aerogramme/src/imap/mailbox_view.rs

1135 lines
41 KiB
Rust
Raw Normal View History

use std::borrow::Cow;
2022-06-29 15:58:31 +00:00
use std::num::NonZeroU32;
use std::sync::Arc;
2022-06-29 17:24:21 +00:00
use anyhow::{anyhow, bail, Error, Result};
2022-06-29 14:05:34 +00:00
use boitalettres::proto::res::body::Data as Body;
use chrono::{Offset, TimeZone, Utc};
2022-06-29 17:24:21 +00:00
use futures::stream::{FuturesOrdered, StreamExt};
2022-06-29 18:00:38 +00:00
use imap_codec::types::address::Address;
use imap_codec::types::body::{BasicFields, Body as FetchBody, BodyStructure, SpecificFields};
use imap_codec::types::core::{AString, Atom, IString, NString};
use imap_codec::types::datetime::MyDateTime;
2022-06-29 18:00:38 +00:00
use imap_codec::types::envelope::Envelope;
2022-07-13 15:05:07 +00:00
use imap_codec::types::fetch_attributes::{
2022-07-15 15:55:04 +00:00
FetchAttribute, MacroOrFetchAttributes, Section as FetchSection,
2022-07-13 15:05:07 +00:00
};
2022-07-12 15:32:57 +00:00
use imap_codec::types::flag::{Flag, StoreResponse, StoreType};
2022-06-29 15:58:31 +00:00
use imap_codec::types::response::{Code, Data, MessageAttribute, Status};
2022-06-29 17:24:21 +00:00
use imap_codec::types::sequence::{self, SequenceSet};
use eml_codec::{
2023-09-21 09:27:33 +00:00
header,
imf,
part::{AnyPart, composite::Message},
2023-07-25 17:08:48 +00:00
mime::r#type::Deductible,
mime,
};
use crate::mail::mailbox::Mailbox;
2022-07-12 13:59:13 +00:00
use crate::mail::uidindex::{ImapUid, ImapUidvalidity, UidIndex};
2022-07-12 15:32:57 +00:00
use crate::mail::unique_ident::UniqueIdent;
const DEFAULT_FLAGS: [Flag; 5] = [
Flag::Seen,
Flag::Answered,
Flag::Flagged,
Flag::Deleted,
Flag::Draft,
];
const BODY_CHECK: &str = "body attribute asked but only header is fetched, logic error";
enum FetchedMail<'a> {
Partial(imf::Imf<'a>),
Full(Message<'a>),
}
impl<'a> FetchedMail<'a> {
fn as_full(&self) -> Result<&Message<'a>> {
match self {
FetchedMail::Full(x) => Ok(&x),
_ => bail!("The full message must be fetched, not only its headers: it's a logic error"),
}
}
fn imf(&self) -> &imf::Imf<'a> {
match self {
FetchedMail::Full(x) => &x.imf,
FetchedMail::Partial(x) => &x,
}
}
}
2023-09-28 09:57:46 +00:00
pub struct MailIdentifiers {
i: NonZeroU32,
uid: ImapUid,
uuid: UniqueIdent,
}
pub struct MailView<'a> {
ids: MailIdentifiers,
meta: MailMeta,
flags: Vec<Flag>,
content: FetchedMail<'a>,
}
pub struct AttributesProxy {
attrs: FetchAttributes
}
impl AttributeProxy {
fn new(attrs: &MacroOrFetchAttributes, is_uid_fetch: bool) -> Self {
// Expand macros
let mut fetch_attrs = match attributes {
MacroOrFetchAttributes::Macro(m) => m.expand(),
MacroOrFetchAttributes::FetchAttributes(a) => a.clone(),
};
// Handle uids
if *is_uid_fetch && !fetch_attrs.contains(&FetchAttribute::Uid) {
fetch_attrs.push(FetchAttribute::Uid);
}
Self { attrs: fetch_attrs }
}
fn need_body(&self) -> bool {
self.attrs.iter().any(|x| {
matches!(
x,
FetchAttribute::Body
| FetchAttribute::BodyExt { .. }
| FetchAttribute::Rfc822
| FetchAttribute::Rfc822Text
| FetchAttribute::BodyStructure
)
})
}
}
#[derive(Default)]
pub BatchMailViewBuilder<'a> struct {
attrs: AttributeProxy,
mi: Vec<MailIdentifiers>,
meta: Vec<MailMeta>,
flags: Vec<Vec<Flag>>,
bodies: Vec<FetchedMail<'a>>,
}
impl BatchMailViewBuilder<'a> {
fn new(req_attr: &MacroOrFetchAttributes) -> Self {
Self {
attrs: AttributeProxy::new(req_attr)
}
}
fn with_mail_identifiers(mut self, mi: Vec<MailIdentifiers>) -> Self {
self.mi = mi;
self
}
fn with_metadata(mut self, meta: Vec<MailMeta>) -> Self {
self.meta = meta;
self
}
fn with_flags(mut self, flags: Vec<Vec<Flag>>) -> Self {
self.flags = flags;
self
}
fn collect_bodies(mut self, FnOnce) -> Self {
if self.attrs.need_body() {
} else {
self.bodies =
}
self
}
fn build(self) -> Result<Vec<MailView>> {
}
}
/// A MailboxView is responsible for giving the client the information
/// it needs about a mailbox, such as an initial summary of the mailbox's
/// content and continuous updates indicating when the content
/// of the mailbox has been changed.
/// To do this, it keeps a variable `known_state` that corresponds to
/// what the client knows, and produces IMAP messages to be sent to the
/// client that go along updates to `known_state`.
pub struct MailboxView {
2022-07-12 14:35:11 +00:00
pub(crate) mailbox: Arc<Mailbox>,
known_state: UidIndex,
}
impl MailboxView {
/// Creates a new IMAP view into a mailbox.
/// Generates the necessary IMAP messages so that the client
/// has a satisfactory summary of the current mailbox's state.
/// These are the messages that are sent in response to a SELECT command.
pub async fn new(mailbox: Arc<Mailbox>) -> Result<(Self, Vec<Body>)> {
let state = mailbox.current_uid_index().await;
let new_view = Self {
mailbox,
known_state: state,
};
let mut data = Vec::<Body>::new();
2022-07-12 13:59:13 +00:00
data.push(new_view.exists_status()?);
data.push(new_view.recent_status()?);
data.extend(new_view.flags_status()?.into_iter());
data.push(new_view.uidvalidity_status()?);
data.push(new_view.uidnext_status()?);
Ok((new_view, data))
}
/// Produces a set of IMAP responses describing the change between
/// what the client knows and what is actually in the mailbox.
/// This does NOT trigger a sync, it bases itself on what is currently
/// loaded in RAM by Bayou.
pub async fn update(&mut self) -> Result<Vec<Body>> {
2022-06-29 15:58:31 +00:00
let new_view = MailboxView {
mailbox: self.mailbox.clone(),
known_state: self.mailbox.current_uid_index().await,
};
let mut data = Vec::<Body>::new();
2022-07-13 09:39:13 +00:00
// Calculate diff between two mailbox states
// See example in IMAP RFC in section on NOOP command:
// we want to produce something like this:
// C: a047 NOOP
// S: * 22 EXPUNGE
// S: * 23 EXISTS
// S: * 14 FETCH (UID 1305 FLAGS (\Seen \Deleted))
// S: a047 OK Noop completed
// In other words:
// - notify client of expunged mails
// - if new mails arrived, notify client of number of existing mails
// - if flags changed for existing mails, tell client
// (for this last step: if uidvalidity changed, do nothing,
// just notify of new uidvalidity and they will resync)
// - notify client of expunged mails
let mut n_expunge = 0;
for (i, (_uid, uuid)) in self.known_state.idx_by_uid.iter().enumerate() {
if !new_view.known_state.table.contains_key(uuid) {
data.push(Body::Data(Data::Expunge(
NonZeroU32::try_from((i + 1 - n_expunge) as u32).unwrap(),
)));
n_expunge += 1;
}
}
// - if new mails arrived, notify client of number of existing mails
if new_view.known_state.table.len() != self.known_state.table.len() - n_expunge
|| new_view.known_state.uidvalidity != self.known_state.uidvalidity
{
data.push(new_view.exists_status()?);
}
2022-06-29 15:58:31 +00:00
if new_view.known_state.uidvalidity != self.known_state.uidvalidity {
// TODO: do we want to push less/more info than this?
2022-07-12 13:59:13 +00:00
data.push(new_view.uidvalidity_status()?);
data.push(new_view.uidnext_status()?);
2022-06-29 15:58:31 +00:00
} else {
// - if flags changed for existing mails, tell client
2022-06-29 17:27:32 +00:00
for (i, (_uid, uuid)) in new_view.known_state.idx_by_uid.iter().enumerate() {
2022-06-29 15:58:31 +00:00
let old_mail = self.known_state.table.get(uuid);
let new_mail = new_view.known_state.table.get(uuid);
if old_mail.is_some() && old_mail != new_mail {
if let Some((uid, flags)) = new_mail {
data.push(Body::Data(Data::Fetch {
seq_or_uid: NonZeroU32::try_from((i + 1) as u32).unwrap(),
attributes: vec![
2023-05-15 16:23:23 +00:00
MessageAttribute::Uid(*uid),
2022-06-29 15:58:31 +00:00
MessageAttribute::Flags(
flags.iter().filter_map(|f| string_to_flag(f)).collect(),
),
],
}));
}
}
}
}
*self = new_view;
Ok(data)
}
2022-07-12 15:32:57 +00:00
pub async fn store(
&mut self,
sequence_set: &SequenceSet,
kind: &StoreType,
_response: &StoreResponse,
flags: &[Flag],
is_uid_store: &bool,
2022-07-12 15:32:57 +00:00
) -> Result<Vec<Body>> {
self.mailbox.opportunistic_sync().await?;
2022-07-12 15:32:57 +00:00
let flags = flags.iter().map(|x| x.to_string()).collect::<Vec<_>>();
let mails = self.get_mail_ids(sequence_set, *is_uid_store)?;
2022-07-13 09:19:08 +00:00
for (_i, _uid, uuid) in mails.iter() {
2022-07-12 15:32:57 +00:00
match kind {
StoreType::Add => {
self.mailbox.add_flags(*uuid, &flags[..]).await?;
}
StoreType::Remove => {
self.mailbox.del_flags(*uuid, &flags[..]).await?;
}
StoreType::Replace => {
2022-07-13 09:00:35 +00:00
self.mailbox.set_flags(*uuid, &flags[..]).await?;
2022-07-12 15:32:57 +00:00
}
}
}
2022-07-21 10:44:58 +00:00
// @TODO: handle _response
2022-07-12 15:32:57 +00:00
self.update().await
}
2022-07-13 09:00:35 +00:00
pub async fn expunge(&mut self) -> Result<Vec<Body>> {
self.mailbox.opportunistic_sync().await?;
2022-07-13 09:19:08 +00:00
let deleted_flag = Flag::Deleted.to_string();
let state = self.mailbox.current_uid_index().await;
let msgs = state
2022-07-13 09:19:08 +00:00
.table
.iter()
.filter(|(_uuid, (_uid, flags))| flags.iter().any(|x| *x == deleted_flag))
.map(|(uuid, _)| *uuid);
for msg in msgs {
self.mailbox.delete(msg).await?;
}
self.update().await
2022-07-13 09:00:35 +00:00
}
2022-07-21 10:44:58 +00:00
pub async fn copy(
&self,
sequence_set: &SequenceSet,
to: Arc<Mailbox>,
is_uid_copy: &bool,
) -> Result<(ImapUidvalidity, Vec<(ImapUid, ImapUid)>)> {
let mails = self.get_mail_ids(sequence_set, *is_uid_copy)?;
let mut new_uuids = vec![];
for (_i, _uid, uuid) in mails.iter() {
new_uuids.push(to.copy_from(&self.mailbox, *uuid).await?);
}
let mut ret = vec![];
let to_state = to.current_uid_index().await;
for ((_i, uid, _uuid), new_uuid) in mails.iter().zip(new_uuids.iter()) {
let dest_uid = to_state
.table
.get(new_uuid)
.ok_or(anyhow!("copied mail not in destination mailbox"))?
.0;
ret.push((*uid, dest_uid));
}
Ok((to_state.uidvalidity, ret))
}
2022-06-29 17:24:21 +00:00
/// Looks up state changes in the mailbox and produces a set of IMAP
/// responses describing the new state.
pub async fn fetch(
&self,
sequence_set: &SequenceSet,
attributes: &MacroOrFetchAttributes,
is_uid_fetch: &bool,
2022-06-29 17:24:21 +00:00
) -> Result<Vec<Body>> {
2023-09-28 09:57:46 +00:00
/*
let mails = self.get_mail_ids(sequence_set, *is_uid_fetch)?;
let mails_uuid = mails
.iter()
.map(|mi| mi.uuid)
.collect::<Vec<_>>();
let mails_meta = self.mailbox.fetch_meta(&mails_uuid).await?;
*/
// mail identifiers
let mails = self.get_mail_ids(sequence_set, *is_uid_fetch)?;
2022-06-29 17:24:21 +00:00
let mails_uuid = mails
.iter()
2022-07-12 15:32:57 +00:00
.map(|(_i, _uid, uuid)| *uuid)
2022-06-29 17:24:21 +00:00
.collect::<Vec<_>>();
2023-09-28 09:57:46 +00:00
// mail metadata
let mails_meta = self.mailbox.fetch_meta(&mails_uuid).await?;
2022-06-29 17:24:21 +00:00
2023-09-28 09:57:46 +00:00
// fectch email body and transform to a state usable by select
2022-06-29 17:24:21 +00:00
let mails = if need_body {
let mut iter = mails
.into_iter()
.zip(mails_meta.into_iter())
2022-07-12 15:32:57 +00:00
.map(|((i, uid, uuid), meta)| async move {
2022-06-29 17:24:21 +00:00
let body = self.mailbox.fetch_full(uuid, &meta.message_key).await?;
Ok::<_, anyhow::Error>((i, uid, uuid, meta, Some(body)))
})
.collect::<FuturesOrdered<_>>();
let mut mails = vec![];
while let Some(m) = iter.next().await {
mails.push(m?);
}
mails
} else {
mails
.into_iter()
.zip(mails_meta.into_iter())
2022-07-12 15:32:57 +00:00
.map(|((i, uid, uuid), meta)| (i, uid, uuid, meta, None))
2022-06-29 17:24:21 +00:00
.collect::<Vec<_>>()
};
2023-09-28 09:57:46 +00:00
// add flags
let mails = mails.into_iter().filter_map(|(i, uid, uuid, meta, body)|
self
2022-06-29 17:24:21 +00:00
.known_state
.table
.get(&uuid)
2023-09-28 09:57:46 +00:00
.map(|(_uid2, flags)| (i, uid, uuid, meta, flags, body)))
.collect::<Vec<_>>();
2022-06-29 17:24:21 +00:00
2023-09-28 09:57:46 +00:00
// parse email body / headers
let fetched = match &body {
2022-06-29 18:00:38 +00:00
Some(m) => {
FetchedMail::Full(eml_codec::parse_message(m).or(Err(anyhow!("Invalid mail body")))?.1)
}
None => {
FetchedMail::Partial(eml_codec::parse_imf(&meta.headers).or(Err(anyhow!("Invalid mail headers")))?.1)
2022-06-29 18:00:38 +00:00
}
};
2023-09-28 09:57:46 +00:00
// do the logic
select_mail_fragments(fetch_attrs, mails)
}
todo!();
let mut ret = vec![];
for (i, uid, uuid, meta, flags, body) in mails {
let mut attributes = vec![];
2022-06-29 17:24:21 +00:00
for attr in fetch_attrs.iter() {
match attr {
2022-07-13 09:32:47 +00:00
FetchAttribute::Uid => attributes.push(MessageAttribute::Uid(uid)),
2022-06-29 17:24:21 +00:00
FetchAttribute::Flags => {
attributes.push(MessageAttribute::Flags(
flags.iter().filter_map(|f| string_to_flag(f)).collect(),
));
}
FetchAttribute::Rfc822Size => {
attributes.push(MessageAttribute::Rfc822Size(meta.rfc822_size as u32))
}
2022-07-15 14:15:48 +00:00
FetchAttribute::Rfc822Header => {
attributes.push(MessageAttribute::Rfc822Header(NString(
meta.headers.to_vec().try_into().ok().map(IString::Literal),
)))
}
2022-06-30 11:36:21 +00:00
FetchAttribute::Rfc822Text => {
2022-07-15 14:15:48 +00:00
attributes.push(MessageAttribute::Rfc822Text(NString(
fetched.as_full()?.raw_body.try_into().ok().map(IString::Literal),
2022-07-15 14:15:48 +00:00
)));
2022-06-29 18:10:42 +00:00
}
2022-07-15 14:15:48 +00:00
FetchAttribute::Rfc822 => attributes.push(MessageAttribute::Rfc822(NString(
body.as_ref()
.expect(BODY_CHECK)
2022-07-15 14:15:48 +00:00
.clone()
.try_into()
.ok()
.map(IString::Literal),
))),
2022-06-29 18:00:38 +00:00
FetchAttribute::Envelope => {
attributes.push(MessageAttribute::Envelope(message_envelope(fetched.imf())))
2022-06-29 18:00:38 +00:00
}
2022-07-04 16:14:19 +00:00
FetchAttribute::Body => attributes.push(MessageAttribute::Body(
build_imap_email_struct(fetched.as_full()?.child.as_ref())?,
2022-07-04 16:14:19 +00:00
)),
FetchAttribute::BodyStructure => attributes.push(MessageAttribute::Body(
build_imap_email_struct(fetched.as_full()?.child.as_ref())?,
2022-07-04 16:14:19 +00:00
)),
// maps to BODY[<section>]<<partial>> and BODY.PEEK[<section>]<<partial>>
// peek does not implicitly set the \Seen flag
//
// eg. BODY[HEADER.FIELDS (DATE FROM)]
// eg. BODY[]<0.2048>
2022-06-30 11:36:21 +00:00
FetchAttribute::BodyExt {
2022-07-13 15:05:07 +00:00
section,
partial,
peek,
2022-06-30 11:36:21 +00:00
} => {
2022-07-13 15:05:07 +00:00
// @TODO Add missing section specifiers
// Extract message section
match get_message_section(fetched.as_full()?, section) {
2022-07-15 14:15:48 +00:00
Ok(text) => {
let seen_flag = Flag::Seen.to_string();
if !peek && !flags.iter().any(|x| *x == seen_flag) {
// Add \Seen flag
self.mailbox.add_flags(uuid, &[seen_flag]).await?;
}
// Handle <<partial>> which cut the message byts
2022-07-15 14:15:48 +00:00
let (text, origin) = match partial {
Some((begin, len)) => {
if *begin as usize > text.len() {
(&[][..], Some(*begin))
} else if (*begin + len.get()) as usize >= text.len() {
(&text[*begin as usize..], Some(*begin))
} else {
(
&text[*begin as usize
..(*begin + len.get()) as usize],
Some(*begin),
)
}
}
None => (&text[..], None),
};
let data =
NString(text.to_vec().try_into().ok().map(IString::Literal));
attributes.push(MessageAttribute::BodyExt {
section: section.clone(),
origin,
data,
})
2022-07-13 15:05:07 +00:00
}
2022-07-15 14:15:48 +00:00
Err(e) => {
tracing::error!(
"Could not get section {:?} of message {}: {}",
section,
uuid,
e
);
2022-07-13 15:05:07 +00:00
}
}
2022-06-30 11:36:21 +00:00
}
FetchAttribute::InternalDate => {
2023-05-15 16:23:23 +00:00
let dt = Utc.fix().timestamp_opt(i64::try_from(meta.internaldate / 1000)?, 0).earliest().ok_or(anyhow!("Unable to parse internal date"))?;
attributes.push(MessageAttribute::InternalDate(MyDateTime(dt)));
2022-06-30 11:36:21 +00:00
}
2022-06-29 17:24:21 +00:00
}
}
ret.push(Body::Data(Data::Fetch {
seq_or_uid: i,
attributes,
}));
}
Ok(ret)
}
// ----
2023-09-28 09:57:46 +00:00
// Gets the IMAP ID, the IMAP UIDs and, the Aerogramme UUIDs of mails identified by a SequenceSet of
// sequence numbers (~ IMAP selector)
2022-07-12 15:32:57 +00:00
fn get_mail_ids(
&self,
sequence_set: &SequenceSet,
by_uid: bool,
2022-07-12 15:32:57 +00:00
) -> Result<Vec<(NonZeroU32, ImapUid, UniqueIdent)>> {
let mail_vec = self
.known_state
.idx_by_uid
.iter()
.map(|(uid, uuid)| (*uid, *uuid))
.collect::<Vec<_>>();
let mut mails = vec![];
if by_uid {
if mail_vec.is_empty() {
return Ok(vec![]);
}
let iter_strat = sequence::Strategy::Naive {
largest: mail_vec.last().unwrap().0,
};
let mut i = 0;
for uid in sequence_set.iter(iter_strat) {
while mail_vec.get(i).map(|mail| mail.0 < uid).unwrap_or(false) {
i += 1;
}
if let Some(mail) = mail_vec.get(i) {
if mail.0 == uid {
mails.push((NonZeroU32::try_from(i as u32 + 1).unwrap(), mail.0, mail.1));
}
} else {
break;
}
}
} else {
if mail_vec.is_empty() {
bail!("No such message (mailbox is empty)");
}
let iter_strat = sequence::Strategy::Naive {
largest: NonZeroU32::try_from((mail_vec.len()) as u32).unwrap(),
};
for i in sequence_set.iter(iter_strat) {
if let Some(mail) = mail_vec.get(i.get() as usize - 1) {
mails.push((i, mail.0, mail.1));
} else {
bail!("No such mail: {}", i);
}
2022-07-12 15:32:57 +00:00
}
}
Ok(mails)
}
// ----
/// Produce an OK [UIDVALIDITY _] message corresponding to `known_state`
2022-07-12 13:59:13 +00:00
fn uidvalidity_status(&self) -> Result<Body> {
let uid_validity = Status::ok(
None,
2022-07-12 13:59:13 +00:00
Some(Code::UidValidity(self.uidvalidity())),
"UIDs valid",
)
.map_err(Error::msg)?;
Ok(Body::Status(uid_validity))
}
2022-07-12 13:59:13 +00:00
pub(crate) fn uidvalidity(&self) -> ImapUidvalidity {
self.known_state.uidvalidity
}
/// Produce an OK [UIDNEXT _] message corresponding to `known_state`
2022-07-12 13:59:13 +00:00
fn uidnext_status(&self) -> Result<Body> {
let next_uid = Status::ok(
None,
2022-07-12 13:59:13 +00:00
Some(Code::UidNext(self.uidnext())),
"Predict next UID",
)
.map_err(Error::msg)?;
Ok(Body::Status(next_uid))
}
2022-07-12 13:59:13 +00:00
pub(crate) fn uidnext(&self) -> ImapUid {
self.known_state.uidnext
}
/// Produce an EXISTS message corresponding to the number of mails
/// in `known_state`
2022-07-12 13:59:13 +00:00
fn exists_status(&self) -> Result<Body> {
Ok(Body::Data(Data::Exists(self.exists()?)))
}
pub(crate) fn exists(&self) -> Result<u32> {
Ok(u32::try_from(self.known_state.idx_by_uid.len())?)
}
/// Produce a RECENT message corresponding to the number of
/// recent mails in `known_state`
2022-07-12 13:59:13 +00:00
fn recent_status(&self) -> Result<Body> {
Ok(Body::Data(Data::Recent(self.recent()?)))
}
pub(crate) fn recent(&self) -> Result<u32> {
let recent = self
.known_state
.idx_by_flag
.get(&"\\Recent".to_string())
.map(|os| os.len())
.unwrap_or(0);
2022-07-12 13:59:13 +00:00
Ok(u32::try_from(recent)?)
}
/// Produce a FLAGS and a PERMANENTFLAGS message that indicates
/// the flags that are in `known_state` + default flags
2022-07-12 13:59:13 +00:00
fn flags_status(&self) -> Result<Vec<Body>> {
let mut flags: Vec<Flag> = self
.known_state
.idx_by_flag
.flags()
2023-05-15 16:23:23 +00:00
.filter_map(|f| string_to_flag(f))
.collect();
for f in DEFAULT_FLAGS.iter() {
if !flags.contains(f) {
flags.push(f.clone());
}
}
let mut ret = vec![Body::Data(Data::Flags(flags.clone()))];
flags.push(Flag::Permanent);
let permanent_flags =
Status::ok(None, Some(Code::PermanentFlags(flags)), "Flags permitted")
.map_err(Error::msg)?;
ret.push(Body::Status(permanent_flags));
Ok(ret)
}
pub(crate) fn unseen_count(&self) -> usize {
let total = self.known_state.table.len();
let seen = self
.known_state
.idx_by_flag
.get(&Flag::Seen.to_string())
.map(|x| x.len())
.unwrap_or(0);
total - seen
}
}
2022-06-29 15:58:31 +00:00
fn string_to_flag(f: &str) -> Option<Flag> {
match f.chars().next() {
2022-07-12 15:32:57 +00:00
Some('\\') => match f {
"\\Seen" => Some(Flag::Seen),
"\\Answered" => Some(Flag::Answered),
"\\Flagged" => Some(Flag::Flagged),
"\\Deleted" => Some(Flag::Deleted),
"\\Draft" => Some(Flag::Draft),
"\\Recent" => Some(Flag::Recent),
2023-05-15 16:23:23 +00:00
_ => match Atom::try_from(f.strip_prefix('\\').unwrap().to_string()) {
2022-07-12 15:32:57 +00:00
Err(_) => {
tracing::error!(flag=%f, "Unable to encode flag as IMAP atom");
None
}
Ok(a) => Some(Flag::Extension(a)),
},
},
2023-05-15 16:23:23 +00:00
Some(_) => match Atom::try_from(f.to_string()) {
2022-06-29 15:58:31 +00:00
Err(_) => {
tracing::error!(flag=%f, "Unable to encode flag as IMAP atom");
None
}
Ok(a) => Some(Flag::Keyword(a)),
},
None => None,
}
}
2022-06-29 18:00:38 +00:00
/// Envelope rules are defined in RFC 3501, section 7.4.2
/// https://datatracker.ietf.org/doc/html/rfc3501#section-7.4.2
///
/// Some important notes:
///
/// If the Sender or Reply-To lines are absent in the [RFC-2822]
/// header, or are present but empty, the server sets the
/// corresponding member of the envelope to be the same value as
/// the from member (the client is not expected to know to do
/// this). Note: [RFC-2822] requires that all messages have a valid
/// From header. Therefore, the from, sender, and reply-to
/// members in the envelope can not be NIL.
///
/// If the Date, Subject, In-Reply-To, and Message-ID header lines
/// are absent in the [RFC-2822] header, the corresponding member
/// of the envelope is NIL; if these header lines are present but
/// empty the corresponding member of the envelope is the empty
/// string.
//@FIXME return an error if the envelope is invalid instead of panicking
//@FIXME some fields must be defaulted if there are not set.
fn message_envelope(msg: &imf::Imf) -> Envelope {
2023-07-25 17:08:48 +00:00
let from = msg.from.iter().map(convert_mbx).collect::<Vec<_>>();
2022-06-29 18:00:38 +00:00
Envelope {
date: NString(
msg.date.as_ref()
2023-02-07 10:27:17 +00:00
.map(|d| IString::try_from(d.to_rfc3339()).unwrap()),
2022-06-29 18:00:38 +00:00
),
subject: NString(
msg.subject.as_ref()
2022-06-29 18:00:38 +00:00
.map(|d| IString::try_from(d.to_string()).unwrap()),
),
2023-07-25 17:08:48 +00:00
sender: msg.sender.as_ref().map(|v| vec![convert_mbx(v)]).unwrap_or(from.clone()),
reply_to: if msg.reply_to.is_empty() {
from.clone()
} else {
convert_addresses(&msg.reply_to)
},
from: from,
to: convert_addresses(&msg.to),
cc: convert_addresses(&msg.cc),
bcc: convert_addresses(&msg.bcc),
in_reply_to: NString(msg.in_reply_to.iter().next().map(|d| IString::try_from(d.to_string()).unwrap())),
2022-06-29 18:00:38 +00:00
message_id: NString(
msg.msg_id.as_ref().map(|d| IString::try_from(d.to_string()).unwrap()),
2022-06-29 18:00:38 +00:00
),
}
}
fn convert_addresses(addrlist: &Vec<imf::address::AddressRef>) -> Vec<Address> {
let mut acc = vec![];
for item in addrlist {
match item {
imf::address::AddressRef::Single(a) => acc.push(convert_mbx(a)),
imf::address::AddressRef::Many(l) => acc.extend(l.participants.iter().map(convert_mbx))
2022-06-29 18:00:38 +00:00
}
}
return acc
2022-06-29 18:00:38 +00:00
}
fn convert_mbx(addr: &imf::mailbox::MailboxRef) -> Address {
2022-06-29 18:00:38 +00:00
Address::new(
NString(addr.name.as_ref().map(|x| IString::try_from(x.to_string()).unwrap())),
// SMTP at-domain-list (source route) seems obsolete since at least 1991
// https://www.mhonarc.org/archive/html/ietf-822/1991-06/msg00060.html
2022-06-29 18:00:38 +00:00
NString(None),
NString(Some(IString::try_from(addr.addrspec.local_part.to_string()).unwrap())),
NString(Some(IString::try_from(addr.addrspec.domain.to_string()).unwrap())),
2022-06-29 18:00:38 +00:00
)
}
2022-07-04 16:14:19 +00:00
/*
--CAPTURE--
b fetch 29878:29879 (BODY)
* 29878 FETCH (BODY (("text" "plain" ("charset" "utf-8") NIL NIL "quoted-printable" 3264 82)("text" "html" ("charset" "utf-8") NIL NIL "quoted-printable" 31834 643) "alternative"))
* 29879 FETCH (BODY ("text" "html" ("charset" "us-ascii") NIL NIL "7bit" 4107 131))
^^^^^^^^^^^^^^^^^^^^^^ ^^^ ^^^ ^^^^^^ ^^^^ ^^^
| | | | | | number of lines
| | | | | size
| | | | content transfer encoding
| | | description
| | id
| parameter list
b OK Fetch completed (0.001 + 0.000 secs).
*/
fn build_imap_email_struct<'a>(part: &AnyPart<'a>) -> Result<BodyStructure> {
match part {
AnyPart::Mult(x) => {
2023-07-25 17:08:48 +00:00
let itype = &x.mime.interpreted_type;
let subtype = IString::try_from(itype.subtype.to_string()).unwrap_or(unchecked_istring("alternative"));
2022-07-20 13:14:34 +00:00
Ok(BodyStructure::Multi {
bodies: x.children
2022-07-20 13:14:34 +00:00
.iter()
.filter_map(|inner| build_imap_email_struct(&inner).ok())
.collect(),
2022-07-20 13:14:34 +00:00
subtype,
extension_data: None,
/*Some(MultipartExtensionData {
parameter_list: vec![],
disposition: None,
language: None,
location: None,
extension: vec![],
})*/
})
}
AnyPart::Txt(x) => {
2023-07-25 17:08:48 +00:00
let mut basic = basic_fields(&x.mime.fields, x.body.len())?;
2023-07-25 17:08:48 +00:00
// Get the interpreted content type, set it
let itype = match &x.mime.interpreted_type {
Deductible::Inferred(v) | Deductible::Explicit(v) => v
};
let subtype = IString::try_from(itype.subtype.to_string()).unwrap_or(unchecked_istring("plain"));
// Add charset to the list of parameters if we know it has been inferred as it will be
// missing from the parsed content.
if let Deductible::Inferred(charset) = &itype.charset {
basic.parameter_list.push((
unchecked_istring("charset"),
IString::try_from(charset.to_string()).unwrap_or(unchecked_istring("us-ascii"))
));
}
2022-07-20 13:14:34 +00:00
Ok(BodyStructure::Single {
body: FetchBody {
basic,
specific: SpecificFields::Text {
subtype,
2023-07-25 17:08:48 +00:00
number_of_lines: nol(x.body),
2022-07-20 13:14:34 +00:00
},
},
extension: None,
})
}
AnyPart::Bin(x) => {
2023-07-25 17:08:48 +00:00
let basic = basic_fields(&x.mime.fields, x.body.len())?;
let default = mime::r#type::NaiveType { main: &b"application"[..], sub: &b"octet-stream"[..], params: vec![] };
let ct = x.mime.fields.ctype.as_ref().unwrap_or(&default);
let type_ = IString::try_from(String::from_utf8_lossy(ct.main).to_string())
.or(Err(anyhow!("Unable to build IString from given Content-Type type given")))?;
let subtype = IString::try_from(String::from_utf8_lossy(ct.sub).to_string())
.or(Err(anyhow!("Unable to build IString from given Content-Type subtype given")))?;
2022-07-20 13:14:34 +00:00
Ok(BodyStructure::Single {
body: FetchBody {
basic,
specific: SpecificFields::Basic { type_, subtype },
},
extension: None,
})
}
AnyPart::Msg(x) => {
2023-07-25 17:08:48 +00:00
let basic = basic_fields(&x.mime.fields, x.raw_part.len())?;
2023-02-07 10:27:17 +00:00
Ok(BodyStructure::Single {
body: FetchBody {
basic,
specific: SpecificFields::Message {
envelope: message_envelope(&x.imf),
body_structure: Box::new(build_imap_email_struct(x.child.as_ref())?),
2023-07-25 17:08:48 +00:00
number_of_lines: nol(x.raw_part),
2023-02-07 10:27:17 +00:00
},
},
extension: None,
})
}
2022-07-20 13:14:34 +00:00
}
}
2023-07-25 17:08:48 +00:00
fn nol(input: &[u8]) -> u32 {
input.iter()
.filter(|x| **x == b'\n')
.count()
.try_into()
.unwrap_or(0)
}
/// s is set to static to ensure that only compile time values
2022-07-05 15:08:12 +00:00
/// checked by developpers are passed.
fn unchecked_istring(s: &'static str) -> IString {
IString::try_from(s).expect("this value is expected to be a valid imap-codec::IString")
}
2023-07-25 17:08:48 +00:00
fn basic_fields(m: &mime::NaiveMIME, sz: usize) -> Result<BasicFields> {
let parameter_list = m.ctype
.as_ref()
.map(|x| x.params.iter()
.map(|p| (IString::try_from(String::from_utf8_lossy(p.name).to_string()), IString::try_from(p.value.to_string())))
.filter(|(k, v)| k.is_ok() && v.is_ok())
.map(|(k, v)| (k.unwrap(), v.unwrap()))
.collect())
.unwrap_or(vec![]);
Ok(BasicFields {
2022-07-05 15:08:12 +00:00
parameter_list,
id: NString(
m.id.as_ref()
2023-05-15 16:23:23 +00:00
.and_then(|ci| IString::try_from(ci.to_string()).ok()),
),
2022-07-05 15:08:12 +00:00
description: NString(
m.description.as_ref()
2023-05-15 16:23:23 +00:00
.and_then(|cd| IString::try_from(cd.to_string()).ok()),
2022-07-05 15:08:12 +00:00
),
content_transfer_encoding: match m.transfer_encoding {
mime::mechanism::Mechanism::_8Bit => unchecked_istring("8bit"),
mime::mechanism::Mechanism::Binary => unchecked_istring("binary"),
mime::mechanism::Mechanism::QuotedPrintable => unchecked_istring("quoted-printable"),
mime::mechanism::Mechanism::Base64 => unchecked_istring("base64"),
_ => unchecked_istring("7bit"),
},
// @FIXME we can't compute the size of the message currently...
2023-07-25 17:08:48 +00:00
size: u32::try_from(sz)?,
})
2022-07-05 15:08:12 +00:00
}
/// Extract message section for section identifier passed by the FETCH BODY[<section>]<<partial>>
/// request
///
/// Example of message sections:
///
/// ```
/// HEADER ([RFC-2822] header of the message)
/// TEXT ([RFC-2822] text body of the message) MULTIPART/MIXED
/// 1 TEXT/PLAIN
/// 2 APPLICATION/OCTET-STREAM
/// 3 MESSAGE/RFC822
/// 3.HEADER ([RFC-2822] header of the message)
/// 3.TEXT ([RFC-2822] text body of the message) MULTIPART/MIXED
/// 3.1 TEXT/PLAIN
/// 3.2 APPLICATION/OCTET-STREAM
/// 4 MULTIPART/MIXED
/// 4.1 IMAGE/GIF
/// 4.1.MIME ([MIME-IMB] header for the IMAGE/GIF)
/// 4.2 MESSAGE/RFC822
/// 4.2.HEADER ([RFC-2822] header of the message)
/// 4.2.TEXT ([RFC-2822] text body of the message) MULTIPART/MIXED
/// 4.2.1 TEXT/PLAIN
/// 4.2.2 MULTIPART/ALTERNATIVE
/// 4.2.2.1 TEXT/PLAIN
/// 4.2.2.2 TEXT/RICHTEXT
/// ```
2022-07-15 14:15:48 +00:00
fn get_message_section<'a>(
parsed: &'a Message<'a>,
section: &Option<FetchSection>,
) -> Result<Cow<'a, [u8]>> {
match section {
2022-07-20 13:14:34 +00:00
Some(FetchSection::Text(None)) => {
Ok(parsed.raw_body.into())
2022-07-20 13:14:34 +00:00
}
2022-07-15 14:15:48 +00:00
Some(FetchSection::Text(Some(part))) => {
map_subpart(parsed.child.as_ref(), part.0.as_slice(), |part_msg| {
Ok(part_msg.as_message().ok_or(Error::msg("Not a message/rfc822 part while expected by request (xxx.TEXT)"))?
.raw_body
2022-07-15 14:15:48 +00:00
.into())
})
}
Some(FetchSection::Header(part)) => map_subpart(
parsed.child.as_ref(),
2022-07-15 14:15:48 +00:00
part.as_ref().map(|p| p.0.as_slice()).unwrap_or(&[]),
|part_msg| {
Ok(part_msg.as_message().ok_or(Error::msg("Not a message/rfc822 part while expected by request (xxx.TEXT)"))?
.raw_headers
2022-07-15 14:15:48 +00:00
.into())
},
),
Some(
FetchSection::HeaderFields(part, fields) | FetchSection::HeaderFieldsNot(part, fields),
) => {
2023-09-21 09:27:33 +00:00
let invert = matches!(section, Some(FetchSection::HeaderFieldsNot(_, _)));
let fields = fields
.iter()
.map(|x| match x {
AString::Atom(a) => a.as_bytes(),
AString::String(IString::Literal(l)) => l.as_slice(),
AString::String(IString::Quoted(q)) => q.as_bytes(),
})
.collect::<Vec<_>>();
2023-09-21 09:27:33 +00:00
map_subpart(
parsed.child.as_ref(),
part.as_ref().map(|p| p.0.as_slice()).unwrap_or(&[]),
|part_msg| {
let mut ret = vec![];
2023-09-21 09:27:33 +00:00
for f in &part_msg.mime().kv {
let (k, v) = match f {
header::Field::Good(header::Kv2(k, v)) => (k, v),
_ => continue,
};
if fields
.as_slice()
.iter()
2023-09-21 09:27:33 +00:00
.any(|x| (x == k) ^ invert)
{
2023-09-21 09:27:33 +00:00
ret.extend(*k);
ret.extend(b": ");
2023-09-21 09:27:33 +00:00
ret.extend(*v);
ret.extend(b"\r\n");
}
}
ret.extend(b"\r\n");
Ok(ret.into())
},
2023-09-21 09:27:33 +00:00
)
}
Some(FetchSection::Part(part)) => map_subpart(parsed.child.as_ref(), part.0.as_slice(), |part| {
let bytes = match &part {
AnyPart::Txt(p) => p.body,
AnyPart::Bin(p) => p.body,
AnyPart::Msg(p) => p.raw_part,
AnyPart::Mult(_) => bail!("Multipart part has no body"),
2022-07-15 14:15:48 +00:00
};
Ok(bytes.to_vec().into())
2022-07-15 14:15:48 +00:00
}),
Some(FetchSection::Mime(part)) => map_subpart(parsed.child.as_ref(), part.0.as_slice(), |part| {
let bytes = match &part {
AnyPart::Txt(p) => p.mime.fields.raw,
AnyPart::Bin(p) => p.mime.fields.raw,
AnyPart::Msg(p) => p.mime.fields.raw,
AnyPart::Mult(p) => p.mime.fields.raw,
};
Ok(bytes.to_vec().into())
2022-07-15 14:15:48 +00:00
}),
None => Ok(parsed.raw_part.into()),
2022-07-15 14:15:48 +00:00
}
}
/// Fetch a MIME SubPart
///
/// eg. FETCH BODY[4.2.2.1] -> [4, 2, 2, 1]
fn map_subpart<'a, F, R>(part: &AnyPart<'a>, path: &[NonZeroU32], f: F) -> Result<R>
2022-07-15 14:15:48 +00:00
where
F: FnOnce(&AnyPart<'a>) -> Result<R>,
2022-07-15 14:15:48 +00:00
{
if path.is_empty() {
f(part)
2022-07-15 14:15:48 +00:00
} else {
match part {
AnyPart::Mult(x) => map_subpart(
x.children
.get(path[0].get() as usize - 1)
.ok_or(anyhow!("Unable to resolve subpath {:?}, current multipart has only {} elements", path, x.children.len()))?,
&path[1..],
f),
AnyPart::Msg(x) => map_subpart(x.child.as_ref(), path, f),
_ => bail!("You tried to access a subpart on an atomic part (text or binary). Unresolved subpath {:?}", path),
2022-07-15 14:15:48 +00:00
}
}
}
#[cfg(test)]
mod tests {
use super::*;
2022-07-05 13:21:14 +00:00
use imap_codec::codec::Encode;
use std::fs;
2022-07-05 16:27:49 +00:00
/// Future automated test. We use lossy utf8 conversion + lowercase everything,
2022-07-05 13:21:14 +00:00
/// so this test might allow invalid results. But at least it allows us to quickly test a
/// large variety of emails.
/// Keep in mind that special cases must still be tested manually!
#[test]
2022-07-05 13:21:14 +00:00
fn fetch_body() -> Result<()> {
2022-07-05 15:48:10 +00:00
let prefixes = [
2023-07-25 17:08:48 +00:00
/* *** MY OWN DATASET *** */
2022-07-05 15:48:10 +00:00
"tests/emails/dxflrs/0001_simple",
"tests/emails/dxflrs/0002_mime",
2022-07-05 16:27:49 +00:00
"tests/emails/dxflrs/0003_mime-in-mime",
2022-07-20 13:14:34 +00:00
"tests/emails/dxflrs/0004_msg-in-msg",
2023-07-25 17:08:48 +00:00
// eml_codec do not support continuation for the moment
2022-07-20 11:58:24 +00:00
//"tests/emails/dxflrs/0005_mail-parser-readme",
2023-03-09 10:30:44 +00:00
"tests/emails/dxflrs/0006_single-mime",
2023-07-25 17:08:48 +00:00
"tests/emails/dxflrs/0007_raw_msg_in_rfc822",
/* *** (STRANGE) RFC *** */
//"tests/emails/rfc/000", // must return text/enriched, we return text/plain
//"tests/emails/rfc/001", // does not recognize the multipart/external-body, breaks the
// whole parsing
//"tests/emails/rfc/002", // wrong date in email
//"tests/emails/rfc/003", // dovecot fixes \r\r: the bytes number is wrong + text/enriched
/* *** THIRD PARTY *** */
//"tests/emails/thirdparty/000", // dovecot fixes \r\r: the bytes number is wrong
//"tests/emails/thirdparty/001", // same
"tests/emails/thirdparty/002", // same
2023-03-09 10:30:44 +00:00
2023-07-25 17:08:48 +00:00
/* *** LEGACY *** */
//"tests/emails/legacy/000", // same issue with \r\r
2022-07-05 15:48:10 +00:00
];
2022-07-05 15:48:10 +00:00
for pref in prefixes.iter() {
println!("{}", pref);
let txt = fs::read(format!("{}.eml", pref))?;
2022-07-08 08:23:07 +00:00
let exp = fs::read(format!("{}.dovecot.body", pref))?;
2023-07-25 17:08:48 +00:00
let message = eml_codec::parse_message(&txt).unwrap().1;
2022-07-05 15:48:10 +00:00
let mut resp = Vec::new();
2023-07-25 17:08:48 +00:00
MessageAttribute::Body(build_imap_email_struct(&message.child)?).encode(&mut resp).unwrap();
2022-07-05 15:48:10 +00:00
let resp_str = String::from_utf8_lossy(&resp).to_lowercase();
2022-07-05 13:21:14 +00:00
2022-07-05 15:48:10 +00:00
let exp_no_parenthesis = &exp[1..exp.len() - 1];
let exp_str = String::from_utf8_lossy(exp_no_parenthesis).to_lowercase();
2023-02-07 11:22:54 +00:00
println!("aerogramme: {}\n\ndovecot: {}\n\n", resp_str, exp_str);
2022-07-08 08:23:07 +00:00
//println!("\n\n {} \n\n", String::from_utf8_lossy(&resp));
2022-07-05 15:48:10 +00:00
assert_eq!(resp_str, exp_str);
}
Ok(())
}
}