fix(toolchain): store multi-token dialect flags as tokens, not a string to split

macOS CI segfaulted on every build.mcpp test. The crash was inside
`contract_env` — a function this branch never touched — corrupting a local
vector before the compile even started, and it reproduced only under
clang/libc++, never under GCC.

Bisecting with a clean target and dep cache each round (a first attempt was
worthless: switching mcpp.toml's toolchain without clearing the dependency
cache links a libstdc++-built mcpplibs.cmdline into a libc++ mcpp, which
crashes for an entirely unrelated reason) narrowed it to a single addition:
the `split_ws` helper. Not a call to it — its mere PRESENCE in the anonymous
namespace. Deleting it fixed the crash; a trivial unused function in the same
spot did not cause one. That is a clang codegen problem, not a logic error,
and no amount of reading the diff would have found it.

The fix is also the better design. `split_ws` only existed because the
dialect table stored "-x c++" and "/nologo /EHsc /utf-8" as ninja-command
strings while the build.mcpp path needs argv tokens — so the token boundary
was being re-derived at the call site. The table now carries both forms, the
argv one as a span over a static array, and the spelling stays in one row.

Verified under BOTH toolchains from a clean target and cache: clang 22.1.8
(e2e 89 / 92 / 179 / 181 all pass, previously all segfaults) and GCC 16.1.0
(unit 49/49).

Author: Sunrisepeak

src/build/build_program.cppm

@@ -104,22 +104,6 @@ struct Directives {
     std::vector<std::string> rerunEnv;      // declared env-var inputs
 };
 
-// Split a whitespace-separated flag string into argv tokens. The dialect
-// table stores some entries as multi-token strings ("-x c++",
-// "/nologo /EHsc /utf-8") because their other consumer is a ninja command
-// line, where a single string is what's wanted; an argv vector is not.
-std::vector<std::string> split_ws(std::string_view s) {
-    std::vector<std::string> out;
-    std::size_t i = 0;
-    while (i < s.size()) {
-        while (i < s.size() && (s[i] == ' ' || s[i] == '\t')) ++i;
-        std::size_t b = i;
-        while (i < s.size() && s[i] != ' ' && s[i] != '\t') ++i;
-        if (i > b) out.emplace_back(s.substr(b, i - b));
-    }
-    return out;
-}
-
 std::string trim(std::string_view s) {
     std::size_t b = 0, e = s.size();
     while (b < e && (s[b] == ' ' || s[b] == '\t' || s[b] == '\r')) ++b;
@@ -859,7 +843,7 @@ std::expected<void, std::string> run_build_program(
     if (msvcHost) {
         // /nologo /EHsc /utf-8 — cl.exe needs these to behave like the other
         // two drivers do by default (quiet, exceptions on, UTF-8 sources).
-        for (auto& f : split_ws(dial.alwaysFlags)) compileArgv.push_back(f);
+        for (auto f : dial.alwaysFlagsArgv) compileArgv.emplace_back(f);
     }
     compileArgv.push_back(std_flag);
     // No optimization: this program runs once per build and its compile time
@@ -871,7 +855,7 @@ std::expected<void, std::string> run_build_program(
     for (auto& sf : stdFlags)    compileArgv.push_back(sf);
     // The `.mcpp` extension is unknown to every driver, so without this the
     // file is handed to the linker as a linker script.
-    for (auto& f : split_ws(dial.forceCxxLang)) compileArgv.push_back(f);
+    for (auto f : dial.forceCxxLangArgv) compileArgv.emplace_back(f);
     compileArgv.push_back(src.string());
     if (usesModule || !stdObjects.empty()) {
         // Link the module objects (GNU: reset the input language first so the

src/toolchain/dialect.cppm

@@ -32,6 +32,9 @@ struct CommandDialect {
     std::string_view optPrefix;        // "-O"     | "/O"
     std::string_view debugFlags;       // "-g"     | "/Zi /FS"
     std::string_view alwaysFlags;      // ""       | "/nologo /EHsc /utf-8"
+    // Same rationale as forceCxxLangArgv: the argv consumer gets tokens, not
+    // a string it has to split. Both spans point at static arrays below.
+    std::span<const std::string_view> alwaysFlagsArgv;
 
     // Link and language-selection spellings.
     //
@@ -42,7 +45,15 @@ struct CommandDialect {
     std::string_view libSearchPrefix;  // "-L"     | "/LIBPATH:"
     // The `.mcpp` extension is unknown to every compiler driver, so the
     // language has to be forced or the driver hands the file to the linker.
+    //
+    // Two forms of the same thing, because the two consumers need different
+    // shapes and neither should re-derive the other's: a ninja command line
+    // wants one string, an argv vector wants tokens. Storing both keeps the
+    // spelling in one row — splitting the string at the call site would put
+    // the token boundary in a second place, and cost a helper this file is
+    // better off without (see the note on `alwaysFlagsArgv`).
     std::string_view forceCxxLang;     // "-x c++" | "/TP"
+    std::span<const std::string_view> forceCxxLangArgv;
     // Static CRT / runtime. On MSVC this is a compile-time CRT model, not a
     // link mode — there is no /MT equivalent of `-static` for the whole image.
     std::string_view staticRuntime;    // "-static"| "/MT"
@@ -95,6 +106,12 @@ namespace mcpp::toolchain {
 
 namespace {
 
+// Token forms of the multi-token rows. Static arrays so the spans above are
+// constexpr-initializable and no consumer has to split a string at runtime.
+constexpr std::string_view kGnuForceCxxArgv[]  = {"-x", "c++"};
+constexpr std::string_view kMsvcForceCxxArgv[] = {"/TP"};
+constexpr std::string_view kMsvcAlwaysArgv[]   = {"/nologo", "/EHsc", "/utf-8"};
+
 constexpr CommandDialect kGnuDialect{
     .id              = "gnu",
     .includePrefix   = "-I",
@@ -105,9 +122,11 @@ constexpr CommandDialect kGnuDialect{
     .optPrefix       = "-O",
     .debugFlags      = "-g",
     .alwaysFlags     = "",
+    .alwaysFlagsArgv = {},
     .libFlag         = "-l{}",
     .libSearchPrefix = "-L",
     .forceCxxLang    = "-x c++",
+    .forceCxxLangArgv = kGnuForceCxxArgv,
     .staticRuntime   = "-static",
     .outputExePrefix = "-o ",
     .objExt          = ".o",
@@ -130,9 +149,11 @@ constexpr CommandDialect kMsvcDialect{
     .optPrefix       = "/O",
     .debugFlags      = "/Zi /FS",
     .alwaysFlags     = "/nologo /EHsc /utf-8",
+    .alwaysFlagsArgv = kMsvcAlwaysArgv,
     .libFlag         = "{}.lib",
     .libSearchPrefix = "/LIBPATH:",
     .forceCxxLang    = "/TP",
+    .forceCxxLangArgv = kMsvcForceCxxArgv,
     .staticRuntime   = "/MT",
     .outputExePrefix = "/Fe:",
     .objExt          = ".obj",