Right now notmuch panics if something goes wrong in the configure event.
This patch checks for that and returns an error instead, so that we can at least
get the UI up and running (and all the other accounts)
The experience will be completely degraded until another configure event occurs.
Configure an oauthbearer source without a token_endpoint
parameter would panic due to nil pointer dereference
Example
source=imaps+oauthbearer://frode.aa%40gmail.com@imap.gmail.com:993
source-cred-cmd=pass oatuh2 frode.aa@gmail.com
token_endpoint is not required as it will use the provided
password as access_token when it is not set
This changes the search flags for maildir and imap backends.
They now no longer use -t for searching all text. This seems to make
more sense as being the targeted recipient. I have similarly added Cc
for -c. The text search now resides under -a for all text.
There is a command and config option. The criteria are a list of the
sort criterion and each can be individually reversed.
This only includes support for sorting in the maildir backend currently.
The other backends are not supported in this patch.
This populates the directory info model properly when requested,
allowing the fields to be relied upon elsewhere.
This also sends the dirinfo when new messages come in.
Basic searching is supported with the following:
- read messages
- unread messages
- from addresses
- text in body
- text in subject
- text in all
The implementation loops through all messages in the selected directory.
It tries to be smart by detecting which parts of each message the search
query needs to use and only loads these from the filesystem.
For some reason the current code frequently segfaults due to an
invalid C memory address. This commit mediates that by never keeping an object
alive longer than absolutely necessary.
Notmuch only allows a single write connection, all other clients trying to
modify the db block. Hence we should only open one when we actually need it.
Apparently we also need to refresh the RO DB connection upon modification,
else we get stale message tag results
This commit introduces the notmuch backend.
The backend is conditionally compiled in if the "notmuch" tag is provided.
Most of the message types are implemented, with the notable exceptions
of DeleteMessages as well as any copy / move / append type.
Reason being, that those aren't normally applicable in a notmuch based workflow.
Changes v2 --> v3, based on review comments
* Use account config for configuration
Things like FetchEntityPartReader etc can be reused by most workers
working with raw email files from disk (or any reader for that matter).
This patch extract that common functionality in a separate package.
This fixes ~sircmpwn/aerc2#245. This sets up the imap client to send
error messages to the logger of the worker. Errors now end up in the
bottom status line.
https://todo.sr.ht/~sircmpwn/aerc2/245
When a directory is opened, start watching its "new" subdirectory for
incoming messages using the fsnotify library. When creation events are
detected, run the Unseen routine to move the message from new to cur and
add new UIDs to the store, updating the UI's list of directory contents
as we go.
Email headers can be encoded with different charsets, which is signalled
using a special character sequence. The go-message package provides two
different methods for accessing header values, Get(key) (actually
inherited from the embedded textproto.Header) which returns the raw
header value and Text(key), which returns the header's value decoded as
UTF-8.
Before, in the maildir backend, we were using the Get method which
sometimes resulted in encoded headers being displayed in the UI. This
patch replaces the incorrect usage of Get() with Text().
Add the initial implementation of a backend for Maildir accounts. Much
of the functionality required is implemented in the go-message and
go-maildir libraries, so we use them as much as possible.
The maildir worker hooks into a new maildir:// URL scheme in the
accounts.conf file which points to a container of several maildir
directories. From there, the OpenDirectory, FetchDirectoryContents, etc
messages work on subdirectories. This is implemented as a Container
struct which handles mapping between the symbolic email folder names and
UIDs to the concrete directories and file names.
imaps+oauthbearer://user:token@host?token_endpoint=...
- the config Source password is used as access token if
no token_endpoint parameter is set
- the config Source password is used as refresh token if
token_endpoint parameter is set, and used to exchange
with an access token
The implementation has only been tested with Gmail.
source = imaps+oauthbearer://{username}:{refersh_token}@imap.gmail.com:993? \
client_id=XX&\
client_secret=XX&\
token_endpoint=https%3A%2F%2Faccounts.google.com%2Fo%2Foauth2%2Ftoken
client credentials created with
https://console.developers.google.com/apis/credentials
refresh token created with
https://github.com/google/gmail-oauth2-tools/blob/master/python/oauth2.py
rel: https://todo.sr.ht/~sircmpwn/aerc2/42
Before, we were using several IMAP-specific concepts to represent
information being displayed in the UI. Factor these structures out of
the IMAP package to make it easier for other backends to provide the
required information.
A sequence-set is an IMAP-specific implementation detail. Throughout the
UI, aerc simply operates using lists of opaque identifiers. In order to
loosen the coupling between the UI and IMAP in particular, replace most
usages of imap.SeqSet with []uint32, leaving the translation to a SeqSet
to the IMAP backend as needed.
Before, the information needed to display different parts of the UI was
tightly coupled to the specific messages being sent back and forth to
the backend worker. Separating out a models package allows us to be more
specific about exactly what a backend is able to and required to
provide for the UI.
Adds an archive command that moves the current message into the folder
specified in the account config entry.
Supports three layouts at this point:
- flat: puts all messages next to each other
- year: creates a folder per year
- month: same as above, plus folders per month
This also adds a "-p" argument to "cp" and "mv" that works like
"--parents" on mkdir(1). We use this to auto-create the directories
for the archive layout.
If a MailboxInfo has the attribute \Noselect,
it is not possible to use this name as a selectable mailbox.
Therefore it should not be passed to the directory handlers.
The issue pops up if one has a hierarchy like this:
INBOX
INBOX/lists/stuff
INBOX/lists/otherStuff
Even though lists is not a valid inbox (doesn't contain mail, only other maildirs)
it will show up in the directory listing, when we iterate over the MailboxInfo.
It does have the corresponding attribute set though and we can simply filter it out.
Worker.callbacks contains two types of callbacks: some are action callbacks,
some are message callbacks. Each of those is access from one side of the
communication channel (UI goroutine vs. worker goroutine).
Instead of using a channel, we can use two different maps for each kind. This
simplifies the code and also ensures we don't call an action callback instead
of a message callback (or the other way around).
Message IDs are allocated for both messages from UI to workers and the other
way around. Hence, the global nextId variable is accessed from multiple
goroutines.
Instead, use atomic to access the global counter.
Unfortunately, the IMAP protocol hasn't been designed to be used from multiple
goroutines at the same time. For instance, if you fetch twice the same message
from two different goroutines, it's not possible to tell whether the response
is for one receiver or the other. For this reason, go-imap clients aren't safe
to use from multiple goroutines.
This commit changes the IMAP workers to be synchronous again (a command is
executed only after the previous one has completed). To use IMAP from different
threads, popular clients (e.g. Thunderbird) typically open multiple
connections.
Worker.Process* functions were called in different goroutines than
Worker.Post*. Protect the map with a mutex. Also make the map unexported to
prevent external unprotected accesses.
Worker.Process* functions used to delete items from the map. However they
didn't delete the element they retrieved: callbacks[msg.InResponseTo()] was
read while callbacks[msg] was deleted. I'm not sure I understand why. I tried
to delete the element that was accessed - but this broke everything (UI froze
at "Connecting..."). I don't believe any elements were actually removed from
the map, so the new code just doesn't remove anything.
