Sereal Archives

Sereal v2 Finally Released as Stable

Just now, I've uploaded three distributions to PAUSE: Sereal-Encoder-2.01, Sereal-Decoder-2.01, and Sereal-2.01.

This means that version 2 of the Sereal serialization protocol is finally considered public, stable, and ready for production use. There was a long article about Sereal v2 on the blog a few months ago. Since then, at least one major new feature was added to Sereal version 2: Object serialization hooks (FREEZE/THAW). Let me summarize the two major user-visible new features.

  • Encoding data in document headers. This feature is useful primarily for large and/or compressed Sereal documents: It allows you to embed meta data into the document header which can be inspected without having to deserialize the entire document. Think routing information in a high-traffic server application.

  • FREEZE/THAW hooks. This much-requested feature led to lots of discussions between the other Sereal contributors and myself and quite some disagreement. The end result is that if you enable an option on the Sereal encoder object, the encoder will check for the existence of a FREEZE method on any object it encounters and then call it with the string "Sereal" as its first argument. Whatever the FREEZE method returns (as its first return value) will be serialized in place of the object. On the decoding side, any object that underwent FREEZE will require a THAW method of the object's class. It's the obvious reverse of FREEZE and whatever THAW returns will be included in the output of the decoder run. This feature slows down encoder and decoder due to method lookups and method calls. But it's optional. The whole thing is designed to only cost you anything if you actually use it. The best part about FREEZE/THAW, however, is that I didn't have to come up with the interface. Marc Lehmann wrote Types::Serializer including a specification of how these FREEZE/THAW hooks can work for multiple serialization libraries. (Basically passing in the name of the serializer for the class to dispatch on.) Thus, you can now often write a single hook that supports all of Sereal, JSON(::XS), and CBOR(::XS).

All in all, Sereal v2 was long in the making. The Perl/XS implementation of the decoder is fully backward-compatible and the encoder can optionally be made to emit v1 documents. We took great care in designing the changes and that simply took some time to percolate.

The Perl/XS implementation is live. I'd be surprised if we didn't see full, production grade v2 implementations from Borislav (Ruby), Damian (Go), and Andrea (Objective C) released within a few days.

Announcement for Sereal, a binary data serialization format, finally live!

It's been long in the making, but finally, I've gotten the Sereal announcement article in a shape that I felt somewhat comfortable with publishing. Designing and implementing Sereal was a true team effort and we really hope to see non-Perl implementations of it in the future. We're virtually committed to finish the Java decoder at least for our data-warehousing infrastructure. Any help and cooperation is welcome, as are patches to improve the actual text of the specification (which is kind of a weak point still).

By the way, for those who worried about the lack of a comment-system on the dev blog before, we've added Disqus-support.

But now, I'm just glad it's out there!

About Steffen Mueller

user-pic Physicist turned software developer. Working on Infrastructure Development at Apparently the only person who thinks the perl internals aren't nearly as bad as people make them. See also: the tech blog, CPAN.