OHAI,
We had planned to slow down a bit to prepare for release: no more feature, no more invasive changes until we tag something out of which we can publish a release. I then discussed with Bret Lambert (blambert@) and he asked me if we used some kind of code scanner to detect defects in our code.
OHAI,
Lately we have stabilized and decided that we’re okay with the features we have for our first release, this time for good, so we focused on making sure that we didn’t introduce regressions and that we were rock solid.
<p>Hey,</p> I wanted to publish something last Friday but I unfortunately got my hands on a copy of Super Nintendo's Zelda which led to an instant loss of motivation.
<p>OHAI,</p> This week has been crazy, my brain melted a little as I worked on filters, it melted a little as I worked on LDAP and it melted a little more as I tried to trick Eric into working on filters with me (yes, that actually worked :-).
OHAI,
This week I had intended to work on filters. After 3 days of pain and swearing, even though it did move forward, I finally decided to step back for a few days and work on something else to preserve my sanity.
OHAI,
First of all, on a completely unrelated note, I’d like to emphasize that this blog post is being written from my text editor launched by a little shell and pushed to poolp using the Ocean API. I can be web2.0 from console and emacs, and that’s definitely worth a mention ;-)
Almost a month since my last post… I know, I know.
So I left my position at Scality two weeks ago. I had been there for a year and a half working on a very interesting and tricky project. I will miss my colleagues and the discussions we had on all kinds of topic, I hope them the best future and that we will be able to share beers again ;-)
The crypto backend that has been committed a couple days ago started a little discussion and it was decided to switch the default cipher from Blowfish to AES-128 and make it a default.
A few days ago, Charles committed the compress_backend API which allowed transparent deflation/inflation of envelopes and messages as they hit the queue.
The compression code is executed before the queue_backend gets the data so that any queue_backend can benefit transparently from compression without having to add any code to handle it. Due to our design, this also means that if the queue_backend stores envelopes and messages remotely, then the compression will take place before sending and after fetching, the remote end only ever sees compressed data.