Catcrafts/Crafter.Build
Watch
2
0
Fork
You've already forked Crafter.Build
 0
Code Issues Pull requests Projects Releases 1 Packages Wiki Activity Actions

feat(lint): fixed-width-types keeps widths that a foreign API chose

Browse source 
The rewrite itself stays token-shaped — it edits type SPELLINGS, which an AST
discards — but what it must not touch now comes from the AST.

The old exemption was per LINE: `int main`, `argc`, `argv`, `extern "`. Being a
transform, a missed exemption here does not over-report, it emits code that no
longer matches the API being called, so this is the rule where guessing from
substrings mattered most. And being per-line, it also disabled the rule for
anything sharing a line with one of those words.

Now a declaration with C language linkage, or one whose initialiser binds to an
entity declared outside the project, contributes a protected byte range and
keeps its spelling. That answers the case directly: a function declared in
somebody else's header taking `unsigned int` keeps `unsigned int`, and a local
initialised from strtoul keeps `unsigned long`, because of where those are
declared rather than because of what the line says.

main is protected from the start of its declaration to the opening brace of its
body, not for its whole extent. Its signature is fixed by the language; its body
is ordinary code. Three findings on this repository came out of that
distinction, all correct:

    for (int i = 1; i < argc; ++i)   ->   for (std::int32_t i = 1; ...)

skipped before only because `argc` appeared on the line, plus Crafter::Run's own
`int argc` and return type, which are ours rather than the language's.

All three AST rules now share one interop test instead of carrying a denylist
each, and it is the same test: whose header dictates this spelling.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Jorijn van der Graaf 2026-07-30 23:59:46 +02:00
commit 6438cb9ebb
4 changed files with 68 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

4
implementations/Crafter.Build-Clang.cpp
View file

@ -1638,7 +1638,7 @@ Exit status:
}
// lint-disable-next-line no-char-pointer
int Crafter::Run(int argc, char** argv) {
std::int32_t Crafter::Run(std::int32_t argc, char** argv) {
try {
std::string_view argv0 = argc > 0 ? argv[0] : "crafter-build";
fs::path projectFile = "./project.cpp";
@ -1654,7 +1654,7 @@ int Crafter::Run(int argc, char** argv) {
RunLintOptions lintOpts;
Progress::Verbosity verbosity = Progress::Verbosity::Default;
for (int i = 1; i < argc; ++i) {
for (std::int32_t i = 1; i < argc; ++i) {
std::string_view arg = argv[i];
if (arg == "-h" || arg == "--help" || (!runTests && !runLint && !runFormat && arg == "help")) {
PrintHelp(argv0);

2
interfaces/Crafter.Build-Clang.cppm
View file

@ -554,7 +554,7 @@ export namespace Crafter {
// Takes main's argv verbatim so the CLI entry point is a one-liner;
// the shape is inherited from the language, not chosen here.
// lint-disable-next-line no-char-pointer
CRAFTER_API int Run(int argc, char** argv);
CRAFTER_API std::int32_t Run(std::int32_t argc, char** argv);
// Delete the bin/ and build/ trees beside `projectFile`, returning the paths
// that existed and were removed. Backs `crafter-build clean`.

40
lint-rules.h
View file

@ -487,11 +487,40 @@ inline void AddProjectLintRules(Crafter::Configuration& cfg) {
// short/int/long → fixed-width types. Lines mentioning main/argc/argv or
// extern "C" keep their C-conventional ints.
cfg.AddLintRule("fixed-width-types", [](LintContext& ctx) {
cfg.AddAstLintRule("fixed-width-types", [](LintContext& ctx) {
if (!IsCppFile(ctx)) return;
const std::string& code = ctx.CommentStripped();
struct Rep { std::size_t pos; std::size_t len; std::string to; };
std::vector<Rep> reps;
// Byte ranges whose integer spelling is not ours to change. A wrong
// rewrite here does not merely over-report — it produces code that no
// longer matches the API it is calling — so this is derived from the
// AST rather than from substrings on the line.
//
// Replaces the old per-LINE textual exemption (`int main`, `argc`,
// `argv`, `extern "`), which both missed cases and disabled the rule
// for everything else sharing a line with one of those words.
std::vector<std::pair<std::size_t, std::size_t>> protectedRanges;
for (const Crafter::LintDecl& decl : ctx.Decls()) {
if (decl.isExternC || decl.isForeignApi) {
protectedRanges.emplace_back(decl.begin, decl.end);
continue;
}
// main's signature is fixed by the language. Only the signature:
// its body is ordinary code, and the declaration's extent covers
// the whole function.
if (decl.kind == Crafter::LintDeclKind::Function && decl.name == "main") {
std::size_t bodyBrace = ctx.content.find('{', decl.begin);
protectedRanges.emplace_back(decl.begin, bodyBrace == std::string::npos ? decl.end : bodyBrace);
}
}
auto isProtected = [&protectedRanges](std::size_t offset) {
return std::any_of(protectedRanges.begin(), protectedRanges.end(),
[offset](const std::pair<std::size_t, std::size_t>& range) {
return offset >= range.first && offset < range.second;
});
};
// Builtin integer type specifiers combine in any order (`unsigned
// long`, `long unsigned int`, ...), so match whole RUNS of these
// keywords and classify the run, rather than the words one by one.
@ -503,13 +532,8 @@ inline void AddProjectLintRules(Crafter::Configuration& cfg) {
std::size_t lineEnd = code.find('\n', lineStart);
if (lineEnd == std::string::npos) lineEnd = code.size();
std::string_view line(code.data() + lineStart, lineEnd - lineStart);
// `int main` / argc / argv keep their C-conventional type; so do
// extern "C" prototypes (the literal body is blanked in the
// stripped text, so match `extern "`).
bool exempt = line.contains("int main") || line.contains("argc") || line.contains("argv")
|| line.contains("extern \"");
std::size_t pos = 0;
while (!exempt && pos < line.size()) {
while (pos < line.size()) {
if (!IsWordChar(line[pos])) { ++pos; continue; }
std::size_t wordEnd = pos;
while (wordEnd < line.size() && IsWordChar(line[wordEnd])) ++wordEnd;
@ -546,6 +570,8 @@ inline void AddProjectLintRules(Crafter::Configuration& cfg) {
if (hasChar && !hasUnsigned && !hasSigned) continue;
if (nextWord == "double") continue;
if (isProtected(lineStart + runBegin)) continue;
std::string_view width = hasChar ? "8" : hasShort ? "16" : hasLong ? "64" : "32";
reps.push_back({lineStart + runBegin, runEnd - runBegin,
std::format("std::{}int{}_t", hasUnsigned ? "u" : "", width)});

32
tests/HouseRules/main.cpp
View file

@ -363,6 +363,38 @@ int main() {
Check(!HasFinding(r.summary, "'Holder'"), "naming: PascalCase type passes");
}
// fixed-width-types must not rewrite an integer whose width somebody else's
// header chose — the rewrite would leave code that no longer matches the API
// it calls. Derived from the AST, so the exemption is per-declaration rather
// than the old per-line textual one, which both missed cases and disabled
// the rule for anything else sharing a line with `argc`/`extern "`.
{
RuleRun r = RunRule("#include <cstdlib>\n"
"extern \"C\" unsigned long CApiCall(unsigned int flags);\n"
"long OurOwnApi(int value) { return value; }\n"
"void F() {\n"
" unsigned long fromLibc = std::strtoul(\"1\", nullptr, 10);\n"
" unsigned ours = 1;\n"
" (void)fromLibc; (void)ours;\n"
"}\n"
"int main(int argc, char** argv) {\n"
" for (int i = 0; i < argc; ++i) { (void)argv[i]; }\n"
" return 0;\n"
"}\n",
"fixed-width-types", LintMode::Apply);
// Somebody else's widths, kept.
Check(r.text.contains("extern \"C\" unsigned long CApiCall(unsigned int flags);"), "fixed-width: extern \"C\" signature keeps its widths");
Check(r.text.contains("unsigned long fromLibc = std::strtoul"), "fixed-width: a value from a C library keeps its width");
Check(r.text.contains("int main(int argc, char** argv)"), "fixed-width: main's signature is untouched");
// Ours, converted.
Check(r.text.contains("std::int64_t OurOwnApi(std::int32_t value)"), "fixed-width: our own signature converts");
Check(r.text.contains("std::uint32_t ours = 1;"), "fixed-width: our own local converts");
// The loop counter inside main's BODY is ordinary code. The old rule
// skipped it because `argc` appeared on the same line; only the
// signature is exempt now, not everything near it.
Check(r.text.contains("for (std::int32_t i = 0;"), "fixed-width: main's body is not exempt, only its signature");
}
if (Failures > 0) {
std::println(std::cerr, "{} assertions failed", Failures);
return 1;