der: (re-)add Reader::read_value with EOC support#1895
Merged
Conversation
This reverts commit 5277fcf.
tarcieri
force-pushed
the
der/restore-read-value
branch
2 times, most recently
from
759d6f1 to
f2f6da0
Compare
One of the blockers for addressing indefinite length handling (#779) has been where to consume the EOC tag. This commit (re)introduces `Reader::read_value` (added in #1877, removed in #1887) and handles decoding the EOC there. This gives us a single place where EOC can be handled for all constructed messages. With these changes, the decoder is able to parse `cms_ber.bin` from `cms/tests`, from which the `cms_der.bin` file has been translated. This file provides a real-world example of nested indefinite lengths from CMS. Note, however, that the example contains a constructed `Any` which isn't yet being correctly handled. A `TODO` for ensuring the BER and DER decode identically has been added.
tarcieri
force-pushed
the
der/restore-read-value
branch
from
f2f6da0 to
586e4c4
Compare
Contributor
|
Note: because read_nested is not needed for primitive types, I suggest to introduce As stated in comment #1877 Therefore call chain would be: |
Member
Author
|
@dishmaker I would love to find somewhere better to put this logic, but it ends up being quite difficult given the current usage |
Merged
tarcieri
added a commit
that referenced
this pull request
Feb 13, 2026
## Added - Custom error types support to the `Decode` and `DecodeValue` traits (#1055) - `EncodingRules` enum (#1321) - `Decode::from_ber` (#1389) - Documentation for field-level `tag_mode` attribute (#1401) - `Hash` implementation for `AlgorithmIdentifier` (#1414) - `SequenceRef::as_bytes` and `AsRef<[u8]>` impl (#1454) - `Reader::peek_into` (#1478) - `GeneralString` variant to `Tag` (#1512) - Conversions from `OctetString(Ref)` to `Vec`/`Bytes` (#1540) - Custom error types in derive macros (#1560) - Support for tags beyond 30 (#1651) - Const `::new` to `Length`, `BytesRef`, and `AnyRef` (#1713) - Const `GeneralizedTime::from_date_time` (#1718) - `Decode::from_der_partial` (#1725) - Conversions between `BitStringRef`/`OctetStringRef` and `[u8; N]` (#1731) - Add class bits consts for Application and Private tag support (#1721) - conversions between `heapless:Vec<u8>` and `OctetStringRef` (#1735) - `IsConstructed` trait, impl'ed on any `FixedTag` (#1744) - Implement `Hash` for `SetOf` (#1764) - Implement `Uint`/`Int` conversions from native types (#1762) - Support for `APPLICATION`, `CONTEXT-SPECIFIC` and `PRIVATE` tags (#1819, #1825, #1944) - Support `Cow<[u8]>` in `derive(Sequence)` (#1850) - `diagnostic::on_unimplemented` attributes (#1876) - `Reader::read_value`, auto-nest `DecodeValue` (#1877, #1895, #1897, #1901) - Indefinite length support for BER (#1884, #1885, #1894, #1900, #1902, #1910) - Constructed OctetString support (#1899, #1922) - string conversions, predicate methods for EncodingRules (#1903, #1953) - `Any::header` (#1935) - `Tag::RelativeOid` (#1942) - `ber` feature (#1948, #1950) - Hash derive for `StringOwned` and `Ia5String` (#1973) - Implement `DecodeValue/EncodeValue/Tagged` for `Cow` (#2093) - Add doc examples for `EncodeValue`, `Encode`, `DecodeValue`, `Decode`, `Error`, `ErrorKind`, `Tagged`, `FixedTag`, `IsConstructed`, `Tag`, `Length`, `Header` (#2075, #2071, #2070, #2064, #2058, #2052, #2053, #2051) - `SequenceRef::new` (#2224) ## Changed - Return `Tag::Integer.length_error()` on empty ints (#1400) - Allow all blanket impls on `?Sized` types (#1451) - Refactor `Tag::peek` (#1479) - Refactor `Header::peek` (#1480) - Use `core::error::Error` (#1553) - Use 2024 edition, bump MSRV to 1.85 (#1670) - Bump `const-oid` to v0.10 (#1676) - Reject zero lengths reads (#1716) - Deprecate `TagNumber::new` (#1727) - Use strict context-specific skipping condition (equal tag numbers only) (#1740) - Const `Any::to_ref`, `BytesOwned::to_ref` (#1797) - Return `ErrorKind::Noncanonical` in `EXPLICIT` when primitive (#1818) - Use `read_nested` to check length of `IMPLICIT` types (#1739) - Simplify `From<&UintRef<'a>>` for `Uint` (#1840) - Make `ObjectIdentifier<MAX_SIZE>` impls generic (#1851) - Extract `reader::position::Position` (#1880) - Make `Reader` cloneable (#1883) - Simplify `Header::peek` and `Tag::peek` (#1886) - Improve constructed bit handling (#1919) - Use fat pointer in `OctetStringRef`, consolidate bytes/string modules, improve internal ref types (#1920, #1921, #1998, #2040) - Change constructor `Header::new`, add `Header::with_length`, tests for constructed octet string (#1931, #1930) - Simplify `der_cmp` for `Length` (#1997) - Rename variable encoder -> writer and improve example (#2078) - Have `PemReader` decode to `Cow::Owned` (#2094) - Only sort `SET OF` types with `EncodingRules::Der` (#2219) - `SetOf` and `SequenceOf` now require `heapless` feature (#2220, #2229) ## Fixed - Append in `Encode::encode_to_vec` (#1760) - Derive optional OCTET/BIT STRING on `Option<&[u8]>` (#1737) - X.680 tag order: compare class and number first (#1790) - BMPString compatibility in derive macros (#1793) - `Tag::peek_optional` for 6-byte and 7+ byte tags (#1804) - `Header::peek` for 11-byte tag-lengths (#1828) - Panic in `value_cmp`: add `Iterator::size_hint` (#1830) - Error position tracking improvements (#1889, #2080, #2079) - Bound `Decode(Value)::Error` on `core::error::Error` (#2137) - Have `SetOf(Vec)::insert` check for duplicates (#2217) ## Removed - `TagNumber::N0..N30` consts (#1724) - 256MiB limit on `Length` (#1726) - Generic `<T>` from `Reader::finish` (#1833) - Deprecated `SetOf(Vec)::add` methods (#2232) - `SequenceRef` lifetime (#2224)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
One of the blockers for addressing indefinite length handling (#779) has been where to consume the EOC tag.
This commit (re)introduces
Reader::read_value(added in #1877, removed in #1887) and handles decoding the EOC there. This gives us a single place where EOC can be handled for all constructed messages.With these changes, the decoder is able to parse
cms_ber.binfromcms/tests, from which thecms_der.binfile has been translated. This file provides a real-world example of nested indefinite lengths from CMS.Note, however, that the example contains a constructed
Anywhich isn't yet being correctly handled. ATODOfor ensuring the BER and DER decode identically has been added.cc @dishmaker