[MC] Skip lower-case integer suffixes

[ksaunders]

Summary

In the GNU assemblers, lower-cased integer suffices are permitted. However, in LLVM, it causes an error.

This is especially important when using the C pre-processor with the assembler: some shared code between C and assembler may use lower-cased suffices.

For example,

.global main
main:
	mov x0, 1024u
	ret

Is permitted in AArch64 binutils 2.35.1. However with Clang it produces the following error:

<source>:3:14: error: unexpected token in argument list
        mov x0, 1024u
                    ^

This patch addresses this issue.

[MaskRay]

Thanks. I changed the subject of this patch. You may call arc amend to amend your local commit message or change it manually.

[MaskRay]

If you provide name <email>, I can push this for you if you don't have push permission.

[ksaunders]

I do not have push permissions, so please push it on my behalf as Keegan Saunders <[e-mail]>. Thank you.

llvm/lib/MC/MCParser/AsmLexer.cpp

return AsmToken(AsmToken::EndOfStatement,
                     StringRef(TokStart, CurPtr - TokStart));
   IsAtStartOfStatement = true;
 
   return AsmToken(AsmToken::EndOfStatement,
                   StringRef(TokStart, CurPtr - 1 - TokStart));
 }
 
 static void SkipIgnoredIntegerSuffix(const char *&CurPtr) {
-  // Skip ULL, UL, U, L and LL suffices.
-  if (CurPtr[0] == 'U')
+  // Skip case-insensitive ULL, UL, U, L and LL suffixes.
+  if (CurPtr[0] == 'U' || CurPtr[0] == 'u')
     ++CurPtr;
-  if (CurPtr[0] == 'L')
+  if (CurPtr[0] == 'L' || CurPtr[0] == 'l')
     ++CurPtr;
-  if (CurPtr[0] == 'L')
+  if (CurPtr[0] == 'L' || CurPtr[0] == 'l')
     ++CurPtr;
 }
 
 // Look ahead to search for first non-hex digit, if it's [hH], then we treat the
 // integer as a hexadecimal, possibly with leading zeroes.
 static unsigned doHexLookAhead(const char *&CurPtr, unsigned DefaultRadix,
                                bool LexHex) {
   const char *FirstNonDec = nullptr;

llvm/test/MC/AsmParser/exprs.s

 k:
         .data
         check_expr !1 + 2, 2
         check_expr !0, 1
         check_expr ~0, -1
         check_expr -1, ~0
         check_expr +1, 1
         check_expr 1 + 2, 3
+        check_expr 1U + 2U, 3U
+        check_expr 1L + 2L, 3L
+        check_expr 1UL + 2UL, 3UL
+        check_expr 1LL + 2L, 3LL
+        check_expr 1ULL + 2ULL, 3ULL
+        check_expr 1u + 2u, 3u
+        check_expr 1l + 2l, 3l
+        check_expr 1ul + 2ul, 3ul
+        check_expr 1ll + 2L, 3ll
+        check_expr 1ull + 2ull, 3ull
         check_expr 1 & 3, 1
         check_expr 4 / 2, 2
         check_expr 4 / -2, -2
         check_expr 1 == 1, -1
         check_expr 1 == 0, 0
         check_expr 1 > 0, -1
         check_expr 1 >= 1, -1
         check_expr 1 < 2, -1