Document and improve abstract reader/writer interface #208

Dependabot can provide automatic pull requests for things in the repository that should be updated. Example PR from dependabot against a repo owned by the github organization: github/opensource.guide#2248 Documentation: https://docs.github.com/en/code-security/dependabot/dependabot-security-updates/about-dependabot-security-updates

facilitate replacement of malloc/free functions open_memstream functions are left out of this change as they require other functions from stdlib.

We don't support this any more. ``` ../../src/cbor.h:255:69: error: '_Bool' is a C99 extension [-Werror,-Wc99-extensions] CBOR_INLINE_API CborError cbor_encode_boolean(CborEncoder *encoder, bool value) ^ ``` Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>

Homebrew no longer carries precompiled versions of Qt for macos-11, so switch to macos-13, the last on x86 CPUs (so we can still run Valgrind). It's also going away after the end of June. Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>

``` ==21147== Valgrind: debuginfo reader: ensure_valid failed: ==21147== Valgrind: during call to ML_(img_get) ==21147== Valgrind: request for range [18446744069408125024, +16) exceeds ==21147== Valgrind: valid image size of 140733057859584 for image: ==21147== Valgrind: "/usr/local/Cellar/icu4c/74.2/lib/libicudata.74.2.dylib" ``` Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>

I don't think we need to worry about Qt 4 any more. Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>

Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>

Intel says this is needed Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>

`cbor_parser_init` and `cbor_parser_init_reader` are substantially similar, however the latter misses clearing out `it->flags`, leaving it uninitialised so possibly unsafe. Rather than copying & pasting that from `cbor_parser_init`, lets just use one routine that does the "common" part, then each routine can focus on the specifics needed.

Describe the input parameters for the function and how they are used as best we understand from on-paper analysis of the C code.

The `token` parameter is not sufficient since it is effectively shared by all `CborValue` instances. Since `tinycbor` often uses a temporary `CborValue` context to perform some operation, we need to store our context inside that `CborValue` so that we don't pollute the global state of the reader.

In its place, put an arbitrary `void *` pointer for reader context. The reader needs to store some context information which is specific to the `CborParser` instance it is serving. Right now, `CborValue::source::token` serves this purpose, but the problem is that we also need a per-`CborValue` context and have nowhere to put it. Better to spend an extra pointer (4 bytes on 32-bit platforms) in the `CborParser` (which there'll be just one of), then to do it in the `CborValue` (which there may be several of) or to use a `CborReader` object that itself carries two pointers (`ops` and the context, thus we'd need an extra 3 pointers).

We simplify this reader in two ways: 1. we remove the `consumed` member of `struct Input`, and instead use the `CborValue`'s `source.token` member, which we treat as an unsigned integer offset into our `QByteArray`. 2. we replace the reader-specific `struct Input` with the `QByteArray` it was wrapping, since that's the only thing now contained in our `struct Input`. If a `CborValue` gets cloned, the pointer referred to by `source.token` similarly gets cloned, thus when we advance the pointer on the clone, it leaves the original alone, so computing the length of unknown-length entities in the CBOR document can be done safely.

What is not known, is what the significance is of `CborEncoderAppendType`. It basically tells the writer the nature of the data being written, but the default implementation ignores this and just blindly appends it no matter what. That raises the question of why it's important enough that the writer function needs to know about it.

This reads a CBOR file piece-wise, seeking backward and forward through the file if needed. Some seeking can be avoided by tuning the block size used in reads so that the read window shifts by smaller amounts.

Not 100% sure of the syntax for documenting struct-members outside of the struct as I'm too used to doing it inline, hopefully this works as expected. :-)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Document and improve abstract reader/writer interface #208

Document and improve abstract reader/writer interface #208

Commits on Jun 21, 2024

Document and improve abstract reader/writer interface #208

Are you sure you want to change the base?

Document and improve abstract reader/writer interface #208

Commits on Jun 21, 2024