Preserve raw bytes in IMAP APPEND

[mpscholten]

Fixes #102.

appendFull currently splits the rendered message with BS.lines, computes a new literal size from those split lines, and sends each line with bsPutCrLf. That mutates already-rendered RFC 5322/MIME bytes: CRLF input becomes CRCRLF on the wire and the literal size no longer matches the original message size.

This PR keeps the public API unchanged and sends the APPEND literal exactly as supplied:

• literal size is BS.length mailData
• mailData is written once with bsPut
• the final IMAP literal terminator CRLF is still sent afterwards

Regression coverage:

• IMAP.append conn "INBOX" "Subject: x\r\n\r\nBody\r\n" sends the exact raw message bytes after the APPEND command.

Before the fix, the new regression failed with {23} and doubled carriage returns. After the fix it passes with {20} and unchanged message bytes.

Validation:

• nix shell nixpkgs#cabal-install nixpkgs#ghc --command cabal test imap-parsers --test-show-details=direct
• nix shell nixpkgs#cabal-install nixpkgs#ghc --command cabal build all
• nix shell nixpkgs#cabal-install nixpkgs#ghc --command cabal check (only the existing CHANGELOG doc-place warning remains)
• git diff --check

Note: this includes the same imap-parsers test-target setup and https:// repository metadata fix as the other small upstream PRs because the branch is independently based on upstream master. It can be rebased to drop that overlap after the first test-target PR merges.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 62fdff5430

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Return a ByteString on scripted read failures

When the new imap-parsers test target is built, these failure branches are type-checked as part of popLine :: IO ByteString/popBytes :: IO ByteString, but assertFailure from HUnit returns Assertion (IO ()). That makes the newly added test suite fail to compile before any tests can run; use a polymorphic throwing helper or follow the failure with an unreachable ByteString result.

Useful? React with 👍 / 👎.