lobo/lambda-zig
1
1
Fork 1
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Compare commits

base: lobo:master
Branches Tags
lobo:master
..
compare: lobo:6a14eee1cfd1e7ac12de0dcf0d152bf7721051ac
Branches Tags
lobo:master

No commits in common. "master" and "6a14eee1cfd1e7ac12de0dcf0d152bf7721051ac" have entirely different histories.
master ... 6a14eee1cf

17 changed files with 275 additions and 1415 deletions
Show Stats Expand all files Collapse all files

29
.github/workflows/zig-build.yaml vendored
View File

@ -1,29 +0,0 @@
name: Generic zig build
on:
workflow_dispatch:
push:
branches:
- '*'
- '!zig-develop*'
jobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: elerch/setup-zig@v3
with:
version: 0.13.0
- uses: elerch/zig-action-cache@v1.1.6
- name: Build project
run: zig build --summary all
- name: Run tests
run: zig build test --summary all
- name: Notify
uses: elerch/action-notify-ntfy@v2.github
if: always() && env.GITEA_ACTIONS == 'true'
with:
host: ${{ secrets.NTFY_HOST }}
topic: ${{ secrets.NTFY_TOPIC }}
status: ${{ job.status }}
user: ${{ secrets.NTFY_USER }}
password: ${{ secrets.NTFY_PASSWORD }}

3
.gitignore vendored
View File

@ -1,3 +1,4 @@
.gyro/
zig-cache/
zig-out/
.zig-cache
deps.zig

21
LICENSE
View File

@ -1,21 +0,0 @@
MIT License
Copyright (c) 2021 Emil Lerch
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

80
README.md
View File

@ -1,70 +1,38 @@
lambda-zig: A Custom Runtime for AWS Lambda
===========================================
This is a sample custom runtime built in zig (0.13). Simple projects will execute
This is a sample custom runtime built in zig. Simple projects will execute
in <1ms, with a cold start init time of approximately 11ms.
Some custom build steps have been added to build.zig, which will only currently appear if compiling from a linux operating system:
Some custom build steps have been added to build.zig:
* `zig build awslambda_iam`: Deploy and record a default IAM role for the lambda function
* `zig build awslambda_package`: Package the lambda function for upload
* `zig build awslambda_deploy`: Deploy the lambda function
* `zig build awslambda_run`: Run the lambda function
* `zig build iam`: Deploy and record a default IAM role for the lambda function
* `zig build package`: Package the lambda function for upload
* `zig build deploy`: Deploy the lambda function
* `zig build run`: Run the lambda function
Custom options:
* **debug**: boolean flag to avoid the debug symbols to be stripped. Useful to see
error return traces in the AWS Lambda logs
* **function-name**: set the name of the AWS Lambda function
* **payload**: Use this to set the payload of the function when run using `zig build awslambda_run`
* **region**: Use this to set the region for the function deployment/run
* **function-role**: Name of the role to use for the function. The system will
look up the arn from this name, and create if it does not exist
* **function-arn**: Role arn to use with the function. This must exist
* **payload**: Use this to set the payload of the function when run using `zig build run`
Additionally, a custom IAM role can be used for the function by appending ``-- --role myawesomerole``
to the `zig build deploy` command. This has not really been tested. The role name
is cached in zig-out/bin/iam_role_name, so you can also just set that to the full
arn of your iam role if you'd like.
The AWS Lambda function is compiled as a linux ARM64 executable. Since the build.zig
calls out to the shell for AWS operations, you will need AWS CLI v2.2.43 or greater.
This project vendors dependencies with [gyro](https://github.com/mattnite/gyro), so
first time build should be done with `gyro build`. This should be working
on zig master - certain build.zig constructs are not available in zig 0.8.1.
The AWS Lambda function can be compiled as a linux x86_64 or linux aarch64
executable. The build script will set the architecture appropriately
Caveats:
* Building on Windows will not yet work, as the package step still uses
system commands due to the need to create a zip file, and the current lack
of zip file creation capabilities in the standard library (you can read, but
not write, zip files with the standard library). A TODO exists with more
information should you wish to file a PR.
A sample project using this runtime can be found at https://git.lerch.org/lobo/lambda-zig-sample
Using the zig package manager
-----------------------------
The zig package manager [works just fine](https://github.com/ziglang/zig/issues/14279)
in build.zig, which works well for use of this runtime.
To add lambda package/deployment steps to another project:
1. `zig build init-exe`
2. Add a `build.zig.zon` similar to the below
3. Add a line to build.zig to add necessary build options, etc. Not the build function
return type should be changed from `void` to `!void`
`build.zig`:
```zig
try @import("lambda-zig").lambdaBuildOptions(b, exe);
```
`build.zig.zon`:
```zig
.{
.name = "lambda-zig",
.version = "0.1.0",
.dependencies = .{
.@"lambda-zig" = .{
.url = "https://git.lerch.org/lobo/lambda-zig/archive/fa13a08c4d91034a9b19d85f8c4c0af4cedaa67e.tar.gz",
.hash = "122037c357f834ffddf7b3a514f55edd5a4d7a3cde138a4021b6ac51be8fd2926000",
},
},
}
```
That's it! Now you should have the 4 custom build steps
* Small inbound lambda payloads seem to be confusing [requestz](https://github.com/ducdetronquito/requestz),
which just never returns, causing timeouts
* Unhandled invocation errors seem to be causing the same problem

226
build.zig
View File

@ -1,88 +1,174 @@
const builtin = @import("builtin");
const std = @import("std");
const pkgs = @import("deps.zig").pkgs;
// Although this function looks imperative, note that its job is to
// declaratively construct a build graph that will be executed by an external
// runner.
pub fn build(b: *std.Build) !void {
pub fn build(b: *std.build.Builder) !void {
// Standard target options allows the person running `zig build` to choose
// what target to build for. Here we do not override the defaults, which
// means any target is allowed, and the default is native. Other options
// for restricting supported target set are available.
const target = b.standardTargetOptions(.{});
// We want the target to be aarch64-linux for deploys
const target = std.zig.CrossTarget{
.cpu_arch = .aarch64,
.os_tag = .linux,
};
// Standard optimization options allow the person running `zig build` to select
// between Debug, ReleaseSafe, ReleaseFast, and ReleaseSmall. Here we do not
// set a preferred release mode, allowing the user to decide how to optimize.
const optimize = b.standardOptimizeOption(.{});
// Standard release options allow the person running `zig build` to select
// between Debug, ReleaseSafe, ReleaseFast, and ReleaseSmall.
// const mode = b.standardReleaseOptions();
const lib = b.addStaticLibrary(.{
.name = "lambda-zig",
// In this case the main source file is merely a path, however, in more
// complicated build scripts, this could be a generated file.
.root_source_file = b.path("src/lambda.zig"),
.target = target,
.optimize = optimize,
});
const exe = b.addExecutable("bootstrap", "src/main.zig");
_ = b.addModule("lambda_runtime", .{
.root_source_file = b.path("src/lambda.zig"),
.target = target,
.optimize = optimize,
});
pkgs.addAllTo(exe);
exe.setTarget(target);
exe.setBuildMode(.ReleaseSafe);
const debug = b.option(bool, "debug", "Debug mode (do not strip executable)") orelse false;
exe.strip = !debug;
exe.install();
// This declares intent for the library to be installed into the standard
// location when the user invokes the "install" step (the default step when
// running `zig build`).
b.installArtifact(lib);
// TODO: We can cross-compile of course, but stripping and zip commands
// may vary
if (std.builtin.os.tag == .linux) {
// Package step
const package_step = b.step("package", "Package the function");
package_step.dependOn(b.getInstallStep());
// strip may not be installed or work for the target arch
// TODO: make this much less fragile
const strip = if (debug)
try std.fmt.allocPrint(b.allocator, "true", .{})
else
try std.fmt.allocPrint(b.allocator, "[ -x /usr/aarch64-linux-gnu/bin/strip ] && /usr/aarch64-linux-gnu/bin/strip {s}", .{b.getInstallPath(exe.install_step.?.dest_dir, exe.install_step.?.artifact.out_filename)});
defer b.allocator.free(strip);
package_step.dependOn(&b.addSystemCommand(&.{ "/bin/sh", "-c", strip }).step);
const function_zip = b.getInstallPath(exe.install_step.?.dest_dir, "function.zip");
const zip = try std.fmt.allocPrint(b.allocator, "zip -qj9 {s} {s}", .{ function_zip, b.getInstallPath(exe.install_step.?.dest_dir, exe.install_step.?.artifact.out_filename) });
defer b.allocator.free(zip);
package_step.dependOn(&b.addSystemCommand(&.{ "/bin/sh", "-c", zip }).step);
// Creates a step for unit testing. This only builds the test executable
// but does not run it.
const main_tests = b.addTest(.{
.root_source_file = b.path("src/lambda.zig"),
.target = target,
.optimize = optimize,
});
// Deployment
const deploy_step = b.step("deploy", "Deploy the function");
var deal_with_iam = true;
if (b.args) |args| {
for (args) |arg| {
if (std.mem.eql(u8, "--role", arg)) {
deal_with_iam = false;
break;
}
}
}
var iam_role: []u8 = &.{};
const iam_step = b.step("iam", "Create/Get IAM role for function");
deploy_step.dependOn(iam_step); // iam_step will either be a noop or all the stuff below
if (deal_with_iam) {
// if someone adds '-- --role arn...' to the command line, we don't
// need to do anything with the iam role. Otherwise, we'll create/
// get the IAM role and stick the name in a file in our destination
// directory to be used later
const iam_role_name_file = b.getInstallPath(exe.install_step.?.dest_dir, "iam_role_name");
iam_role = try std.fmt.allocPrint(b.allocator, "--role $(cat {s})", .{iam_role_name_file});
// defer b.allocator.free(iam_role);
if (!fileExists(iam_role_name_file)) {
// Role get/creation command
const ifstatement_fmt =
\\ if aws iam get-role --role-name lambda_basic_execution 2>&1 |grep -q NoSuchEntity; then aws iam create-role --output text --query Role.Arn --role-name lambda_basic_execution --assume-role-policy-document '{
\\ "Version": "2012-10-17",
\\ "Statement": [
\\ {
\\ "Sid": "",
\\ "Effect": "Allow",
\\ "Principal": {
\\ "Service": "lambda.amazonaws.com"
\\ },
\\ "Action": "sts:AssumeRole"
\\ }
\\ ]}' > /dev/null; fi && \
\\ aws iam attach-role-policy --policy-arn arn:aws:iam::aws:policy/AWSLambdaExecute --role-name lambda_basic_execution && \
\\ aws iam get-role --role-name lambda_basic_execution --query Role.Arn --output text >
;
const run_main_tests = b.addRunArtifact(main_tests);
const ifstatement = try std.mem.concat(b.allocator, u8, &[_][]const u8{ ifstatement_fmt, iam_role_name_file });
defer b.allocator.free(ifstatement);
iam_step.dependOn(&b.addSystemCommand(&.{ "/bin/sh", "-c", ifstatement }).step);
}
}
const function_name = b.option([]const u8, "function-name", "Function name for Lambda [zig-fn]") orelse "zig-fn";
const function_name_file = b.getInstallPath(exe.install_step.?.dest_dir, function_name);
const ifstatement = "if [ ! -f {s} ] || [ {s} -nt {s} ]; then if aws lambda get-function --function-name {s} 2>&1 |grep -q ResourceNotFoundException; then echo not found > /dev/null; {s}; else echo found > /dev/null; {s}; fi; fi";
// The architectures option was introduced in 2.2.43 released 2021-10-01
// We want to use arm64 here because it is both faster and cheaper for most
// Amazon Linux 2 is the only arm64 supported option
const not_found = "aws lambda create-function --architectures arm64 --runtime provided.al2 --function-name {s} --zip-file fileb://{s} --handler not_applicable {s} && touch {s}";
const not_found_fmt = try std.fmt.allocPrint(b.allocator, not_found, .{ function_name, function_zip, iam_role, function_name_file });
defer b.allocator.free(not_found_fmt);
const found = "aws lambda update-function-code --function-name {s} --zip-file fileb://{s} && touch {s}";
const found_fmt = try std.fmt.allocPrint(b.allocator, found, .{ function_name, function_zip, function_name_file });
defer b.allocator.free(found_fmt);
var found_final: []const u8 = undefined;
var not_found_final: []const u8 = undefined;
if (b.args) |args| {
found_final = try addArgs(b.allocator, found_fmt, args);
not_found_final = try addArgs(b.allocator, not_found_fmt, args);
} else {
found_final = found_fmt;
not_found_final = not_found_fmt;
}
const cmd = try std.fmt.allocPrint(b.allocator, ifstatement, .{
function_name_file,
std.fs.path.dirname(exe.root_src.?.path),
function_name_file,
function_name,
not_found_fmt,
found_fmt,
});
// This creates a build step. It will be visible in the `zig build --help` menu,
// and can be selected like this: `zig build test`
// This will evaluate the `test` step rather than the default, which is "install".
const test_step = b.step("test", "Run library tests");
test_step.dependOn(&run_main_tests.step);
defer b.allocator.free(cmd);
const exe = b.addExecutable(.{
.name = "custom",
.root_source_file = b.path("src/sample-main.zig"),
.target = target,
.optimize = optimize,
});
// std.debug.print("{s}\n", .{cmd});
deploy_step.dependOn(package_step);
deploy_step.dependOn(&b.addSystemCommand(&.{ "/bin/sh", "-c", cmd }).step);
b.installArtifact(exe);
try lambdaBuildOptions(b, exe);
// TODO: Looks like IquanaTLS isn't playing nicely with payloads this small
// const payload = b.option([]const u8, "payload", "Lambda payload [{\"foo\":\"bar\"}]") orelse
// \\ {"foo": "bar"}"
// ;
const payload = b.option([]const u8, "payload", "Lambda payload [{\"foo\":\"bar\", \"baz\": \"qux\"}]") orelse
\\ {"foo": "bar", "baz": "qux"}"
;
const run_script =
\\ f=$(mktemp) && \
\\ logs=$(aws lambda invoke \
\\ --cli-binary-format raw-in-base64-out \
\\ --invocation-type RequestResponse \
\\ --function-name {s} \
\\ --payload '{s}' \
\\ --log-type Tail \
\\ --query LogResult \
\\ --output text "$f" |base64 -d) && \
\\ cat "$f" && rm "$f" && \
\\ echo && echo && echo "$logs"
;
const run_script_fmt = try std.fmt.allocPrint(b.allocator, run_script, .{ function_name, payload });
defer b.allocator.free(run_script_fmt);
const run_cmd = b.addSystemCommand(&.{ "/bin/sh", "-c", run_script_fmt });
run_cmd.step.dependOn(deploy_step);
if (b.args) |args| {
run_cmd.addArgs(args);
}
const run_step = b.step("run", "Run the app");
run_step.dependOn(&run_cmd.step);
}
}
/// lambdaBuildOptions will add three build options to the build (if compiling
/// the code on a Linux host):
///
/// * package: Packages the function for deployment to Lambda
/// (dependencies are the zip executable and a shell)
/// * iam: Gets an IAM role for the Lambda function, and creates it if it does not exist
/// (dependencies are the AWS CLI, grep and a shell)
/// * deploy: Deploys the lambda function to a live AWS environment
/// (dependencies are the AWS CLI, and a shell)
/// * remoterun: Runs the lambda function in a live AWS environment
/// (dependencies are the AWS CLI, and a shell)
///
/// remoterun depends on deploy
/// deploy depends on iam and package
///
/// iam and package do not have any dependencies
pub fn lambdaBuildOptions(b: *std.Build, exe: *std.Build.Step.Compile) !void {
const function_name = b.option([]const u8, "function-name", "Function name for Lambda [zig-fn]") orelse "zig-fn";
try @import("lambdabuild.zig").configureBuild(b, exe, function_name);
fn fileExists(file_name: []const u8) bool {
const file = std.fs.openFileAbsolute(file_name, .{}) catch return false;
defer file.close();
return true;
}
pub fn configureBuild(b: *std.Build, exe: *std.Build.Step.Compile, function_name: []const u8) !void {
try @import("lambdabuild.zig").configureBuild(b, exe, function_name);
fn addArgs(allocator: *std.mem.Allocator, original: []const u8, args: [][]const u8) ![]const u8 {
var rc = original;
for (args) |arg| {
rc = try std.mem.concat(allocator, u8, &.{ rc, " ", arg });
}
return rc;
}

33
build.zig.zon
View File

@ -1,33 +0,0 @@
.{
.name = "lambda-zig",
// This is a [Semantic Version](https://semver.org/).
// In a future version of Zig it will be used for package deduplication.
.version = "0.0.0",
// This field is optional.
// This is currently advisory only; Zig does not yet do anything
// with this value.
.dependencies = .{
.aws = .{
.url = "https://git.lerch.org/api/packages/lobo/generic/aws-sdk-with-models/908c9d2d429b1f38c835363db566aa17bf1742fd/908c9d2d429b1f38c835363db566aa17bf1742fd-with-models.tar.gz",
.hash = "122022770a177afb2ee46632f88ad5468a5dea8df22170d1dea5163890b0a881399d",
},
},
.minimum_zig_version = "0.12.0",
// Specifies the set of files and directories that are included in this package.
// Only files and directories listed here are included in the `hash` that
// is computed for this package.
// Paths are relative to the build root. Use the empty string (`""`) to refer to
// the build root itself.
// A directory listed here means that all files within, recursively, are included.
.paths = .{
"build.zig",
"build.zig.zon",
"lambdabuild.zig",
"src",
"lambdabuild",
"LICENSE",
"README.md",
},
}

5
gyro.lock Normal file
View File

@ -0,0 +1,5 @@
pkg default ducdetronquito http 0.1.3
pkg default ducdetronquito h11 0.1.1
github nektro iguanaTLS 953ad821fae6c920fb82399493663668cd91bde7 src/main.zig 953ad821fae6c920fb82399493663668cd91bde7
github MasterQ32 zig-network 15b88658809cac9022ec7d59449b0cd3ebfd0361 network.zig 15b88658809cac9022ec7d59449b0cd3ebfd0361
github elerch requestz 1fa8157641300805b9503f98cd201d0959d19631 src/main.zig 1fa8157641300805b9503f98cd201d0959d19631

7
gyro.zzz Normal file
View File

@ -0,0 +1,7 @@
deps:
requestz:
src:
github:
user: elerch
repo: requestz
ref: 1fa8157641300805b9503f98cd201d0959d19631

168
lambdabuild.zig
View File

@ -1,168 +0,0 @@
const std = @import("std");
const builtin = @import("builtin");
const Package = @import("lambdabuild/Package.zig");
const Iam = @import("lambdabuild/Iam.zig");
const Deploy = @import("lambdabuild/Deploy.zig");
const Invoke = @import("lambdabuild/Invoke.zig");
fn fileExists(file_name: []const u8) bool {
const file = std.fs.openFileAbsolute(file_name, .{}) catch return false;
defer file.close();
return true;
}
fn addArgs(allocator: std.mem.Allocator, original: []const u8, args: [][]const u8) ![]const u8 {
var rc = original;
for (args) |arg| {
rc = try std.mem.concat(allocator, u8, &.{ rc, " ", arg });
}
return rc;
}
/// lambdaBuildSteps will add four build steps to the build (if compiling
/// the code on a Linux host):
///
/// * awslambda_package: Packages the function for deployment to Lambda
/// (dependencies are the zip executable and a shell)
/// * awslambda_iam: Gets an IAM role for the Lambda function, and creates it if it does not exist
/// (dependencies are the AWS CLI, grep and a shell)
/// * awslambda_deploy: Deploys the lambda function to a live AWS environment
/// (dependencies are the AWS CLI, and a shell)
/// * awslambda_run: Runs the lambda function in a live AWS environment
/// (dependencies are the AWS CLI, and a shell)
///
/// awslambda_run depends on deploy
/// awslambda_deploy depends on iam and package
///
/// iam and package do not have any dependencies
pub fn configureBuild(b: *std.Build, exe: *std.Build.Step.Compile, function_name: []const u8) !void {
// The rest of this function is currently reliant on the use of Linux
// system being used to build the lambda function
//
// It is likely that much of this will work on other Unix-like OSs, but
// we will work this out later
//
// TODO: support other host OSs
if (builtin.os.tag != .linux) return;
@import("aws").aws.globalLogControl(.info, .warn, .info, false);
const package_step = Package.create(b, .{ .exe = exe });
const step = b.step("awslambda_package", "Package the function");
step.dependOn(&package_step.step);
package_step.step.dependOn(b.getInstallStep());
// Doing this will require that the aws dependency be added to the downstream
// build.zig.zon
// const lambdabuild = b.addExecutable(.{
// .name = "lambdabuild",
// .root_source_file = .{
// // we use cwd_relative here because we need to compile this relative
// // to whatever directory this file happens to be. That is likely
// // in a cache directory, not the base of the build.
// .cwd_relative = try std.fs.path.join(b.allocator, &[_][]const u8{
// std.fs.path.dirname(@src().file).?,
// "lambdabuild/src/main.zig",
// }),
// },
// .target = b.host,
// });
// const aws_dep = b.dependency("aws", .{
// .target = b.host,
// .optimize = lambdabuild.root_module.optimize orelse .Debug,
// });
// const aws_module = aws_dep.module("aws");
// lambdabuild.root_module.addImport("aws", aws_module);
//
const iam_role_name = b.option(
[]const u8,
"function-role",
"IAM role name for function (will create if it does not exist) [lambda_basic_execution]",
) orelse "lambda_basic_execution_blah2";
const iam_role_arn = b.option(
[]const u8,
"function-arn",
"Preexisting IAM role arn for function",
);
const iam = Iam.create(b, .{
.role_name = iam_role_name,
.role_arn = iam_role_arn,
});
const iam_step = b.step("awslambda_iam", "Create/Get IAM role for function");
iam_step.dependOn(&iam.step);
const region = try b.allocator.create(@import("lambdabuild/Region.zig"));
region.* = .{
.allocator = b.allocator,
.specified_region = b.option([]const u8, "region", "Region to use [default is autodetect from environment/config]"),
};
// Deployment
const deploy = Deploy.create(b, .{
.name = function_name,
.arch = exe.root_module.resolved_target.?.result.cpu.arch,
.iam_step = iam,
.package_step = package_step,
.region = region,
});
const deploy_step = b.step("awslambda_deploy", "Deploy the function");
deploy_step.dependOn(&deploy.step);
const payload = b.option([]const u8, "payload", "Lambda payload [{\"foo\":\"bar\", \"baz\": \"qux\"}]") orelse
\\ {"foo": "bar", "baz": "qux"}"
;
const invoke = Invoke.create(b, .{
.name = function_name,
.payload = payload,
.region = region,
});
invoke.step.dependOn(&deploy.step);
const run_step = b.step("awslambda_run", "Run the app in AWS lambda");
run_step.dependOn(&invoke.step);
}
// AWS_CONFIG_FILE (default is ~/.aws/config
// AWS_DEFAULT_REGION
fn findRegionFromSystem(allocator: std.mem.Allocator) ![]const u8 {
const env_map = try std.process.getEnvMap(allocator);
if (env_map.get("AWS_DEFAULT_REGION")) |r| return r;
const config_file_path = env_map.get("AWS_CONFIG_FILE") orelse
try std.fs.path.join(allocator, &[_][]const u8{
env_map.get("HOME") orelse env_map.get("USERPROFILE").?,
".aws",
"config",
});
const config_file = try std.fs.openFileAbsolute(config_file_path, .{});
defer config_file.close();
const config_bytes = try config_file.readToEndAlloc(allocator, 1024 * 1024);
const profile = env_map.get("AWS_PROFILE") orelse "default";
var line_iterator = std.mem.split(u8, config_bytes, "\n");
var in_profile = false;
while (line_iterator.next()) |line| {
const trimmed = std.mem.trim(u8, line, " \t\r");
if (trimmed.len == 0 or trimmed[0] == '#') continue;
if (!in_profile) {
if (trimmed[0] == '[' and trimmed[trimmed.len - 1] == ']') {
// this is a profile directive!
// std.debug.print("profile: {s}, in file: {s}\n", .{ profile, trimmed[1 .. trimmed.len - 1] });
if (std.mem.eql(u8, profile, trimmed[1 .. trimmed.len - 1])) {
in_profile = true;
}
}
continue; // we're only looking for a profile at this point
}
// look for our region directive
if (trimmed[0] == '[' and trimmed[trimmed.len - 1] == ']')
return error.RegionNotFound; // we've hit another profile without getting our region
if (!std.mem.startsWith(u8, trimmed, "region")) continue;
var equalityiterator = std.mem.split(u8, trimmed, "=");
_ = equalityiterator.next() orelse return error.RegionNotFound;
const raw_val = equalityiterator.next() orelse return error.RegionNotFound;
return try allocator.dupe(u8, std.mem.trimLeft(u8, raw_val, " \t"));
}
return error.RegionNotFound;
}

165
lambdabuild/Deploy.zig
View File

@ -1,165 +0,0 @@
const std = @import("std");
const Region = @import("Region.zig");
const aws = @import("aws").aws;
const Deploy = @This();
step: std.Build.Step,
options: Options,
const base_id: std.Build.Step.Id = .custom;
pub const Options = struct {
/// Function name to be used for the function
name: []const u8,
/// Architecture for Lambda function
arch: std.Target.Cpu.Arch,
/// Iam step. This will be a dependency of the deployment
iam_step: *@import("Iam.zig"),
/// Packaging step. This will be a dependency of the deployment
package_step: *@import("Package.zig"),
/// Region for deployment
region: *Region,
};
pub fn create(owner: *std.Build, options: Options) *Deploy {
const name = owner.dupe(options.name);
const step_name = owner.fmt("{s} {s}{s}", .{
"aws lambda",
"deploy ",
name,
});
const self = owner.allocator.create(Deploy) catch @panic("OOM");
self.* = .{
.step = std.Build.Step.init(.{
.id = base_id,
.name = step_name,
.owner = owner,
.makeFn = make,
}),
.options = options,
};
self.step.dependOn(&options.iam_step.step);
self.step.dependOn(&options.package_step.step);
return self;
}
/// gets the last time we deployed this function from the name in cache.
/// If not in cache, null is returned. Note that cache is not account specific,
/// so if you're banging around multiple accounts, you'll want to use different
/// local zig caches for each
fn getlastDeployedTime(step: *std.Build.Step, name: []const u8) !?[]const u8 {
try step.owner.cache_root.handle.makePath("iam");
// we should be able to use the role name, as only the following characters
// are allowed: _+=,.@-.
const cache_file = try std.fmt.allocPrint(
step.owner.allocator,
"deploy{s}{s}",
.{ std.fs.path.sep_str, name },
);
const buff = try step.owner.allocator.alloc(u8, 64);
const time = step.owner.cache_root.handle.readFile(cache_file, buff) catch return null;
return time;
}
fn make(step: *std.Build.Step, node: std.Progress.Node) anyerror!void {
_ = node;
const self: *Deploy = @fieldParentPtr("step", step);
if (self.options.arch != .aarch64 and self.options.arch != .x86_64)
return step.fail("AWS Lambda can only deploy aarch64 and x86_64 functions ({} not allowed)", .{self.options.arch});
const last_packaged_sha256 = blk: {
// file should always be there, but we shouldn't break if the cache doesn't exist
const last_deployed_id_file = std.fs.openFileAbsolute(try self.options.package_step.shasumFilePath(), .{}) catch break :blk null;
defer last_deployed_id_file.close();
break :blk try last_deployed_id_file.readToEndAlloc(step.owner.allocator, 2048);
};
var client = aws.Client.init(self.step.owner.allocator, .{});
defer client.deinit();
const services = aws.Services(.{.lambda}){};
const function = blk: {
var diagnostics = aws.Diagnostics{
.http_code = undefined,
.response_body = undefined,
.allocator = self.step.owner.allocator,
};
const options = aws.Options{
.client = client,
.diagnostics = &diagnostics,
.region = try self.options.region.region(),
};
aws.globalLogControl(.info, .warn, .info, true);
defer aws.globalLogControl(.info, .warn, .info, false);
const call = aws.Request(services.lambda.get_function).call(.{
.function_name = self.options.name,
}, options) catch |e| {
// There seems an issue here, but realistically, we have an arena
// so there's no leak leaving this out
defer diagnostics.deinit();
if (diagnostics.http_code == 404) break :blk null;
return step.fail(
"Unknown error {} from Lambda GetFunction. HTTP code {}, message: {s}",
.{ e, diagnostics.http_code, diagnostics.response_body },
);
};
defer call.deinit();
break :blk .{
.last_modified = try step.owner.allocator.dupe(u8, call.response.configuration.?.last_modified.?),
.revision_id = try step.owner.allocator.dupe(u8, call.response.configuration.?.revision_id.?),
.sha256 = try step.owner.allocator.dupe(u8, call.response.configuration.?.code_sha256.?),
};
};
if (last_packaged_sha256) |s|
if (function) |f|
if (std.mem.eql(u8, s, f.sha256)) {
step.result_cached = true;
return;
};
const encoder = std.base64.standard.Encoder;
const file = try std.fs.openFileAbsolute(self.options.package_step.packagedFileLazyPath().getPath2(step.owner, step), .{});
defer file.close();
const bytes = try file.readToEndAlloc(step.owner.allocator, 100 * 1024 * 1024);
const base64_buf = try step.owner.allocator.alloc(u8, encoder.calcSize(bytes.len));
const base64_bytes = encoder.encode(base64_buf, bytes);
const options = aws.Options{
.client = client,
.region = try self.options.region.region(),
};
const arm64_arch = [_][]const u8{"arm64"};
const x86_64_arch = [_][]const u8{"x86_64"};
const architectures = (if (self.options.arch == .aarch64) arm64_arch else x86_64_arch);
const arches: [][]const u8 = @constCast(architectures[0..]);
if (function) |f| {
// TODO: make sure our zipfile newer than the lambda function
const update_call = try aws.Request(services.lambda.update_function_code).call(.{
.function_name = self.options.name,
.architectures = arches,
.revision_id = f.revision_id,
.zip_file = base64_bytes,
}, options);
defer update_call.deinit();
} else {
// New function - we need to create from scratch
const create_call = try aws.Request(services.lambda.create_function).call(.{
.function_name = self.options.name,
.architectures = arches,
.code = .{ .zip_file = base64_bytes },
.handler = "not_applicable",
.package_type = "Zip",
.runtime = "provided.al2",
.role = self.options.iam_step.resolved_arn,
}, options);
defer create_call.deinit();
}
}

146
lambdabuild/Iam.zig
View File

@ -1,146 +0,0 @@
const std = @import("std");
const aws = @import("aws").aws;
const Iam = @This();
step: std.Build.Step,
options: Options,
/// resolved_arn will be set only after make is run
resolved_arn: []const u8 = undefined,
arn_buf: [2048]u8 = undefined, // https://docs.aws.amazon.com/IAM/latest/APIReference/API_Role.html has 2k limit
const base_id: std.Build.Step.Id = .custom;
pub const Options = struct {
name: []const u8 = "",
role_name: []const u8,
role_arn: ?[]const u8,
};
pub fn create(owner: *std.Build, options: Options) *Iam {
const name = owner.dupe(options.name);
const step_name = owner.fmt("{s} {s}{s}", .{
"aws lambda",
"iam",
name,
});
const self = owner.allocator.create(Iam) catch @panic("OOM");
self.* = .{
.step = std.Build.Step.init(.{
.id = base_id,
.name = step_name,
.owner = owner,
.makeFn = make,
}),
.options = options,
};
return self;
}
/// gets an IamArn from the name in cache. If not in cache, null is returned
/// Note that cache is not account specific, so if you're banging around multiple
/// accounts, you'll want to use different local zig caches for each
pub fn getIamArnFromName(step: *std.Build.Step, name: []const u8) !?[]const u8 {
try step.owner.cache_root.handle.makePath("iam");
// we should be able to use the role name, as only the following characters
// are allowed: _+=,.@-.
const iam_file = try std.fmt.allocPrint(
step.owner.allocator,
"iam{s}{s}",
.{ std.fs.path.sep_str, name },
);
const buff = try step.owner.allocator.alloc(u8, 64);
const arn = step.owner.cache_root.handle.readFile(iam_file, buff) catch return null;
return arn;
}
fn make(step: *std.Build.Step, node: std.Progress.Node) anyerror!void {
_ = node;
const self: *Iam = @fieldParentPtr("step", step);
if (try getIamArnFromName(step, self.options.role_name)) |a| {
step.result_cached = true;
@memcpy(self.arn_buf[0..a.len], a);
self.resolved_arn = self.arn_buf[0..a.len];
return; // exists in cache - nothing to do
}
var client = aws.Client.init(self.step.owner.allocator, .{});
defer client.deinit();
const services = aws.Services(.{.iam}){};
var arn = blk: {
var diagnostics = aws.Diagnostics{
.http_code = undefined,
.response_body = undefined,
.allocator = self.step.owner.allocator,
};
const options = aws.Options{
.client = client,
.diagnostics = &diagnostics,
};
const call = aws.Request(services.iam.get_role).call(.{
.role_name = self.options.role_name, // TODO: if we have a role_arn, we should use it and skip
}, options) catch |e| {
defer diagnostics.deinit();
if (diagnostics.http_code == 404) break :blk null;
return step.fail(
"Unknown error {} from IAM GetRole. HTTP code {}, message: {s}",
.{ e, diagnostics.http_code, diagnostics.response_body },
);
};
defer call.deinit();
break :blk try step.owner.allocator.dupe(u8, call.response.role.arn);
};
// Now ARN will either be null (does not exist), or a value
if (arn == null) {
// we need to create the role before proceeding
const options = aws.Options{
.client = client,
};
const create_call = try aws.Request(services.iam.create_role).call(.{
.role_name = self.options.role_name,
.assume_role_policy_document =
\\{
\\ "Version": "2012-10-17",
\\ "Statement": [
\\ {
\\ "Sid": "",
\\ "Effect": "Allow",
\\ "Principal": {
\\ "Service": "lambda.amazonaws.com"
\\ },
\\ "Action": "sts:AssumeRole"
\\ }
\\ ]
\\}
,
}, options);
defer create_call.deinit();
arn = try step.owner.allocator.dupe(u8, create_call.response.role.arn);
const attach_call = try aws.Request(services.iam.attach_role_policy).call(.{
.policy_arn = "arn:aws:iam::aws:policy/AWSLambdaExecute",
.role_name = self.options.role_name,
}, options);
defer attach_call.deinit();
}
@memcpy(self.arn_buf[0..arn.?.len], arn.?);
self.resolved_arn = self.arn_buf[0..arn.?.len];
// NOTE: This must match getIamArnFromName
const iam_file = try std.fmt.allocPrint(
step.owner.allocator,
"iam{s}{s}",
.{ std.fs.path.sep_str, self.options.role_name },
);
try step.owner.cache_root.handle.writeFile(.{
.sub_path = iam_file,
.data = arn.?,
});
}

90
lambdabuild/Invoke.zig
View File

@ -1,90 +0,0 @@
const std = @import("std");
const aws = @import("aws").aws;
const Region = @import("Region.zig");
const Invoke = @This();
step: std.Build.Step,
options: Options,
const base_id: std.Build.Step.Id = .custom;
pub const Options = struct {
/// Function name to invoke
name: []const u8,
/// Payload to send to the function
payload: []const u8,
/// Region for deployment
region: *Region,
};
pub fn create(owner: *std.Build, options: Options) *Invoke {
const name = owner.dupe(options.name);
const step_name = owner.fmt("{s} {s}{s}", .{
"aws lambda",
"invoke",
name,
});
const self = owner.allocator.create(Invoke) catch @panic("OOM");
self.* = .{
.step = std.Build.Step.init(.{
.id = base_id,
.name = step_name,
.owner = owner,
.makeFn = make,
}),
.options = options,
};
return self;
}
fn make(step: *std.Build.Step, node: std.Progress.Node) anyerror!void {
_ = node;
const self: *Invoke = @fieldParentPtr("step", step);
var client = aws.Client.init(self.step.owner.allocator, .{});
defer client.deinit();
const services = aws.Services(.{.lambda}){};
const options = aws.Options{
.client = client,
.region = try self.options.region.region(),
};
var inx: usize = 10; // 200ms * 10
while (inx > 0) : (inx -= 1) {
var diagnostics = aws.Diagnostics{
.http_code = undefined,
.response_body = undefined,
.allocator = self.step.owner.allocator,
};
const call = aws.Request(services.lambda.get_function).call(.{
.function_name = self.options.name,
}, options) catch |e| {
// There seems an issue here, but realistically, we have an arena
// so there's no leak leaving this out
defer diagnostics.deinit();
if (diagnostics.http_code == 404) continue; // function was just created...it's ok
return step.fail(
"Unknown error {} from Lambda GetFunction. HTTP code {}, message: {s}",
.{ e, diagnostics.http_code, diagnostics.response_body },
);
};
defer call.deinit();
if (!std.mem.eql(u8, "InProgress", call.response.configuration.?.last_update_status.?))
break; // We're ready to invoke!
const ms: usize = if (inx == 5) 500 else 50;
std.time.sleep(ms * std.time.ns_per_ms);
}
if (inx == 0)
return step.fail("Timed out waiting for lambda to update function", .{});
const call = try aws.Request(services.lambda.invoke).call(.{
.function_name = self.options.name,
.payload = self.options.payload,
.log_type = "Tail",
.invocation_type = "RequestResponse",
}, options);
defer call.deinit();
std.debug.print("{?s}\n", .{call.response.payload});
}

158
lambdabuild/Package.zig
View File

@ -1,158 +0,0 @@
const std = @import("std");
const Package = @This();
step: std.Build.Step,
options: Options,
/// This is set as part of the make phase, and is the location in the cache
/// for the lambda package. The package will also be copied to the output
/// directory, but this location makes for a good cache key for deployments
zipfile_cache_dest: ?[]const u8 = null,
zipfile_dest: ?[]const u8 = null,
const base_id: std.Build.Step.Id = .install_file;
pub const Options = struct {
name: []const u8 = "",
exe: *std.Build.Step.Compile,
zipfile_name: []const u8 = "function.zip",
};
pub fn create(owner: *std.Build, options: Options) *Package {
const name = owner.dupe(options.name);
const step_name = owner.fmt("{s} {s}{s}", .{
"aws lambda",
"package",
name,
});
const package = owner.allocator.create(Package) catch @panic("OOM");
package.* = .{
.step = std.Build.Step.init(.{
.id = base_id,
.name = step_name,
.owner = owner,
.makeFn = make,
}),
.options = options,
};
return package;
}
pub fn shasumFilePath(self: Package) ![]const u8 {
return try std.fmt.allocPrint(
self.step.owner.allocator,
"{s}{s}{s}",
.{ std.fs.path.dirname(self.zipfile_cache_dest.?).?, std.fs.path.sep_str, "sha256sum.txt" },
);
}
pub fn packagedFilePath(self: Package) []const u8 {
return self.step.owner.getInstallPath(.prefix, self.options.zipfile_name);
}
pub fn packagedFileLazyPath(self: Package) std.Build.LazyPath {
return .{ .src_path = .{
.owner = self.step.owner,
.sub_path = self.step.owner.getInstallPath(.prefix, self.options.zipfile_name),
} };
}
fn make(step: *std.Build.Step, node: std.Progress.Node) anyerror!void {
_ = node;
const self: *Package = @fieldParentPtr("step", step);
// get a hash of the bootstrap and whatever other files we put into the zip
// file (because a zip is not really reproducible). That hash becomes the
// cache directory, similar to the way rest of zig works
//
// Otherwise, create the package in our cache indexed by hash, and copy
// our bootstrap, zip things up and install the file into zig-out
const bootstrap = bootstrapLocation(self.*) catch |e| {
if (@errorReturnTrace()) |trace| {
std.debug.dumpStackTrace(trace.*);
}
return step.fail("Could not copy output to bootstrap: {}", .{e});
};
const bootstrap_dirname = std.fs.path.dirname(bootstrap).?;
const zipfile_src = try std.fs.path.join(step.owner.allocator, &[_][]const u8{ bootstrap_dirname, self.options.zipfile_name });
self.zipfile_cache_dest = zipfile_src;
self.zipfile_dest = self.step.owner.getInstallPath(.prefix, self.options.zipfile_name);
if (std.fs.copyFileAbsolute(zipfile_src, self.zipfile_dest.?, .{})) |_| {
// we're good here. The zip file exists in cache and has been copied
step.result_cached = true;
} else |_| {
// error, but this is actually the normal case. We will zip the file
// using system zip and store that in cache with the output file for later
// use
// TODO: For Windows, tar.exe can actually do zip files.
// tar -a -cf function.zip file1 [file2...]
//
// See: https://superuser.com/questions/201371/create-zip-folder-from-the-command-line-windows#comment2725283_898508
var child = std.process.Child.init(&[_][]const u8{
"zip",
"-qj9X",
zipfile_src,
bootstrap,
}, self.step.owner.allocator);
child.stdout_behavior = .Ignore;
child.stdin_behavior = .Ignore; // we'll allow stderr through
switch (try child.spawnAndWait()) {
.Exited => |rc| if (rc != 0) return step.fail("Non-zero exit code {} from zip", .{rc}),
.Signal, .Stopped, .Unknown => return step.fail("Abnormal termination from zip step", .{}),
}
try std.fs.copyFileAbsolute(zipfile_src, self.zipfile_dest.?, .{}); // It better be there now
// One last thing. We want to get a Sha256 sum of the zip file, and
// store it in cache. This will help the deployment process compare
// to what's out in AWS, since revision id is apparently trash for these
// purposes
const zipfile = try std.fs.openFileAbsolute(zipfile_src, .{});
defer zipfile.close();
const zip_bytes = try zipfile.readToEndAlloc(step.owner.allocator, 100 * 1024 * 1024);
var hash: [std.crypto.hash.sha2.Sha256.digest_length]u8 = undefined;
std.crypto.hash.sha2.Sha256.hash(zip_bytes, &hash, .{});
const base64 = std.base64.standard.Encoder;
var encoded: [base64.calcSize(std.crypto.hash.sha2.Sha256.digest_length)]u8 = undefined;
const shaoutput = try std.fs.createFileAbsolute(try self.shasumFilePath(), .{});
defer shaoutput.close();
try shaoutput.writeAll(base64.encode(encoded[0..], hash[0..]));
}
}
fn bootstrapLocation(package: Package) ![]const u8 {
const output = package.step.owner.getInstallPath(.bin, package.options.exe.out_filename);
// We will always copy the output file, mainly because we also need the hash...
// if (std.mem.eql(u8, "bootstrap", package.options.exe.out_filename))
// return output; // easy path
// Not so easy...read the file, get a hash of contents, see if it's in cache
const output_file = try std.fs.openFileAbsolute(output, .{});
defer output_file.close();
const output_bytes = try output_file.readToEndAlloc(package.step.owner.allocator, 100 * 1024 * 1024); // 100MB file
// std.Build.Cache.Hasher
// std.Buidl.Cache.hasher_init
var hasher = std.Build.Cache.HashHelper{}; // We'll reuse the same file hasher from cache
hasher.addBytes(output_bytes);
const hash = std.fmt.bytesToHex(hasher.hasher.finalResult(), .lower);
const dest_path = try package.step.owner.cache_root.join(
package.step.owner.allocator,
&[_][]const u8{ "p", hash[0..], "bootstrap" },
);
const dest_file = std.fs.openFileAbsolute(dest_path, .{}) catch null;
if (dest_file) |d| {
d.close();
return dest_path;
}
const pkg_path = try package.step.owner.cache_root.join(
package.step.owner.allocator,
&[_][]const u8{"p"},
);
// Destination file does not exist. Write the bootstrap (after creating the directory)
std.fs.makeDirAbsolute(pkg_path) catch {};
std.fs.makeDirAbsolute(std.fs.path.dirname(dest_path).?) catch {};
const write_file = try std.fs.createFileAbsolute(dest_path, .{});
defer write_file.close();
try write_file.writeAll(output_bytes);
return dest_path;
}

55
lambdabuild/Region.zig
View File

@ -1,55 +0,0 @@
const std = @import("std");
specified_region: ?[]const u8,
allocator: std.mem.Allocator,
/// internal state, please do not use
_calculated_region: ?[]const u8 = null,
const Region = @This();
pub fn region(self: *Region) ![]const u8 {
if (self.specified_region) |r| return r; // user specified
if (self._calculated_region) |r| return r; // cached
self._calculated_region = try findRegionFromSystem(self.allocator);
return self._calculated_region.?;
}
// AWS_CONFIG_FILE (default is ~/.aws/config
// AWS_DEFAULT_REGION
fn findRegionFromSystem(allocator: std.mem.Allocator) ![]const u8 {
const env_map = try std.process.getEnvMap(allocator);
if (env_map.get("AWS_DEFAULT_REGION")) |r| return r;
const config_file_path = env_map.get("AWS_CONFIG_FILE") orelse
try std.fs.path.join(allocator, &[_][]const u8{
env_map.get("HOME") orelse env_map.get("USERPROFILE").?,
".aws",
"config",
});
const config_file = try std.fs.openFileAbsolute(config_file_path, .{});
defer config_file.close();
const config_bytes = try config_file.readToEndAlloc(allocator, 1024 * 1024);
const profile = env_map.get("AWS_PROFILE") orelse "default";
var line_iterator = std.mem.split(u8, config_bytes, "\n");
var in_profile = false;
while (line_iterator.next()) |line| {
const trimmed = std.mem.trim(u8, line, " \t\r");
if (trimmed.len == 0 or trimmed[0] == '#') continue;
if (!in_profile) {
if (trimmed[0] == '[' and trimmed[trimmed.len - 1] == ']') {
// this is a profile directive!
// std.debug.print("profile: {s}, in file: {s}\n", .{ profile, trimmed[1 .. trimmed.len - 1] });
if (std.mem.eql(u8, profile, trimmed[1 .. trimmed.len - 1])) {
in_profile = true;
}
}
continue; // we're only looking for a profile at this point
}
// look for our region directive
if (trimmed[0] == '[' and trimmed[trimmed.len - 1] == ']')
return error.RegionNotFound; // we've hit another profile without getting our region
if (!std.mem.startsWith(u8, trimmed, "region")) continue;
var equalityiterator = std.mem.split(u8, trimmed, "=");
_ = equalityiterator.next() orelse return error.RegionNotFound;
const raw_val = equalityiterator.next() orelse return error.RegionNotFound;
return try allocator.dupe(u8, std.mem.trimLeft(u8, raw_val, " \t"));
}
return error.RegionNotFound;
}

BIN
lambdabuild/function.zip
View File

Binary file not shown.

500
src/lambda.zig
View File

@ -1,438 +1,96 @@
const std = @import("std");
const builtin = @import("builtin");
const requestz = @import("requestz");
pub const HandlerFn = *const fn (std.mem.Allocator, []const u8) anyerror![]const u8;
const log = std.log.scoped(.lambda);
var client: ?std.http.Client = null;
const prefix = "http://";
const postfix = "/2018-06-01/runtime/invocation";
pub fn deinit() void {
if (client) |*c| c.deinit();
client = null;
}
/// Starts the lambda framework. Handler will be called when an event is processing
/// If an allocator is not provided, an approrpriate allocator will be selected and used
/// This function is intended to loop infinitely. If not used in this manner,
/// make sure to call the deinit() function
pub fn run(allocator: ?std.mem.Allocator, event_handler: HandlerFn) !void { // TODO: remove inferred error set?
const lambda_runtime_uri = std.posix.getenv("AWS_LAMBDA_RUNTIME_API") orelse test_lambda_runtime_uri.?;
// TODO: If this is null, go into single use command line mode
pub fn run(event_handler: fn (*std.mem.Allocator, []const u8) anyerror![]const u8) !void { // TODO: remove inferred error set?
const prefix = "http://";
const postfix = "/2018-06-01/runtime/invocation";
const lambda_runtime_uri = std.os.getenv("AWS_LAMBDA_RUNTIME_API");
var gpa = std.heap.GeneralPurposeAllocator(.{}){};
defer _ = gpa.deinit();
const alloc = allocator orelse gpa.allocator();
const allocator = &gpa.allocator;
const url = try std.fmt.allocPrint(alloc, "{s}{s}{s}/next", .{ prefix, lambda_runtime_uri, postfix });
defer alloc.free(url);
const uri = try std.Uri.parse(url);
const url = try std.fmt.allocPrint(allocator, "{s}{s}{s}/next", .{ prefix, lambda_runtime_uri, postfix });
defer allocator.free(url);
// TODO: Simply adding this line without even using the client is enough
// to cause seg faults!?
// client = client orelse .{ .allocator = alloc };
// so we'll do this instead
if (client != null) return error.MustDeInitBeforeCallingRunAgain;
client = .{ .allocator = alloc };
log.info("tid {d} (lambda): Bootstrap initializing with event url: {s}", .{ std.Thread.getCurrentId(), url });
std.log.notice("Bootstrap initializing with event url: {s}", .{url});
while (lambda_remaining_requests == null or lambda_remaining_requests.? > 0) {
if (lambda_remaining_requests) |*r| {
// we're under test
log.debug("lambda remaining requests: {d}", .{r.*});
r.* -= 1;
}
var req_alloc = std.heap.ArenaAllocator.init(alloc);
while (true) {
var req_alloc = std.heap.ArenaAllocator.init(allocator);
defer req_alloc.deinit();
const req_allocator = req_alloc.allocator();
// Fundamentally we're doing 3 things:
// 1. Get the next event from Lambda (event data and request id)
// 2. Call our handler to get the response
// 3. Post the response back to Lambda
var ev = getEvent(req_allocator, uri) catch |err| {
// Well, at this point all we can do is shout at the void
log.err("Error fetching event details: {}", .{err});
std.posix.exit(1);
// continue;
};
if (ev == null) continue; // this gets logged in getEvent, and without
// a request id, we still can't do anything
// reasonable to report back
const event = ev.?;
defer ev.?.deinit();
const event_response = event_handler(req_allocator, event.event_data) catch |err| {
event.reportError(@errorReturnTrace(), err, lambda_runtime_uri) catch unreachable;
continue;
};
event.postResponse(lambda_runtime_uri, event_response) catch |err| {
event.reportError(@errorReturnTrace(), err, lambda_runtime_uri) catch unreachable;
continue;
};
}
}
const Event = struct {
allocator: std.mem.Allocator,
event_data: []const u8,
request_id: []const u8,
const Self = @This();
pub fn init(allocator: std.mem.Allocator, event_data: []const u8, request_id: []const u8) Self {
return .{
.allocator = allocator,
.event_data = event_data,
.request_id = request_id,
};
}
pub fn deinit(self: *Self) void {
self.allocator.free(self.event_data);
self.allocator.free(self.request_id);
}
fn reportError(
self: Self,
return_trace: ?*std.builtin.StackTrace,
err: anytype,
lambda_runtime_uri: []const u8,
) !void {
// If we fail in this function, we're pretty hosed up
if (return_trace) |rt|
log.err("Caught error: {}. Return Trace: {any}", .{ err, rt })
else
log.err("Caught error: {}. No return trace available", .{err});
const err_url = try std.fmt.allocPrint(
self.allocator,
"{s}{s}{s}/{s}/error",
.{ prefix, lambda_runtime_uri, postfix, self.request_id },
);
defer self.allocator.free(err_url);
const err_uri = try std.Uri.parse(err_url);
const content =
\\{{
\\ "errorMessage": "{s}",
\\ "errorType": "HandlerReturnedError",
\\ "stackTrace": [ "{any}" ]
\\}}
;
const content_fmt = if (return_trace) |rt|
try std.fmt.allocPrint(self.allocator, content, .{ @errorName(err), rt })
else
try std.fmt.allocPrint(self.allocator, content, .{ @errorName(err), "no return trace available" });
defer self.allocator.free(content_fmt);
log.err("Posting to {s}: Data {s}", .{ err_url, content_fmt });
// TODO: There is something up with using a shared client in this way
// so we're taking a perf hit in favor of stability. In a practical
// sense, without making HTTPS connections (lambda environment is
// non-ssl), this shouldn't be a big issue
var cl = std.http.Client{ .allocator = self.allocator };
defer cl.deinit();
const res = cl.fetch(.{
.method = .POST,
.payload = content_fmt,
.location = .{ .uri = err_uri },
.extra_headers = &.{
.{
.name = "Lambda-Runtime-Function-Error-Type",
.value = "HandlerReturned",
},
},
}) catch |post_err| { // Well, at this point all we can do is shout at the void
log.err("Error posting response (start) for request id {s}: {}", .{ self.request_id, post_err });
std.posix.exit(1);
};
// TODO: Determine why this post is not returning
if (res.status != .ok) {
const req_allocator = &req_alloc.allocator;
var client = try requestz.Client.init(req_allocator);
// defer client.deinit();
// Lambda freezes the process at this line of code. During warm start,
// the process will unfreeze and data will be sent in response to client.get
var response = client.get(url, .{}) catch |err| {
std.log.err("Get fail: {}", .{err});
// Documentation says something about "exit immediately". The
// Lambda infrastrucutre restarts, so it's unclear if that's necessary.
// It seems as though a continue should be fine, and slightly faster
log.err("Post fail: {} {s}", .{
@intFromEnum(res.status),
res.status.phrase() orelse "",
});
std.posix.exit(1);
// std.os.exit(1);
continue;
};
defer response.deinit();
var request_id: ?[]const u8 = null;
for (response.headers.items()) |h| {
if (std.mem.indexOf(u8, h.name.value, "Lambda-Runtime-Aws-Request-Id")) |_|
request_id = h.value;
// TODO: XRay uses an environment variable to do its magic. It's our
// responsibility to set this, but no zig-native setenv(3)/putenv(3)
// exists. I would kind of rather not link in libc for this,
// so we'll hold for now and think on this
// if (std.mem.indexOf(u8, h.name.value, "Lambda-Runtime-Trace-Id")) |_|
// std.process.
// std.os.setenv("AWS_LAMBDA_RUNTIME_API");
}
log.err("Error reporting post complete", .{});
}
if (request_id == null) {
// We can't report back an issue because the runtime error reporting endpoint
// uses request id in its path. So the best we can do is log the error and move
// on here.
std.log.err("Could not find request id: skipping request", .{});
continue;
}
const req_id = request_id.?;
fn postResponse(self: Self, lambda_runtime_uri: []const u8, event_response: []const u8) !void {
const response_url = try std.fmt.allocPrint(
self.allocator,
"{s}{s}{s}/{s}/response",
.{ prefix, lambda_runtime_uri, postfix, self.request_id },
);
defer self.allocator.free(response_url);
var cl = std.http.Client{ .allocator = self.allocator };
defer cl.deinit();
// Lambda does different things, depending on the runtime. Go 1.x takes
// any return value but escapes double quotes. Custom runtimes can
// do whatever they want. node I believe wraps as a json object. We're
// going to leave the return value up to the handler, and they can
// use a seperate API for normalization so we're explicit. As a result,
// we can just post event_response completely raw here
const res = try cl.fetch(.{
.method = .POST,
.payload = event_response,
.location = .{ .url = response_url },
});
if (res.status != .ok) return error.UnexpectedStatusFromPostResponse;
}
};
fn getEvent(allocator: std.mem.Allocator, event_data_uri: std.Uri) !?Event {
// TODO: There is something up with using a shared client in this way
// so we're taking a perf hit in favor of stability. In a practical
// sense, without making HTTPS connections (lambda environment is
// non-ssl), this shouldn't be a big issue
var cl = std.http.Client{ .allocator = allocator };
defer cl.deinit();
var response_bytes = std.ArrayList(u8).init(allocator);
defer response_bytes.deinit();
var server_header_buffer: [16 * 1024]u8 = undefined;
// Lambda freezes the process at this line of code. During warm start,
// the process will unfreeze and data will be sent in response to client.get
var res = try cl.fetch(.{
.server_header_buffer = &server_header_buffer,
.location = .{ .uri = event_data_uri },
.response_storage = .{ .dynamic = &response_bytes },
});
if (res.status != .ok) {
// Documentation says something about "exit immediately". The
// Lambda infrastrucutre restarts, so it's unclear if that's necessary.
// It seems as though a continue should be fine, and slightly faster
// std.os.exit(1);
log.err("Lambda server event response returned bad error code: {} {s}", .{
@intFromEnum(res.status),
res.status.phrase() orelse "",
});
return error.EventResponseNotOkResponse;
}
var request_id: ?[]const u8 = null;
var header_it = std.http.HeaderIterator.init(server_header_buffer[0..]);
while (header_it.next()) |h| {
if (std.ascii.eqlIgnoreCase(h.name, "Lambda-Runtime-Aws-Request-Id"))
request_id = h.value;
// TODO: XRay uses an environment variable to do its magic. It's our
// responsibility to set this, but no zig-native setenv(3)/putenv(3)
// exists. I would kind of rather not link in libc for this,
// so we'll hold for now and think on this
// if (std.mem.indexOf(u8, h.name.value, "Lambda-Runtime-Trace-Id")) |_|
// std.process.
// std.os.setenv("AWS_LAMBDA_RUNTIME_API");
}
if (request_id == null) {
// We can't report back an issue because the runtime error reporting endpoint
// uses request id in its path. So the best we can do is log the error and move
// on here.
log.err("Could not find request id: skipping request", .{});
return null;
}
const req_id = request_id.?;
log.debug("got lambda request with id {s}", .{req_id});
return Event.init(
allocator,
try response_bytes.toOwnedSlice(),
try allocator.dupe(u8, req_id),
);
}
////////////////////////////////////////////////////////////////////////
// All code below this line is for testing
////////////////////////////////////////////////////////////////////////
var server_port: ?u16 = null;
var server_remaining_requests: usize = 0;
var lambda_remaining_requests: ?usize = null;
var server_response: []const u8 = "unset";
var server_request_aka_lambda_response: []u8 = "";
var test_lambda_runtime_uri: ?[]u8 = null;
var server_ready = false;
/// This starts a test server. We're not testing the server itself,
/// so the main tests will start this thing up and create an arena around the
/// whole thing so we can just deallocate everything at once at the end,
/// leaks be damned
fn startServer(allocator: std.mem.Allocator) !std.Thread {
return try std.Thread.spawn(
.{},
threadMain,
.{allocator},
);
}
fn threadMain(allocator: std.mem.Allocator) !void {
const address = try std.net.Address.parseIp("127.0.0.1", 0);
var http_server = try address.listen(.{ .reuse_address = true });
server_port = http_server.listen_address.in.getPort();
test_lambda_runtime_uri = try std.fmt.allocPrint(allocator, "127.0.0.1:{d}", .{server_port.?});
log.debug("server listening at {s}", .{test_lambda_runtime_uri.?});
defer test_lambda_runtime_uri = null;
defer server_port = null;
log.info("starting server thread, tid {d}", .{std.Thread.getCurrentId()});
var arena = std.heap.ArenaAllocator.init(allocator);
defer arena.deinit();
const aa = arena.allocator();
// We're in control of all requests/responses, so this flag will tell us
// when it's time to shut down
while (server_remaining_requests > 0) {
server_remaining_requests -= 1;
// defer {
// if (!arena.reset(.{ .retain_capacity = {} })) {
// // reallocation failed, arena is degraded
// log.warn("Arena reset failed and is degraded. Resetting arena", .{});
// arena.deinit();
// arena = std.heap.ArenaAllocator.init(allocator);
// aa = arena.allocator();
// }
// }
processRequest(aa, &http_server) catch |e| {
log.err("Unexpected error processing request: {any}", .{e});
if (@errorReturnTrace()) |trace| {
std.debug.dumpStackTrace(trace.*);
}
const event_response = event_handler(req_allocator, response.body) catch |err| {
// Stack trace will return null if stripped
const return_trace = @errorReturnTrace();
std.log.err("Caught error: {}. Return Trace: {}", .{ err, return_trace });
const err_url = try std.fmt.allocPrint(req_allocator, "{s}{s}/runtime/invocation/{s}/error", .{ prefix, lambda_runtime_uri, req_id });
defer req_allocator.free(err_url);
const content =
\\ {s}
\\ "errorMessage": "{s}",
\\ "errorType": "HandlerReturnedError",
\\ "stackTrace": [ "{}" ]
\\ {s}
;
const content_fmt = try std.fmt.allocPrint(req_allocator, content, .{ "{", @errorName(err), return_trace, "}" });
defer req_allocator.free(content_fmt);
std.log.err("Posting to {s}: Data {s}", .{ err_url, content_fmt });
var headers = .{.{ "Lambda-Runtime-Function-Error-Type", "HandlerReturned" }};
// TODO: Determine why this post is not returning
var err_resp = client.post(err_url, .{
.content = content_fmt,
.headers = headers,
}) catch |post_err| { // Well, at this point all we can do is shout at the void
std.log.err("Error posting response for request id {s}: {}", .{ req_id, post_err });
std.os.exit(0);
continue;
};
std.log.err("Post complete", .{});
defer err_resp.deinit();
continue;
};
const response_url = try std.fmt.allocPrint(req_allocator, "{s}{s}{s}/{s}/response", .{ prefix, lambda_runtime_uri, postfix, req_id });
// defer req_allocator.free(response_url);
var resp_resp = client.post(response_url, .{ .content = event_response }) catch |err| {
// TODO: report error
std.log.err("Error posting response for request id {s}: {}", .{ req_id, err });
continue;
};
defer resp_resp.deinit();
}
}
fn processRequest(allocator: std.mem.Allocator, server: *std.net.Server) !void {
server_ready = true;
errdefer server_ready = false;
log.debug(
"tid {d} (server): server waiting to accept. requests remaining: {d}",
.{ std.Thread.getCurrentId(), server_remaining_requests + 1 },
);
var connection = try server.accept();
defer connection.stream.close();
server_ready = false;
var read_buffer: [1024 * 16]u8 = undefined;
var http_server = std.http.Server.init(connection, &read_buffer);
if (http_server.state == .ready) {
var request = http_server.receiveHead() catch |err| switch (err) {
error.HttpConnectionClosing => return,
else => {
std.log.err("closing http connection: {s}", .{@errorName(err)});
std.log.debug("Error occurred from this request: \n{s}", .{read_buffer[0..http_server.read_buffer_len]});
return;
},
};
server_request_aka_lambda_response = try (try request.reader()).readAllAlloc(allocator, std.math.maxInt(usize));
var respond_options = std.http.Server.Request.RespondOptions{};
const response_bytes = serve(allocator, request, &respond_options) catch |e| brk: {
respond_options.status = .internal_server_error;
// TODO: more about this particular request
log.err("Unexpected error from executor processing request: {any}", .{e});
if (@errorReturnTrace()) |trace| {
std.debug.dumpStackTrace(trace.*);
}
break :brk "Unexpected error generating request to lambda";
};
try request.respond(response_bytes, respond_options);
log.debug(
"tid {d} (server): sent response: {s}",
.{ std.Thread.getCurrentId(), response_bytes },
);
}
}
fn serve(allocator: std.mem.Allocator, request: std.http.Server.Request, respond_options: *std.http.Server.Request.RespondOptions) ![]const u8 {
_ = allocator;
_ = request;
respond_options.extra_headers = &.{
.{ .name = "Lambda-Runtime-Aws-Request-Id", .value = "69" },
};
return server_response;
}
fn handler(allocator: std.mem.Allocator, event_data: []const u8) ![]const u8 {
_ = allocator;
return event_data;
}
pub fn test_lambda_request(allocator: std.mem.Allocator, request: []const u8, request_count: usize, handler_fn: HandlerFn) ![]u8 {
var arena = std.heap.ArenaAllocator.init(allocator);
defer arena.deinit();
const aa = arena.allocator();
// Setup our server to run, and set the response for the server to the
// request. There is a cognitive disconnect here between mental model and
// physical model.
//
// Mental model:
//
// Lambda request -> λ -> Lambda response
//
// Physcial Model:
//
// 1. λ requests instructions from server
// 2. server provides "Lambda request"
// 3. λ posts response back to server
//
// So here we are setting up our server, then our lambda request loop,
// but it all needs to be in seperate threads so we can control startup
// and shut down. Both server and Lambda are set up to watch global variable
// booleans to know when to shut down. This function is designed for a
// single request/response pair only
lambda_remaining_requests = request_count;
server_remaining_requests = lambda_remaining_requests.? * 2; // Lambda functions
// fetch from the server,
// then post back. Always
// 2, no more, no less
server_response = request; // set our instructions to lambda, which in our
// physical model above, is the server response
defer server_response = "unset"; // set it back so we don't get confused later
// when subsequent tests fail
const server_thread = try startServer(aa); // start the server, get it ready
while (!server_ready)
std.time.sleep(100);
log.debug("tid {d} (main): server reports ready", .{std.Thread.getCurrentId()});
// we aren't testing the server,
// so we'll use the arena allocator
defer server_thread.join(); // we'll be shutting everything down before we exit
// Now we need to start the lambda framework
try run(allocator, handler_fn); // We want our function under test to report leaks
return try allocator.dupe(u8, server_request_aka_lambda_response);
}
test "basic request" {
// std.testing.log_level = .debug;
const allocator = std.testing.allocator;
const request =
\\{"foo": "bar", "baz": "qux"}
;
// This is what's actually coming back. Is this right?
const expected_response =
\\{"foo": "bar", "baz": "qux"}
;
const lambda_response = try test_lambda_request(allocator, request, 1, handler);
defer deinit();
defer allocator.free(lambda_response);
try std.testing.expectEqualStrings(expected_response, lambda_response);
}
test "several requests do not fail" {
// std.testing.log_level = .debug;
const allocator = std.testing.allocator;
const request =
\\{"foo": "bar", "baz": "qux"}
;
// This is what's actually coming back. Is this right?
const expected_response =
\\{"foo": "bar", "baz": "qux"}
;
const lambda_response = try test_lambda_request(allocator, request, 5, handler);
defer deinit();
defer allocator.free(lambda_response);
try std.testing.expectEqualStrings(expected_response, lambda_response);
}

4
src/sample-main.zig → src/main.zig
View File

@ -2,10 +2,10 @@ const std = @import("std");
const lambda = @import("lambda.zig");
pub fn main() anyerror!void {
try lambda.run(null, handler);
try lambda.run(handler);
}
fn handler(allocator: std.mem.Allocator, event_data: []const u8) ![]const u8 {
fn handler(allocator: *std.mem.Allocator, event_data: []const u8) ![]const u8 {
_ = allocator;
return event_data;
}