
(domain portion only)
* SparkPost: becomes `return_path` API param
* Other ESPs: not believed to be supported
Also support undocumented Django SMTP backend behavior, where envelope
sender is given by `message.from_email` when
`message.extra_headers["From"]` is set. Fixes#91.
Django allows setting the reply address with either message.reply_to
or message.extra_headers["Reply-To"]. If both are supplied, the extra
headers version takes precedence. (See EmailMessage.message().)
Several Anymail backends had duplicate logic to handle conflicting
properties. Move that logic into the base Payload.
(Also prepares for common handling of extra_headers['From'], later.)
Related changes:
* Use CaseInsensitiveDict for processing extra_headers.
This is potentially a breaking change, but any code that was trying
to send multiple headers differing only in case was likely already
broken. (Email header field names are case-insensitive, per RFC-822.)
* Handle CaseInsensitiveDict in RequestsPayload.serialize_json().
(Several backends had duplicate code for handling this, too.)
* Fixes SparkPost backend, which had been incorrectly treating
message.reply_to and message.extra_headers['Reply-To'] differently.
Add support for sending transactional email through SendinBlue. (Thanks to @RignonNoel.)
Partially implements #84. (Tracking webhooks will be a separate PR. SendinBlue doesn't support inbound handling.)
* Un-hardcode status message_id in test backend
For the test EmailBackend, get message ID's based on array position in
`mail.outbox`, so that tests can predict the message ID.
* Add a console backend for use in development
Adds an EmailBackend derived from both Anymail's test backend and
Django's console backend, to provide anymail statuses and signal
handling while printing messages to the console. For use during
development on localhost.
Closes#87
Use a default timeout of 30 seconds for all requests, and add a
REQUESTS_TIMEOUT Anymail setting to override.
(I'm making a judgement call that this is not a breaking change in the
real world, and not bumping the major version. Theoretically, it could
affect you if your network somehow takes >30s to connect to your ESP,
but eventually succeeds. If so, set REQUESTS_TIMEOUT to None to restore
the earlier behavior.)
Fixes#80.
Within an EmailAddress (previously ParsedEmail object), properties
now match Python 3.6 email.headerregistry.Address naming:
* .email --> .addr_spec
* .name --> .display_name
* .localpart --> .username
(Completes work started in 386668908423d1d4eade90cf7a21a546a1e96514;
this updates remaining uses of old names and removes them.)
Update internal-use ParsedEmail to be more like Python 3.6+
email.headerregistry.Address, and remove "internal use only"
recommendation.
(Prep for exposing inbound email headers in a convenient form.
Old names remain temporarily available for internal use;
should clean up at some point.)
SendGrid requires extra headers and metadata values be strings.
Anymail has always coerced int and float; this treats Python 2's
`long` integer type the same.
Fixes#74
* Change Anymail's test EmailBackend to collect sent messages in
django.core.mail.outbox, same as Django's own locmem EmailBackend.
(So Django's test runner will automatically clear accumulated mail
between test cases.)
* Rename EmailMessage `test_response` attr to `anymail_test_response`
to avoid conflicts, and record merged ESP send params in
new `anymail_send_params` attr.
* Add docs
Closes#36.
[RFC-5322 allows](https://tools.ietf.org/html/rfc5322#section-3.6.2)
multiple addresses in the From header.
Django's SMTP backend supports this, as a single comma-separated
string (*not* a list of strings like the recipient params):
from_email='one@example.com, two@example.com'
to=['one@example.com', 'two@example.com']
Both Mailgun and SparkPost support multiple From addresses
(and Postmark accepts them, though truncates to the first one
on their end). For compatibility with Django -- and because
Anymail attempts to support all ESP features -- Anymail now
allows multiple From addresses, too, for ESPs that support it.
Note: as a practical matter, deliverability with multiple
From addresses is pretty bad. (Google outright rejects them.)
This change also reworks Anymail's internal ParsedEmail object,
and approach to parsing addresses, for better consistency with
Django's SMTP backend and improved error messaging.
In particular, Django (and now Anymail) allows multiple email
addresses in a single recipient string:
to=['one@example.com', 'two@example.com, three@example.com']
len(to) == 2 # but there will be three recipients
Fixes#60
Issue a better error message if message.reply_to
is set to a single string.
(Would also like to do this for to, cc, and bcc,
but Django core EmailMessage.recipients is called
and stumbles over thoses cases before Anymail's
backend gets involved.)
Fixes#57
* **Future breaking change:**
Rename all Anymail backends to just `EmailBackend`,
matching Django's naming convention.
(E.g., switch to "anymail.backends.mailgun.EmailBackend"
rather than "anymail.backends.mailgun.MailgunBackend".)
The old names still work, but will issue a DeprecationWarning
and will be removed in some future release.
(Apologies for this change; the old naming convention was
a holdover from Djrill, and I wanted consistency with
other Django EmailBackends before hitting 1.0.)
Fixes#49.
SendGrid: update to v3 send API
**SendGrid:** **[possibly-breaking]** Update SendGrid backend to newer Web API v3. This should be a transparent change for most projects. Exceptions: if you use SendGrid username/password auth, esp_extra with "x-smtpapi", or multiple Reply-To addresses, please review the [porting notes](http://anymail.readthedocs.io/en/latest/esps/sendgrid/#sendgrid-v3-upgrade).
Closes#28
In BasePayload, ensure any Django ugettext_lazy
(or similar) are converted to real strings before
handing off to ESP code. This resolves problems where
calling code expects it can use lazy strings "anywhere",
but non-Django code (requests, ESP packages) don't
always handle them correctly.
* Add utils helpers for lazy objects (is_lazy, force_non_lazy*)
* Add lazy object handling to utils.Attachment
* Add lazy object handling converters to BasePayload attr
processing where appropriate. (This ends up varying by
the expected attribute type.)
Fixes#34.
