xref: /freebsd/contrib/llvm-project/llvm/lib/Support/DJB.cpp (revision 0b57cec536236d46e3dba9bd041533462f33dbb7)
1*0b57cec5SDimitry Andric //===-- Support/DJB.cpp ---DJB Hash -----------------------------*- C++ -*-===//
2*0b57cec5SDimitry Andric //
3*0b57cec5SDimitry Andric // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4*0b57cec5SDimitry Andric // See https://llvm.org/LICENSE.txt for license information.
5*0b57cec5SDimitry Andric // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6*0b57cec5SDimitry Andric //
7*0b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
8*0b57cec5SDimitry Andric //
9*0b57cec5SDimitry Andric // This file contains support for the DJ Bernstein hash function.
10*0b57cec5SDimitry Andric //
11*0b57cec5SDimitry Andric //===----------------------------------------------------------------------===//
12*0b57cec5SDimitry Andric 
13*0b57cec5SDimitry Andric #include "llvm/Support/DJB.h"
14*0b57cec5SDimitry Andric #include "llvm/ADT/ArrayRef.h"
15*0b57cec5SDimitry Andric #include "llvm/Support/Compiler.h"
16*0b57cec5SDimitry Andric #include "llvm/Support/ConvertUTF.h"
17*0b57cec5SDimitry Andric #include "llvm/Support/Unicode.h"
18*0b57cec5SDimitry Andric 
19*0b57cec5SDimitry Andric using namespace llvm;
20*0b57cec5SDimitry Andric 
21*0b57cec5SDimitry Andric static UTF32 chopOneUTF32(StringRef &Buffer) {
22*0b57cec5SDimitry Andric   UTF32 C;
23*0b57cec5SDimitry Andric   const UTF8 *const Begin8Const =
24*0b57cec5SDimitry Andric       reinterpret_cast<const UTF8 *>(Buffer.begin());
25*0b57cec5SDimitry Andric   const UTF8 *Begin8 = Begin8Const;
26*0b57cec5SDimitry Andric   UTF32 *Begin32 = &C;
27*0b57cec5SDimitry Andric 
28*0b57cec5SDimitry Andric   // In lenient mode we will always end up with a "reasonable" value in C for
29*0b57cec5SDimitry Andric   // non-empty input.
30*0b57cec5SDimitry Andric   assert(!Buffer.empty());
31*0b57cec5SDimitry Andric   ConvertUTF8toUTF32(&Begin8, reinterpret_cast<const UTF8 *>(Buffer.end()),
32*0b57cec5SDimitry Andric                      &Begin32, &C + 1, lenientConversion);
33*0b57cec5SDimitry Andric   Buffer = Buffer.drop_front(Begin8 - Begin8Const);
34*0b57cec5SDimitry Andric   return C;
35*0b57cec5SDimitry Andric }
36*0b57cec5SDimitry Andric 
37*0b57cec5SDimitry Andric static StringRef toUTF8(UTF32 C, MutableArrayRef<UTF8> Storage) {
38*0b57cec5SDimitry Andric   const UTF32 *Begin32 = &C;
39*0b57cec5SDimitry Andric   UTF8 *Begin8 = Storage.begin();
40*0b57cec5SDimitry Andric 
41*0b57cec5SDimitry Andric   // The case-folded output should always be a valid unicode character, so use
42*0b57cec5SDimitry Andric   // strict mode here.
43*0b57cec5SDimitry Andric   ConversionResult CR = ConvertUTF32toUTF8(&Begin32, &C + 1, &Begin8,
44*0b57cec5SDimitry Andric                                            Storage.end(), strictConversion);
45*0b57cec5SDimitry Andric   assert(CR == conversionOK && "Case folding produced invalid char?");
46*0b57cec5SDimitry Andric   (void)CR;
47*0b57cec5SDimitry Andric   return StringRef(reinterpret_cast<char *>(Storage.begin()),
48*0b57cec5SDimitry Andric                    Begin8 - Storage.begin());
49*0b57cec5SDimitry Andric }
50*0b57cec5SDimitry Andric 
51*0b57cec5SDimitry Andric static UTF32 foldCharDwarf(UTF32 C) {
52*0b57cec5SDimitry Andric   // DWARF v5 addition to the unicode folding rules.
53*0b57cec5SDimitry Andric   // Fold "Latin Small Letter Dotless I" and "Latin Capital Letter I With Dot
54*0b57cec5SDimitry Andric   // Above" into "i".
55*0b57cec5SDimitry Andric   if (C == 0x130 || C == 0x131)
56*0b57cec5SDimitry Andric     return 'i';
57*0b57cec5SDimitry Andric   return sys::unicode::foldCharSimple(C);
58*0b57cec5SDimitry Andric }
59*0b57cec5SDimitry Andric 
60*0b57cec5SDimitry Andric static Optional<uint32_t> fastCaseFoldingDjbHash(StringRef Buffer, uint32_t H) {
61*0b57cec5SDimitry Andric   bool AllASCII = true;
62*0b57cec5SDimitry Andric   for (unsigned char C : Buffer) {
63*0b57cec5SDimitry Andric     H = H * 33 + ('A' <= C && C <= 'Z' ? C - 'A' + 'a' : C);
64*0b57cec5SDimitry Andric     AllASCII &= C <= 0x7f;
65*0b57cec5SDimitry Andric   }
66*0b57cec5SDimitry Andric   if (AllASCII)
67*0b57cec5SDimitry Andric     return H;
68*0b57cec5SDimitry Andric   return None;
69*0b57cec5SDimitry Andric }
70*0b57cec5SDimitry Andric 
71*0b57cec5SDimitry Andric uint32_t llvm::caseFoldingDjbHash(StringRef Buffer, uint32_t H) {
72*0b57cec5SDimitry Andric   if (Optional<uint32_t> Result = fastCaseFoldingDjbHash(Buffer, H))
73*0b57cec5SDimitry Andric     return *Result;
74*0b57cec5SDimitry Andric 
75*0b57cec5SDimitry Andric   std::array<UTF8, UNI_MAX_UTF8_BYTES_PER_CODE_POINT> Storage;
76*0b57cec5SDimitry Andric   while (!Buffer.empty()) {
77*0b57cec5SDimitry Andric     UTF32 C = foldCharDwarf(chopOneUTF32(Buffer));
78*0b57cec5SDimitry Andric     StringRef Folded = toUTF8(C, Storage);
79*0b57cec5SDimitry Andric     H = djbHash(Folded, H);
80*0b57cec5SDimitry Andric   }
81*0b57cec5SDimitry Andric   return H;
82*0b57cec5SDimitry Andric }
83