From dbee23c564542fbe2ac8e22e86d172aaa4563323 Mon Sep 17 00:00:00 2001
From: Jorijn van der Graaf <jorijnvdgraaf@catcrafts.net>
Date: Mon, 8 Jun 2026 19:28:20 +0200
Subject: [PATCH] lto

---
 implementations/Crafter.Build-Clang.cpp | 28 ++++++++++++++++++++++---
 1 file changed, 25 insertions(+), 3 deletions(-)

diff --git a/implementations/Crafter.Build-Clang.cpp b/implementations/Crafter.Build-Clang.cpp
index ce3ea4e..7dec117 100644
--- a/implementations/Crafter.Build-Clang.cpp
+++ b/implementations/Crafter.Build-Clang.cpp
@@ -635,6 +635,20 @@ BuildResult Crafter::Build(Configuration& config, std::unordered_map<fs::path, s
 
     std::string cmakeBuildType;
 
+    // Behaviour-neutral release performance for the binaries we emit: ThinLTO
+    // (cross-TU inlining) plus dead-section GC and safe identical-code folding.
+    // Default-on in Release, never Debug (keeps builds fast and debuggable).
+    // Excluded for wasm32 — its -mllvm/sjlj codegen flags don't compose with
+    // LTO here — and for nvcc .cu objects below (they can't emit LLVM bitcode,
+    // so they link in as ordinary objects alongside the bitcode ones). ThinLTO
+    // rather than monolithic -flto so link time and memory scale with arbitrary
+    // user project sizes. Compile-side flags (-flto, -ffunction/-fdata-sections)
+    // ride on `command`, which is reused as the link base; the link-only -Wl
+    // flags go on linkExtras so they don't warn on each -c compile.
+    const bool useLto = !config.debug && !isWasm;
+    const std::string ltoCompileFlags = useLto ? " -flto=thin -ffunction-sections -fdata-sections" : "";
+    const std::string ltoLinkFlags = useLto ? " -flto=thin -Wl,--gc-sections -Wl,--icf=safe" : "";
+
     if(config.debug) {
         cmakeBuildType = "Debug";
         command += " -g -D CRAFTER_BUILD_CONFIGURATION_DEBUG";
@@ -642,6 +656,7 @@ BuildResult Crafter::Build(Configuration& config, std::unordered_map<fs::path, s
         cmakeBuildType = "Release";
         command += " -O3";
     }
+    command += ltoCompileFlags;
 
     // Same target-aware setup as the C++ compile path (line 459-): wasm32
     // rejects -march, silently ignores -mtune, and needs --sysroot to find
@@ -664,11 +679,11 @@ BuildResult Crafter::Build(Configuration& config, std::unordered_map<fs::path, s
         const std::string objPath = (buildDir / cFile.filename()).string() + "_source.o";
         const std::string srcPath = cFile.string() + ".c";
         if (!fs::exists(objPath) || (fs::exists(srcPath) && fs::last_write_time(srcPath) > fs::last_write_time(objPath))) {
-            threads.emplace_back([&cFile, &buildDir, &buildError, &buildCancelled, &config, &includeFlags, &defineFlags, &userFlags, &cArchFlags]() {
+            threads.emplace_back([&cFile, &buildDir, &buildError, &buildCancelled, &config, &includeFlags, &defineFlags, &userFlags, &cArchFlags, &ltoCompileFlags]() {
                 Progress::Task task(std::format("Compiling {}.c", cFile.filename().string()));
                 if (buildCancelled.load(std::memory_order_relaxed)) return;
 
-                std::string result = RunCommand(std::format("clang {}.c --target={}{} -O3 -c{}{}{} -o {}_source.o", cFile.string(), config.target, cArchFlags, includeFlags, defineFlags, userFlags, (buildDir / cFile.filename()).string()));
+                std::string result = RunCommand(std::format("clang {}.c --target={}{} -O3{} -c{}{}{} -o {}_source.o", cFile.string(), config.target, cArchFlags, ltoCompileFlags, includeFlags, defineFlags, userFlags, (buildDir / cFile.filename()).string()));
                 if (result.empty()) return;
 
                 bool expected = false;
@@ -876,6 +891,10 @@ BuildResult Crafter::Build(Configuration& config, std::unordered_map<fs::path, s
     for(const std::string& flag : buildResult.libs) {
         linkExtras += " " + flag;
     }
+    // Link-only LTO/section flags (empty in Debug and for wasm). Every exe/lib
+    // link below ends with linkExtras, so this reaches them all; the static-lib
+    // archive path uses ar instead and is handled separately.
+    linkExtras += ltoLinkFlags;
     // mingw uses libstdc++; C++26 std::print/format extras live in libstdc++exp.
     // libstdc++ on mingw uses winpthreads for std::atomic_wait /
     // counting_semaphore / stop_token, so -lpthread is required as soon as
@@ -1058,7 +1077,10 @@ BuildResult Crafter::Build(Configuration& config, std::unordered_map<fs::path, s
             #endif
         } else if(config.type == ConfigurationType::LibraryStatic) {
             #ifdef CRAFTER_BUILD_CONFIGURATION_TARGET_x86_64_pc_linux_gnu
-            buildResult.result = RunCommand(std::format("ar rcs {}.a {}", (outputDir/fs::path(std::string("lib")+config.outputName)).string(), files));
+            // ThinLTO emits LLVM bitcode objects; plain `ar` writes an archive
+            // index that omits their symbols, so a consumer's lld link can't
+            // pull the needed members. llvm-ar writes a bitcode-aware index.
+            buildResult.result = RunCommand(std::format("{} rcs {}.a {}", useLto ? "llvm-ar" : "ar", (outputDir/fs::path(std::string("lib")+config.outputName)).string(), files));
             #endif
 
             #if defined(CRAFTER_BUILD_CONFIGURATION_TARGET_x86_64_pc_windows_msvc) || defined(CRAFTER_BUILD_CONFIGURATION_TARGET_x86_64_w64_mingw32)