llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT--> @llvm/pr-subscribers-clang Author: Aaron Ballman (AaronBallman) <details> <summary>Changes</summary> The utilities we use for lexing string and character literals can be run in a mode where we pass a null pointer for the diagnostics engine. This mode is used by the format string checkers, for example. However, there were two places that failed to account for a null diagnostic engine pointer: `\o{}` and `\x{}`. This patch adds a check for a null pointer and correctly handles fallback behavior. Fixes #<!-- -->102218 --- Full diff: https://github.com/llvm/llvm-project/pull/102339.diff 3 Files Affected: - (modified) clang/docs/ReleaseNotes.rst (+2) - (modified) clang/lib/Lex/LiteralSupport.cpp (+8-6) - (modified) clang/test/Lexer/char-escapes-delimited.c (+13) ``````````diff diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst index 978b4ac8ea2e3..565812e9b503f 100644 --- a/clang/docs/ReleaseNotes.rst +++ b/clang/docs/ReleaseNotes.rst @@ -170,6 +170,8 @@ Bug Fixes in This Version be used in C++. - Fixed a failed assertion when checking required literal types in C context. (#GH101304). - Fixed a crash when trying to transform a dependent address space type. Fixes #GH101685. +- Fixed a crash when diagnosing format strings and encountering an empty + delimited escape sequence (e.g., ``"\o{}"``). #GH102218 Bug Fixes to Compiler Builtins ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ diff --git a/clang/lib/Lex/LiteralSupport.cpp b/clang/lib/Lex/LiteralSupport.cpp index 9d2720af5dbd9..225a6c2d15baa 100644 --- a/clang/lib/Lex/LiteralSupport.cpp +++ b/clang/lib/Lex/LiteralSupport.cpp @@ -190,9 +190,10 @@ static unsigned ProcessCharEscape(const char *ThisTokBegin, Delimited = true; ThisTokBuf++; if (*ThisTokBuf == '}') { - Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, - diag::err_delimited_escape_empty); - return ResultChar; + HadError = true; + if (Diags) + Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, + diag::err_delimited_escape_empty); } } else if (ThisTokBuf == ThisTokEnd || !isHexDigit(*ThisTokBuf)) { if (Diags) @@ -283,9 +284,10 @@ static unsigned ProcessCharEscape(const char *ThisTokBegin, Delimited = true; ++ThisTokBuf; if (*ThisTokBuf == '}') { - Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, - diag::err_delimited_escape_empty); - return ResultChar; + HadError = true; + if (Diags) + Diag(Diags, Features, Loc, ThisTokBegin, EscapeBegin, ThisTokBuf, + diag::err_delimited_escape_empty); } while (ThisTokBuf != ThisTokEnd) { diff --git a/clang/test/Lexer/char-escapes-delimited.c b/clang/test/Lexer/char-escapes-delimited.c index 5327ef700b0e2..7a8986bc5f867 100644 --- a/clang/test/Lexer/char-escapes-delimited.c +++ b/clang/test/Lexer/char-escapes-delimited.c @@ -123,3 +123,16 @@ void separators(void) { static_assert('\N??<DOLLAR SIGN??>' == '$'); // expected-warning 2{{trigraph converted}} \ // ext-warning {{extension}} cxx23-warning {{C++23}} #endif + +void GH102218(void) { + // The format specifier checking code runs the lexer with diagnostics + // disabled. This used to crash Clang for malformed \o and \x because the + // lexer missed a null pointer check for the diagnostics engine in that case. + extern int printf(const char *, ...); + printf("\o{}"); // expected-error {{delimited escape sequence cannot be empty}} + printf("\x{}"); // expected-error {{delimited escape sequence cannot be empty}} + + // These cases always worked but are here for completeness. + printf("\u{}"); // expected-error {{delimited escape sequence cannot be empty}} + printf("\N{}"); // expected-error {{delimited escape sequence cannot be empty}} +} `````````` </details> https://github.com/llvm/llvm-project/pull/102339 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits