An SMTP client and server library for Zig implementing RFC 5321.
0

Configure Feed

Select the types of activity you want to include in your feed.

zig-smtp / build.zig
11 kB 224 lines
1// SPDX-FileCopyrightText: © 2026 Jeffrey C. Ollie <jeff@ocjtech.us> 2// SPDX-License-Identifier: MIT 3 4const std = @import("std"); 5 6// Although this function looks imperative, it does not perform the build 7// directly and instead it mutates the build graph (`b`) that will be then 8// executed by an external runner. The functions in `std.Build` implement a DSL 9// for defining build steps and express dependencies between them, allowing the 10// build runner to parallelize the build automatically (and the cache system to 11// know when a step doesn't need to be re-run). 12pub fn build(b: *std.Build) void { 13 // Standard target options allow the person running `zig build` to choose 14 // what target to build for. Here we do not override the defaults, which 15 // means any target is allowed, and the default is native. Other options 16 // for restricting supported target set are available. 17 const target = b.standardTargetOptions(.{}); 18 // Standard optimization options allow the person running `zig build` to select 19 // between Debug, ReleaseSafe, ReleaseFast, and ReleaseSmall. Here we do not 20 // set a preferred release mode, allowing the user to decide how to optimize. 21 const optimize = b.standardOptimizeOption(.{}); 22 // It's also possible to define more custom flags to toggle optional features 23 // of this build script using `b.option()`. All defined flags (including 24 // target and optimize options) will be listed when running `zig build --help` 25 // in this directory. 26 27 // This creates a module, which represents a collection of source files alongside 28 // some compilation options, such as optimization mode and linked system libraries. 29 // Zig modules are the preferred way of making Zig code available to consumers. 30 // addModule defines a module that we intend to make available for importing 31 // to our consumers. We must give it a name because a Zig package can expose 32 // multiple modules and consumers will need to be able to specify which 33 // module they want to access. 34 const tls_dep = b.dependency("tls", .{ 35 .target = target, 36 .optimize = optimize, 37 }); 38 39 const mod = b.addModule("zsmtp", .{ 40 // The root source file is the "entry point" of this module. Users of 41 // this module will only be able to access public declarations contained 42 // in this file, which means that if you have declarations that you 43 // intend to expose to consumers that were defined in other files part 44 // of this module, you will have to make sure to re-export them from 45 // the root file. 46 .root_source_file = b.path("src/root.zig"), 47 // Later on we'll use this module as the root module of a test executable 48 // which requires us to specify a target. 49 .target = target, 50 .imports = &.{ 51 .{ .name = "tls", .module = tls_dep.module("tls") }, 52 }, 53 }); 54 55 // Here we define an executable. An executable needs to have a root module 56 // which needs to expose a `main` function. While we could add a main function 57 // to the module defined above, it's sometimes preferable to split business 58 // logic and the CLI into two separate modules. 59 // 60 // If your goal is to create a Zig library for others to use, consider if 61 // it might benefit from also exposing a CLI tool. A parser library for a 62 // data serialization format could also bundle a CLI syntax checker, for example. 63 // 64 // If instead your goal is to create an executable, consider if users might 65 // be interested in also being able to embed the core functionality of your 66 // program in their own executable in order to avoid the overhead involved in 67 // subprocessing your CLI tool. 68 // 69 // If neither case applies to you, feel free to delete the declaration you 70 // don't need and to put everything under a single module. 71 const exe = b.addExecutable(.{ 72 .name = "zsmtp", 73 .root_module = b.createModule(.{ 74 // b.createModule defines a new module just like b.addModule but, 75 // unlike b.addModule, it does not expose the module to consumers of 76 // this package, which is why in this case we don't have to give it a name. 77 .root_source_file = b.path("src/main.zig"), 78 // Target and optimization levels must be explicitly wired in when 79 // defining an executable or library (in the root module), and you 80 // can also hardcode a specific target for an executable or library 81 // definition if desireable (e.g. firmware for embedded devices). 82 .target = target, 83 .optimize = optimize, 84 // List of modules available for import in source files part of the 85 // root module. 86 .imports = &.{ 87 // Here "zsmtp" is the name you will use in your source code to 88 // import this module (e.g. `@import("zsmtp")`). The name is 89 // repeated because you are allowed to rename your imports, which 90 // can be extremely useful in case of collisions (which can happen 91 // importing modules from different packages). 92 .{ .name = "zsmtp", .module = mod }, 93 }, 94 }), 95 }); 96 97 // This declares intent for the executable to be installed into the 98 // install prefix when running `zig build` (i.e. when executing the default 99 // step). By default the install prefix is `zig-out/` but can be overridden 100 // by passing `--prefix` or `-p`. 101 b.installArtifact(exe); 102 103 // This creates a top level step. Top level steps have a name and can be 104 // invoked by name when running `zig build` (e.g. `zig build run`). 105 // This will evaluate the `run` step rather than the default step. 106 // For a top level step to actually do something, it must depend on other 107 // steps (e.g. a Run step, as we will see in a moment). 108 const run_step = b.step("run", "Run the app"); 109 110 // This creates a RunArtifact step in the build graph. A RunArtifact step 111 // invokes an executable compiled by Zig. Steps will only be executed by the 112 // runner if invoked directly by the user (in the case of top level steps) 113 // or if another step depends on it, so it's up to you to define when and 114 // how this Run step will be executed. In our case we want to run it when 115 // the user runs `zig build run`, so we create a dependency link. 116 const run_cmd = b.addRunArtifact(exe); 117 run_step.dependOn(&run_cmd.step); 118 119 // By making the run step depend on the default step, it will be run from the 120 // installation directory rather than directly from within the cache directory. 121 run_cmd.step.dependOn(b.getInstallStep()); 122 123 // This allows the user to pass arguments to the application in the build 124 // command itself, like this: `zig build run -- arg1 arg2 etc` 125 if (b.args) |args| { 126 run_cmd.addArgs(args); 127 } 128 129 const lib = b.addLibrary(.{ 130 .name = "zsmtp", 131 .root_module = mod, 132 }); 133 134 const install_docs = b.addInstallDirectory(.{ 135 .source_dir = lib.getEmittedDocs(), 136 .install_dir = .prefix, 137 .install_subdir = "docs", 138 }); 139 140 const docs_step = b.step("docs", "Build the API docs"); 141 docs_step.dependOn(&install_docs.step); 142 143 // The is_email address corpus test embeds its XML test files from the 144 // lazy isemail dependency, fetched only on demand: 145 // zig build test -Disemail-corpus 146 const isemail_corpus = b.option(bool, "isemail-corpus", "Fetch Dominic Sayers' is_email suite and run the address corpus test") orelse false; 147 var isemail_available = false; 148 if (isemail_corpus) { 149 if (b.lazyDependency("isemail", .{})) |isemail_dep| { 150 mod.addAnonymousImport("isemail_tests_xml", .{ 151 .root_source_file = isemail_dep.path("test/tests.xml"), 152 }); 153 mod.addAnonymousImport("isemail_tests_original_xml", .{ 154 .root_source_file = isemail_dep.path("test/tests-original.xml"), 155 }); 156 isemail_available = true; 157 } 158 } 159 const test_options = b.addOptions(); 160 test_options.addOption(bool, "isemail_corpus", isemail_available); 161 mod.addOptions("build_options", test_options); 162 163 // Exim's scriptable SMTP test client, handy for driving the zsmtp 164 // server through raw protocol dialogues with reply expectations (see 165 // test/protocol-torture.script). Guarded by an option so the lazy exim 166 // dependency is only fetched on demand: 167 // zig build -Dexim-client && ./zig-out/bin/exim-client <host> <port> 168 if (b.option(bool, "exim-client", "Build exim's scriptable SMTP test client (fetches the exim source)") orelse false) { 169 if (b.lazyDependency("exim", .{})) |exim_dep| { 170 const exim_client = b.addExecutable(.{ 171 .name = "exim-client", 172 .root_module = b.createModule(.{ 173 .target = target, 174 .optimize = optimize, 175 .link_libc = true, 176 }), 177 }); 178 exim_client.root_module.addCSourceFile(.{ 179 .file = exim_dep.path("test/src/client.c"), 180 .flags = &.{"-w"}, 181 }); 182 b.installArtifact(exim_client); 183 } 184 } 185 186 // Creates an executable that will run `test` blocks from the provided module. 187 // Here `mod` needs to define a target, which is why earlier we made sure to 188 // set the releative field. 189 const mod_tests = b.addTest(.{ 190 .root_module = mod, 191 }); 192 193 // A run step that will run the test executable. 194 const run_mod_tests = b.addRunArtifact(mod_tests); 195 196 // Creates an executable that will run `test` blocks from the executable's 197 // root module. Note that test executables only test one module at a time, 198 // hence why we have to create two separate ones. 199 const exe_tests = b.addTest(.{ 200 .root_module = exe.root_module, 201 }); 202 203 // A run step that will run the second test executable. 204 const run_exe_tests = b.addRunArtifact(exe_tests); 205 206 // A top level step for running all tests. dependOn can be called multiple 207 // times and since the two run steps do not depend on one another, this will 208 // make the two of them run in parallel. 209 const test_step = b.step("test", "Run tests"); 210 test_step.dependOn(&run_mod_tests.step); 211 test_step.dependOn(&run_exe_tests.step); 212 213 // Just like flags, top level steps are also listed in the `--help` menu. 214 // 215 // The Zig build system is entirely implemented in userland, which means 216 // that it cannot hook into private compiler APIs. All compilation work 217 // orchestrated by the build system will result in other Zig compiler 218 // subcommands being invoked with the right flags defined. You can observe 219 // these invocations when one fails (or you pass a flag to increase 220 // verbosity) to validate assumptions and diagnose problems. 221 // 222 // Lastly, the Zig build system is relatively simple and self-contained, 223 // and reading its source code will allow you to master it. 224}