From d6ff8b4f8585cb765ef433e0af9f74d30418525e Mon Sep 17 00:00:00 2001 From: Joey Hess Date: Sun, 3 Aug 2014 20:32:26 -0400 Subject: devblog --- doc/devblog/day_210__conversion_and_digression.mdwn | 14 ++++++++++++++ 1 file changed, 14 insertions(+) create mode 100644 doc/devblog/day_210__conversion_and_digression.mdwn (limited to 'doc/devblog') diff --git a/doc/devblog/day_210__conversion_and_digression.mdwn b/doc/devblog/day_210__conversion_and_digression.mdwn new file mode 100644 index 000000000..2629eb917 --- /dev/null +++ b/doc/devblog/day_210__conversion_and_digression.mdwn @@ -0,0 +1,14 @@ +Just finished converting both rsync and gcrypt to the new API, +and testing them. Still need to fix 2 test suite failures for gcrypt. +Otherwise, only WebDAV remains unconverted. + +Earlier today, I investigated switching from hS3 to +. Learned its API, which seemed a +lot easier to comprehend than the other two times I looked at it. Wrote +some test programs, which are in the `s3-aws` branch. I was able to stream +in large files to S3, without ever buffering them in memory (which hS3's +API precludes). And for chunking, it can reuse an http connection. +This seems very promising. (Also, it might eventually get Glacier support..) + +I have uploaded haskell-aws to Debian, and once it gets into testing and +backports, I plan to switch git-annex over to it. -- cgit v1.2.3