blob server protocol and spec

[Nick Gerakines]

Hi everyone,

I started a c++ reference implementation of some of the camlistore services and have some questions on the protocol and expected features.

The project code is all on GitHub, licensed under the MIT:

    https://github.com/ngerakines/camlispp


I'm comfortable with both perl, go and python, but am having a hard time find a solid feature list for the blob service. Other than the blob related docs folder within the main repo, is there anything that describes the expected features of a blob storage daemon? I ask this with the end-goal of being able to have the go/python impl connect with and sync to the c++ one without any issues or hiccups.

# Nick Gerakines

[Brad Fitzpatrick]

There's also the doc/protocol directory. It should be accurate.

You can test that it all works with the bs-test.pl script (maybe out of date?) or more accurately: with camget, camput, and "camtool sync" to sync between blobservers.

Any particular motivation for a C++ server? I see few advantages over the Go version.

# Nick Gerakines

--
You received this message because you are subscribed to the Google Groups "Camlistore" group.
To unsubscribe from this group and stop receiving emails from it, send an email to camlistore+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Nick Gerakines]

Good to know, thanks.

I like Go and I'm pretty comfortable with it, but at work we primarily use c++ and java (generally speaking).

[Brad Fitzpatrick]

Yeah, but I use tools all the time written in languages my employer doesn't typically use, but I don't typically rewrite them.  Hence my curiosity about your motivation.  Just for fun, or ... ?

[Nick Gerakines]

This is a research project for me to prototype a CAS service spec and see how it can be applied to what I do at work. Battle.net deals with gobs and gobs of data of different types and from many difference sources and we are always looking for better ways to improve how we work with that data.

[Brad Fitzpatrick]

Gotcha. The blob server should be the easiest part. Do you intend to reimplement all the indexing and search stuff too?

[Nick Gerakines]

Yeah, that is the plan. I'm also thinking about doing some sort of view transform to retrieve blobs with different compression types or create torrent definitions on-the-fly.