585 Commits
Author SHA1 Message Date
Daniel Eggert 6b1b196cd6 Use NFC for ModifiedUTF7.encode (#884) 2026-11-13 13:38:17 +01:00
Daniel Eggert b2c94a8bce Only skip an orphaned LF after a response-ending CR (#882) 2026-11-10 17:49:13 +01:00
Ravindu Danthanarayana 95590ebc81 Fix corruption of FETCH literals when the header CRLF is split (#880) 2026-11-04 20:57:20 +01:00
Daniel Eggert d67ffdf153 Fix encode/parse round-trip bugs (#868)
* Separate LIST from its independent select options

`.listIndependent([.remote], …)` encoded as `LIST(REMOTE) …`.

RFC 5258: `list = "LIST" [SP list-select-opts] SP mailbox SP mbox-or-pat`.

* Keep leading zeros in fractional seconds

`12:34:56.05` parsed as `fraction: 5` and re-encoded as `12:34:56.5`,
changing the value.

RFC 3339: `time-secfrac = "." 1*DIGIT`. `FullTime.fraction` becomes
`FullTime.Fraction`, which stores the digits; `init?(_:)` validates,
literals trap, `String.init` reads.

* Model the RFC 3339 time offset

`;EXPIRE=2025-12-31T23:59:59Z` failed to parse inside a URL: the
parser stopped before `Z`. The encoder never wrote an offset.

RFC 4467: `expire = ";EXPIRE=" date-time`; RFC 3339: `full-time =
partial-time time-offset`, `time-offset = "Z" / time-numoffset`. Adds
the required `FullTime.offset`.

* Encode LIST extended data

`MailboxInfo.extensions` was parsed but never written, so
`* LIST (\Marked) "/" "Fruit" ("CHILDINFO" ("SUBSCRIBED"))` lost
`("CHILDINFO" …)` on re-encode.

RFC 5258: `mailbox-list = … mailbox [SP mbox-list-extended]`.

* Round-trip multi-value LIST option values

`.array(["a", "b"])` encoded as `(("a" "b"))`; the parser rejected both
that and `("a" "b")`: it took the `astring` branch first and could not
backtrack.

RFC 5258: `option-value = "(" option-val-comp ")"`, `option-val-comp =
astring / option-val-comp *(SP option-val-comp) / "(" option-val-comp
")"`. The parser reads a sequence of elements; the encoder writes
several values inside the `option-value` parentheses.

* Validate JMAPACCESS URLs

`.jmapAccess(URL(string: "jmap", relativeTo: base))` encoded as
`JMAPACCESS "jmap -- https://example.com/"` (`URL.description`); `http`
URLs were written though the parser rejects them.

The payload becomes `ResponsePayload.JMAPAccessURL`: `init?(_:)` resolves
against the base and requires `https`, a host, and no escaping inside
`quoted`; `URL.init` reads. The parser uses the same check.

* Make RESETKEY mechanisms require a mailbox

`.resetKey(mailbox: nil, mechanisms: [.internal])` encoded as bare
`RESETKEY`, silently dropping the mechanisms and resetting every key.

RFC 4467: `resetkey = "RESETKEY" [SP mailbox *(SP mechanism)]`. The
payload becomes `Command.ResetKeyScope`: `.allMailboxes` or
`.mailbox(_:mechanisms:)`.

* Hoist a #require the compiler flagged as redundant

* Cite RFC 7162 for SEARCH MODSEQ entry types

The docs described these as RFC 5464 GETMETADATA options, with
GETMETADATA examples. The grammar is RFC 7162 `search-modseq-ext`.

* Fix doc nits

- `ExtendedSearchScopeOptions.content` is a dictionary, not an array.
- `NamespaceDescription` init in the `NamespaceResponse` example needs
  `try` and `responseExtensions:`.
- `TaggedExtensionLabel` example: `SELECT INBOX (CONDSTORE)` parses as
  `.condStore`; use `RELEVANCY`.
- Stray `’` in a parser hint.
2026-11-04 17:41:53 +00:00
Daniel Eggert 9f61149d6d Validate IMAP URL components (#867)
* Validate percent-encoded IMAP URL components

`EncodedMailbox(mailbox: "x] x\r\n* BYE")` was written verbatim inside
the bare URL of `[REFERRAL …]`, injecting a response; likewise
`EncodedUser`, `EncodedAuthenticationType`, `EncodedSearch`,
`EncodedSection`.

RFC 5092: `enc-user`/`enc-auth-type = 1*achar`, `enc-mailbox`/
`enc-search`/`enc-section = 1*bchar`. Each type's public `var` and
labelled init become `init?(_:)`, literals trap, `String.init` reads.
One helper validates by character class and `pct-encoded` triplet, and
uppercases hex digits as the parser does, so values round-trip.

* Validate URLAUTH tokens

`EncodedAuthenticatedURL(data: "x\r\n* BYE")` was written verbatim at the
end of a URLAUTH URL, injecting a response.

RFC 4467: `enc-urlauth = 32*HEXDIG`. `init(data:)` and the public `var`
become `init?(_:)`; literals trap; `String.init` reads. The parser read
exactly 32 digits, so longer tokens didn't round-trip; it now reads all.

* Add IMAPServer.Host and keep IP-literal brackets

`IMAPServer(host: "h] x\r\n* BYE")` was written verbatim in the bare URL
of `[REFERRAL …]`, injecting a response. `imap://[::1]/` parsed as host
`::1` and re-encoded as `imap://::1/`. A negative port encoded as `:-1`.

RFC 3986: `host = IP-literal / IPv4address / reg-name`. New nested
`IMAPServer.Host` replaces the `String`: `init?(_:)` validates IPv6,
IPvFuture and reg-name (IPv4 is a subset), literals trap. The parser
keeps the brackets, accepts all of reg-name, and validates via `init?`;
its unreachable IPv4 branch is gone. `port` is now `UInt16?`.

* Don't parenthesize URLFETCH responses

`.urlFetch([.init(url: "url", data: nil)])` encoded as
`URLFETCH("url" NIL)`, which our parser rejects.

RFC 4467: `urlfetch-data = "URLFETCH" 1*(SP url-full SP nstring)`. Write
each pair after a space, unparenthesized. Fix fixtures that enshrined
the old output.

* Remove dead URL and date writers

`writeAuthIMAPURLRump`, `writeAuthIMAPURLFull`, `writeRelativeIMAPURL`,
`writeURLFetchType`, `writeInternetMessageDate` and
`writeListSelectBaseOptionQuoted` had no callers outside their own
encode tests, nor did the writers only they reached
(`writeAuthenticatedURL(NetworkMessagePath)`, `writeNetworkPath`,
`writeAbsoluteMessagePath`, `writeIUIDOnly`,
`writeURLMessageSectionOnly`, `writeMessagePathByteRangeOnly`). Delete
them and those tests; parse tests stay.

* Keep a section's trailing slash unless a partial follows

`imap://h/INBOX/;UID=1/;SECTION=1/` parsed with section `1`, so
`EncodedSection("1/")` didn't round-trip: the parser dropped the `/`
even when no `;PARTIAL=` followed it. Drop it only when one does.

* Reject `*` as a URL authentication type

`.type("*")` encoded as `;AUTH=*`, which parses as `.any`; `.type("*x")`
failed to parse, since `*` matched first.

RFC 5092: `iauth = ";AUTH=" ( "*" / enc-auth-type )`. `init?(_:)`
rejects `*`; the parser tries the longer `enc-auth-type` first.
2026-11-04 17:41:53 +00:00
Daniel Eggert 4606da62d2 Validate response text, command tags and namespace delimiters (#866)
* Reject invalid response text when writing

`.no(.init(text: "No mailbox x\r\n* BYE pwned"))` was written verbatim,
injecting an untagged response; likewise tagged, BYE, PREAUTH and
continuation responses.

RFC 9051: `text = 1*(TEXT-CHAR / UTF8-2 / UTF8-3 / UTF8-4)`. `writeText`
now asserts no NUL, CR or LF. Empty stays valid. The parser now accepts
UTF-8 text.

* Add Command.Tag for command and response tags

`TaggedCommand(tag: "A1 NOOP\r\nA2 DELETE INBOX\r\nA3", command: .noop)`
was written verbatim, injecting a command; likewise `APPEND` start.
`TaggedResponse(tag: "* BYE x\r\nA1", …)` injected a response.

RFC 3501: `tag = 1*<any ASTRING-CHAR except "+">`. New `Command.Tag`
replaces the `String` in `TaggedCommand`, `TaggedResponse`, `Response`,
`CommandStreamPart`, `AppendCommand`, `BadCommand` and NIOIMAP's state
machine errors; `init?(_:)` validates, literals trap. The parser already
enforces the production. IMAPCommands converts its numeric tags.

* Validate and escape namespace delimiters

`NamespaceDescription(string: "x", char: "\r\n", …)` wrote a raw CRLF
inside `"…"`, injecting a response. `"\""` encoded as `"""`, and `"\\"`
as `"\"` (unterminated); non-ASCII was accepted.

RFC 2342: `DQUOTE QUOTED-CHAR DQUOTE / nil`. `init(string:delimiter:
responseExtensions:)` (was `char:`) throws `InvalidDelimiterError` for
non-ASCII, CR, LF, NUL; the delimiter is stored as a byte, read-only.
Quoted-specials are escaped on write and parsed when escaped.

* Gate a capturing exit test on Swift 6.3
2026-11-04 17:41:52 +00:00
Daniel Eggert f77de8e279 Add ExtensionKey for tagged-extension parameters (#865)
Generic over the extended type, so a STORE key can't be used where a FETCH key is expected.
2026-11-04 17:41:52 +00:00
Daniel Eggert ded4ff7245 Validate atom-shaped types at construction (#864)
* Reject invalid URLAUTH mechanisms

`.resetKey(mailbox: .inbox, mechanisms: [.init("X\r\nA2 DELETE INBOX")])`
was written verbatim, injecting a command into RESETKEY / GENURLAUTH.

RFC 4467: `mechanism = "INTERNAL" / 1*(ALPHA / DIGIT / "-" / ".")`.
`init(_:)` becomes `init?(_:)`; literals trap on invalid input. The
parser already enforces the production.

* Reject invalid mailbox attributes

`MailboxInfo.Attribute("\\X)\r\n* BYE")` was written verbatim, injecting
an untagged response into LIST. It also feeds CREATE … USE via
`UseAttribute(_:)`.

RFC 3501: `mbx-list-flag` is `"\" atom` in every branch. `init(_:)`
becomes `init?(_:)`; literals trap on invalid input. Constants become
`static let` literals; storage is `let`. Fix the `String.init` doc,
which claimed lowercase storage.

* Reject invalid use attributes

`UseAttribute("\\All))\r\nA2 DELETE INBOX\r\nA3 NOOP ((")` was written
verbatim, injecting a command into CREATE … (USE (…)). It also feeds
`GmailLabel(useAttribute:)`.

RFC 6154: `use-attr-ext = "\" atom`. `init(_:)` becomes `init?(_:)`;
literals trap on invalid input. `init(_: MailboxInfo.Attribute)` stays
non-failable: same production. Fix docs that claimed lowercase storage.

* Reject invalid attribute flags

`AttributeFlag("x\" all 1\r\nA2 DELETE INBOX\r\nA3 NOOP \"")` was
written verbatim inside SEARCH MODSEQ "/flags/…", injecting a command.

RFC 7162: `attr-flag = "\\" atom / atom`, where `"\\"` is two literal
backslashes (the stored form is already escaped). `init(_:)` becomes
`init?(_:)`, validating after lowercasing; literals trap on invalid
input. Fix docs that claimed casing is preserved.

* Add FilterName for SEARCH FILTER

`.filter("x\r\nA2 DELETE INBOX")` was written verbatim, injecting a
command into SEARCH.

RFC 5466: `filter-name = 1*<any ATOM-CHAR except "/">`. New `FilterName`
type replaces the `String` payload of `SearchKey.filter`; `init?(_:)`
validates, literals trap on invalid input.

* Reject invalid compression kinds

`.compress(.init("DEFLATE\r\nA2 DELETE INBOX"))` was written verbatim,
injecting a command after COMPRESS.

RFC 4978: `algorithm = "DEFLATE"`, an atom. `init(_:)` becomes
`init?(_:)`, validating after uppercasing; literals trap on invalid
input. Add `String.init(_:)`.

* Reject invalid capabilities

`Capability("X\r\nA2 DELETE INBOX")` was written verbatim, injecting a
command into ENABLE, or (`"X\r\n* OK [ALERT] hi"`) a response into
CAPABILITY / ENABLED. Same for the kinds a capability is built from,
e.g. `.context(.init("X\r\n…"))`.

RFC 3501: `capability = ("AUTH=" auth-type) / atom`. `init(_:)` becomes
`init?(_:)` for `Capability` and each kind (kinds validate after
uppercasing); literals trap on invalid input. Storage is `let`. Add
`String.init(_:)` for the kinds.

* Add validated LIST option tag types

`.standard("X)\r\nA2 DELETE INBOX\r\nA3 LIST (")` was written verbatim,
injecting a command into LIST select/return options. Same for
`.vendor(KeyValue<String, String>)`.

RFC 5258: `option-standard-tag = atom`, `option-vendor-tag =
vendor-token "-" atom`. New `OptionExtensionKind.Standard` and
`.Vendor` replace the payloads; `init?(_:)` validates, literals trap.
`Vendor` stores `ACME-SORT` and derives `token` / `name`.

* Add CharacterSetName for SEARCH, SORT, and BADCHARSET

`.sort(…, charset: "UTF-8 ALL\r\nA2 DELETE INBOX\r\nA3 SEARCH", …)` was
written verbatim, injecting a command into SORT / SEARCH / ESEARCH.
`.badCharset(["a)] x\r\n* BYE"])` injected a response. Quoted names
also didn't round-trip: `"a b"` parsed, then encoded bare.

RFC 3501: `charset = atom / quoted`. New `CharacterSetName` replaces
`String`; valid = non-empty `TEXT-CHAR`. Written as an atom where it is
one (`CHARSET UTF-8` unchanged), quoted otherwise, never a literal. The
parser rejects quoted names it couldn't write back.

* Assert keyword validity when writing a flag keyword

`writeFlagKeyword` wrote verbatim with no check. `writeAtom` would be
wrong: Gmail keywords contain `]`, which no atom does. Assert the
keyword's own production instead; test that a `]` keyword encodes.

* Hoist a #require the compiler flagged as redundant
2026-11-04 17:41:52 +00:00
Daniel Eggert e7729053c0 Fix writers that emit untrusted bytes unquoted (#863)
* Test basic list-mailbox parsing

`parseListMailbox` was only tested against DEL. Cover wildcards,
resp-specials, the quoted-string branch, and an empty input.

* Write GmailLabel bare only if it is a system label

`GmailLabel("\\x)\r\nA2 DELETE INBOX\r\nA3 NOOP (")` was written raw
because it starts with `\`, injecting a command into STORE X-GM-LABELS.
Reachable from server data via `init(mailboxName:)`.

Write bare only for `"\" atom`; otherwise `writeIMAPString`. Also fixes
round-trip of astring labels such as `"\\Foo bar"`.

* Write URLFETCH URLs as astring

`.urlFetch([ByteBuffer(string: "x\r\nA2 DELETE INBOX")])` was written
verbatim, injecting a command.

RFC 4467: `url-full = astring`. Write bare if `1*ASTRING-CHAR` (keeps
long URLAUTH URLs out of literals), else `writeIMAPString` (`%`, `*`).

* Write body encoding as string

`Encoding("x\" 0 0) UID 9 (")` was spliced into `"…"` unescaped, so a
MIME `Content-Transfer-Encoding` header spoofs FETCH items. The parser
accepts any `string` (incl. literals with CRLF), so proxies re-emitted it.

`body-fld-enc = (DQUOTE ("7BIT"/…) DQUOTE) / string`. Use
`writeIMAPString`; standard values are unchanged on the wire.

* Write media type and subtype as string

`Media.Subtype("x\"\r\n* 1 EXPUNGE")` was spliced into `"…"` unescaped;
a MIME `Content-Type` header could inject an untagged response.

`media-basic = ((DQUOTE (…) DQUOTE) / string) SP media-subtype`,
`media-subtype = string`. Use `writeIMAPString` (still uppercased).

* Write search correlator tag as string

`SearchCorrelator(tag: "a\") * BYE")` was spliced into `"…"` unescaped,
so ESEARCH / UIDBATCHES responses could carry injected data.

RFC 4466: `tag-string = string`. Use `writeIMAPString`, and parse a
`string` rather than `"` tag `"` so every tag round-trips.

* Fix escaping of mailbox path separator

The quoted-special cases were swapped: `\` encoded as `"\"`
(unterminated), `"` as `"\\"` (a backslash). `pathSeparator: "\n"` was
written as a raw LF inside `"…"`, and NUL silently became NIL.

`(DQUOTE QUOTED-CHAR DQUOTE / nil)`, `QUOTED-CHAR = <TEXT-CHAR except
quoted-specials> / "\" quoted-specials`. Swap the cases; parse the
escaped forms; `MailboxPath.init` throws for CR, LF, CRLF, NUL.

* Base64-encode URLMECH mechanism data on write

The parser base64-decodes `mechanism-base64`, but the encoder wrote the
decoded bytes raw: parse `INTERNAL=YQ==` → "a" → `INTERNAL=a`. A proxy
re-encoding `[URLMECH INTERNAL X=<base64 of CRLF…>]` emitted raw CRLF.

RFC 4467: `mechanism-base64 = mechanism ["=" base64]`. Encode on write;
fix the fixtures that locked in the mismatch.
2026-11-04 17:41:51 +00:00
Daniel Eggert df4f53ea8b Run macOS Swift 6.2/6.3 tests serially (#881) 2026-11-04 16:38:08 +01:00
Daniel Eggert d7018634d2 Explicit “invalid” case for invalid mailbox IDs (#876) 2026-11-02 21:53:00 +01:00
Daniel Eggert 1d32ed43bf Disable Swift Testing error backtraces on Swift 6.3 CI (#879) 2026-11-02 09:59:28 +01:00
Matt Diephouse f5fa7d3691 Implement OBJECTID+ extension (#840) 2026-10-30 13:47:17 +00:00
Daniel Eggert fbec780de3 Normalize malformed ENVELOPE address groups (#872) 2026-10-27 15:06:24 +01:00
Matt Diephouse 17929ed44f Handle invalid mailbox IDs (#873) 2026-10-16 09:07:48 -04:00
AnthonyandDaniel Eggert d456d32908 Accept the responses a server may send during IDLE (#855)
Co-authored-by: Daniel Eggert <deggert@apple.com>
2026-10-13 17:08:26 +02:00
Matt Diephouse b9e08e13cf Fix some issues with RFC 2087 (QUOTA extension) (#871)
* Separate resources in QUOTA responses with spaces

RFC 2087 defines quota_list as "(" #quota_resource ")", with a single
space as the delimiter. writeQuotaResources concatenated resources with
no separator, producing e.g. (STORAGE 10 512MESSAGE 1 100).

* Support zero or more quota roots in QUOTAROOT responses

RFC 2087 defines quotaroot_response as "QUOTAROOT" SP astring
*(SP astring): a mailbox has zero or more quota roots. The parser
required exactly one, so `* QUOTAROOT comp.mail.mime` (the RFC's own
example) failed to parse and any roots after the first were dropped.

ResponsePayload.quotaRoot now carries [QuotaRoot]. This is a source-
breaking change to the public API.
2026-10-13 10:51:38 -04:00
Daniel Eggert 088b6eb105 Fix race in ClientBootstrap (#870) 2026-10-12 15:32:49 +02:00
finagolfin 8283076ac2 Disable exit tests on Android too (#861) 2026-10-11 18:55:15 +02:00
apple-ospo[bot] 30b48985ac Housekeeping: replace repo-level CODE_OF_CONDUCT (#857)
Co-authored-by: apple-ospo[bot] <326325757+apple-ospo[bot]@users.noreply.github.com>
2026-10-10 19:27:55 +02:00
Ross Cooperman 544677dbaa Add Gmail X-GM-MSGID, X-GM-THRID, and X-GM-RAW search keys (#853) 2026-10-10 18:48:53 +02:00
Daniel Eggert 14ad3887da More “Modified UTF-7” tests (#856) 2026-09-25 11:23:20 -04:00
Daniel Eggert 6c8cfaddcc Add a imperative IMAP client and an imap-tool CLI (#841) 2026-09-25 16:16:41 +01:00
Daniel Eggert 8fd34eaff7 Validate flag and quota resource names as atoms (#854) 2026-09-14 21:16:50 +00:00
Christian 1c7076fab6 Fix framing of zero-length literals (#851) 2026-09-14 14:13:19 -07:00
Daniel Eggert 3507d6da24 Don't build macOS for Swift 6.1 (#847) 2026-09-14 13:43:48 -07:00
Eirik Botten Nicolaysen 2e4eee1f31 Treat DEL (0x7F) as an atom-special (#848) 2026-09-13 10:40:42 -07:00
Rafael Cepeda 6d8e7bd6e5 Delete SECURITY.md (#852)
Deleting file in favour of org-wide https://github.com/apple/.github/blob/main/SECURITY.md.
2026-09-13 10:21:25 -07:00
K Gopal Krishnaandkaygdotorg 26a69caf84 Allow configuring IMAP client line buffer size (#849)
Co-authored-by: kaygdotorg <106914721+kaygdotorg@users.noreply.github.com>
2026-09-09 08:51:12 +02:00
Daniel Eggert e536be599e Reject invalid AuthenticationMechanism (#846) 2026-09-01 10:11:11 +02:00
Daniel Eggert c4d4d23e29 Drop pre-Swift 6.2 (#845) 2026-08-29 09:23:34 +02:00
Daniel Eggert 273b5602db Update UIDBATCHES RFC number (#842) 2026-08-10 14:46:51 -04:00
Matt Diephouse 4b83b70e04 Enable MemberImportVisibility (#844)
https://docs.swift.org/compiler/documentation/diagnostics/member-import-visibility/
2026-07-30 11:03:01 -04:00
Matt Diephouse 009e18e60f Make tests compile for iOS/tvOS/watchOS/visionOS (#843) 2026-07-30 16:18:16 +02:00
Daniel Eggert 90124af395 Add proxy readme (#839) 2026-07-30 09:57:47 -04:00
Oliver Drobnik d67f1243f7 Replace C ctype calls with pure-Swift ASCII checks in UInt8+ParseTypeMembership (#838) 0.4.0 2026-07-20 13:28:36 +02:00
Daniel Eggert dcba159054 Modernize the Proxy example (#835) 2026-07-18 21:54:14 +02:00
Daniel Eggert e67fc89e63 RFC 9979 (#833) 2026-07-18 20:57:21 +02:00
Daniel Eggert 1057994e93 Move examples into sub-dir (#834) 2026-07-18 12:00:06 +02:00
Daniel Eggert 880366a417 Add macOS CI job (#836) 0.3.0 2026-07-07 11:37:57 +02:00
Léon Simmons 79c4237b5c docs: fix typo in AuthenticationStateMachine comment (#832)
authentiction → authentication
2026-07-03 16:33:52 +02:00
Oliver DrobnikandClaude Opus 4.8 0da5049bc9 NIOIMAPCore: import the Android overlay module instead of Bionic (#829)
* NIOIMAPCore: import the Android overlay module, not Bionic

#826 added a `canImport(Bionic)` / `import Bionic` arm to NIOIMAPCore's
per-file libc-import guards. Switch it to the `Android` umbrella module,
which re-exports Bionic plus the Swift overlay, so the import is the more
complete/idiomatic libc surface on the Swift Android SDK. No functional
change — both resolve the libc symbols (isalnum/isalpha) used here.

* CI: enable InlineAlways experimental feature for Android nightly build

swift-collections 1.6.0 guards `@inline(always)` with `#if compiler(>=6.3)`,
but that attribute is still behind the experimental `InlineAlways` feature in
the nightly `main` toolchain the Android Swift SDK job builds with, so the build
fails with "'@inline(always)' is an experimental feature". Pass
`-enable-experimental-feature InlineAlways` to swift build until the upstream
guard is fixed (cf. apple/swift-collections #580/#584).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-25 09:04:30 +01:00
Daniel Eggert 0eaa07f877 Fix cr-lf handling in FramingParser (#830) 2026-06-22 08:10:24 -04:00
Si Beaumont b63f163a4e Add CI for building with Android SDK (#831) 2026-06-12 16:26:16 +02:00
Daniel Eggert faaa15ee50 Add explicit test for credential redaction (#825) 2026-06-03 10:46:06 -04:00
Oliver DrobnikandClaude Opus 4.8 bcf875610c Add Android to NIOIMAPCore's libc imports (#826)
### Motivation:

NIOIMAPCore's parser guards the libc import per file with
Darwin/Glibc/Musl/Windows and otherwise falls into
`#else let badOS = { fatalError("unsupported OS") }()`. On the Swift
Android SDK none of those arms match, so every guarded file takes the
`#else` branch and the module fails to build with "invalid redeclaration
of 'badOS'" (the symbol is declared at file scope in ~17 sources).

### Modifications:

Add an `#elseif canImport(Android)` arm importing `Android` to the per-file
libc-import guards in NIOIMAPCore. For `UInt8+ParseTypeMembership.swift`,
which imports specific ctype functions, add `import func Android.isalnum` /
`import func Android.isalpha`.

### Result:

NIOIMAPCore compiles for Android. No change on any other platform.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-02 16:06:29 +00:00
Daniel Eggert 983e0f1aff InvalidCommandForState error may leak PII (#828) 2026-06-02 16:37:19 +01:00
Daniel Eggert abfc170ff1 Only parse continuation responses while an AUTHENTICATE is in progress (#827) 2026-06-01 14:13:22 -04:00
Daniel Eggert 01de1f9a29 Allow () in place of NIL for empty addr-list fields. Fixes #821 (#823) 2026-04-30 08:12:53 +02:00
Daniel Eggert 34883bcba6 Update style in documentation (#822) 2026-04-28 18:06:43 +02:00