From 4633f516a7328ce5208673c224a1841fcb63ce83 Mon Sep 17 00:00:00 2001 From: Emil Lerch Date: Thu, 28 Oct 2021 10:51:07 -0700 Subject: [PATCH] add project update to readme --- README.md | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/README.md b/README.md index 4e5a39a..6fa30ad 100644 --- a/README.md +++ b/README.md @@ -19,6 +19,25 @@ Running strip on the executable after compilation (it seems zig strip only goes so far), reduces this to 4.3M. This is for x86_linux, (which is all that's tested at the moment). +# 2021-10-28 SDK Update + +To get smaller executable size and better portability with faster compilation, +my intent is to rework the http communications using +[requestz](https://github.com/ducdetronquito/requestz). This relies on a +couple other projects, and will require the creation of a zig implementation +for [Signature Version 4](https://docs.aws.amazon.com/general/latest/gr/signature-version-4.html), +along with support for some S3 quirks, etc. It will also reduce compatibility +with some edge cases. Long term I think this is a better approach, however and +will remove (or replace) a ton of the C dependencies as well as avoid a bunch +of complexity such as the event loop C implementation found in the AWS +libraries, which eventually could be replaced by zig async. I will create a new branch +for this work as master is currently working fairly well. My hope is to +converge on Zig 0.9.0 so this SDK doesn't have to keep up with changes going on +in zig master. I'd also love to have an official package manager in zig, but I +know this is high on the priority list for the foundation. My goal is to **not** require +[gyro](https://github.com/mattnite/gyro) as requestz does, but we'll see how +that develops. + ## Building I am assuming here that if you're playing with zig, you pretty much know