Nicholas Fraser

Yes, unfortunately the projects have similar names. If you look at the first commit of each project you'll find that this project came first and "libmpack" project came second, but...

The stream parser needs to store all of the data coming from the stream. The whole tree must be kept in memory in order to parse a complete message and...

Fixed, sorry about this. It's an unfortunate off-by-one bug which is triggered on the rare occasion that a build ends up exactly at the end of a 4096 page boundary....

Your example doesn't trigger the bug in the unit test suite because the unit tests use small page sizes to test various wrapping behaviors. I need to change the unit...

Thanks for the bug report. It should be fixed now. The writer will clean up properly when destroyed with an open builder. I don't have great tests for it at...

Hmm. You're partially right. An error can't happen from I/O; it can only happen from a malloc failure. When a builder is open all writes are diverted to a buffer...

Hmm, that's really strange. I don't see any bugs with continuous mode or any unusual interactions with gzip/zcat, as long as there is data in the file: ``` $ echo...

Hmm. The relevant line is here: https://github.com/ludocode/mpack/blob/df17e83f0fa8571b9cd0d8ccf38144fa90e244d1/src/mpack/mpack-common.c#L322 This is hitting an assertion, which I guess means you're compiling in debug mode. This does look like a bug in the code,...

Indeed! I am currently adding timestamp support to mpack in the [timestamp branch](https://github.com/ludocode/mpack/tree/timestamp). Once I get a decent battery of unit tests I'll start working on timestamp support here. The...

Upon further testing it appears `__attribute__((__noreturn__))` does not throw the warning under NVidia/PGI if it comes before `static`. So maybe the solution is to simply use that instead of `_Noreturn`...