source: icGREP/icgrep-devel/icgrep/re/re_toolchain.cpp @ 5816

Last change on this file since 5816 was 5816, checked in by cameron, 13 months ago

Supporting multiple alphabets in RE compilation - initial check-in

File size: 6.9 KB
Line 
1/*
2 *  Copyright (c) 2017 International Characters.
3 *  This software is licensed to the public under the Open Software License 3.0.
4 *  icgrep is a trademark of International Characters.
5 */
6
7#include <toolchain/toolchain.h>
8#include <grep_interface.h>
9#include <re/re_toolchain.h>
10#include <cc/cc_compiler.h>            // for CC_Compiler
11#include <llvm/Support/CommandLine.h>  // for clEnumVal, clEnumValEnd, Optio...
12#include <re/re_compiler.h>            // for RE_Compiler
13#include <re/re_nullable.h>            // for RE_Nullable
14#include <re/re_star_normal.h>         // for RE_Star_Normal
15#include <re/re_simplifier.h>          // for RE_Simplifier
16#include <re/re_minimizer.h>
17#include <re/re_local.h>
18#include <re/printer_re.h>
19#include <re/re_analysis.h>
20#include <re/re_cc.h>
21#include <re/casing.h>
22#include <re/exclude_CC.h>
23#include <re/re_name_resolve.h>
24#include <re/grapheme_clusters.h>
25#include <llvm/Support/raw_ostream.h>
26
27using namespace pablo;
28using namespace llvm;
29
30namespace re {
31
32static cl::OptionCategory RegexOptions("Regex Toolchain Options",
33                                              "These options control the regular expression transformation and compilation.");
34const cl::OptionCategory * re_toolchain_flags() {
35    return &RegexOptions;
36}
37
38static cl::bits<RE_PrintFlags> 
39    PrintOptions(cl::values(clEnumVal(ShowREs, "Print parsed or generated regular expressions"),
40                            clEnumVal(ShowAllREs, "Print all regular expression passes"),
41                            clEnumVal(ShowStrippedREs, "Print REs with nullable prefixes/suffixes removed"),
42                            clEnumVal(ShowSimplifiedREs, "Print final simplified REs")
43                            CL_ENUM_VAL_SENTINEL), cl::cat(RegexOptions));
44
45static cl::bits<RE_AlgorithmFlags>
46    AlgorithmOptions(cl::values(clEnumVal(DisableLog2BoundedRepetition, "disable log2 optimizations for bounded repetition of bytes"),
47                              clEnumVal(DisableIfHierarchy, "disable nested if hierarchy for generated Unicode classes (not recommended)"), 
48                              clEnumVal(DisableMatchStar, "disable MatchStar optimization"), 
49                              clEnumVal(DisableUnicodeMatchStar, "disable Unicode MatchStar optimization"),
50                              clEnumVal(DisableUnicodeLineBreak, "disable Unicode line breaks - use LF only")
51                              CL_ENUM_VAL_SENTINEL), cl::cat(RegexOptions));
52
53bool AlgorithmOptionIsSet(RE_AlgorithmFlags flag) {
54    return AlgorithmOptions.isSet(flag);
55}
56
57int IfInsertionGap;
58static cl::opt<int, true> 
59    IfInsertionGapOption("if-insertion-gap",  cl::location(IfInsertionGap), cl::init(3),
60                         cl::desc("minimum number of nonempty elements between inserted if short-circuit tests"), 
61                         cl::cat(RegexOptions));
62
63RE * resolveModesAndExternalSymbols(RE * r) {
64    r = resolveGraphemeMode(r, false /* not in grapheme mode at top level*/);
65    if (PrintOptions.isSet(ShowAllREs)) {
66        errs() << "resolveGraphemeMode:\n" << Printer_RE::PrintRE(r) << '\n';
67    }
68    r = re::resolveUnicodeProperties(r);
69    if (PrintOptions.isSet(ShowAllREs)) {
70        errs() << "resolveUnicodeProperties:\n" << Printer_RE::PrintRE(r) << '\n';
71    }
72    r = resolveCaseInsensitiveMode(r, grep::IgnoreCaseFlag);
73    if (PrintOptions.isSet(ShowAllREs)) {
74        errs() << "resolveCaseInsensitiveMode:\n" << Printer_RE::PrintRE(r) << '\n';
75    }
76    return r;
77}
78
79RE * excludeUnicodeLineBreak(RE * r) {
80    r = exclude_CC(r, re::makeCC(re::makeCC(0x0A, 0x0D), re::makeCC(re::makeCC(0x85), re::makeCC(0x2028, 0x2029))));
81    if (PrintOptions.isSet(ShowAllREs)) {
82        errs() << "excludeUnicodeLineBreak:\n" << Printer_RE::PrintRE(r) << '\n';
83    }
84    return r;
85}
86
87RE * multiplexing_prepasses(RE * r) {
88    std::vector<re::CC *> charclasses;
89    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowREs)) {
90        errs() << "Parser:\n" << Printer_RE::PrintRE(r) << '\n';
91    }
92    //Optimization passes to simplify the AST.
93    r = RE_Nullable::removeNullablePrefix(r);
94    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowStrippedREs)) {
95        errs() << "RemoveNullablePrefix:\n" << Printer_RE::PrintRE(r) << '\n';
96    }
97    r = RE_Nullable::removeNullableSuffix(r);
98    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowStrippedREs)) {
99        errs() << "RemoveNullableSuffix:\n" << Printer_RE::PrintRE(r) << '\n';
100    }
101    r = RE_Star_Normal::star_normal(r);
102    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowSimplifiedREs)) {
103        //Print to the terminal the AST that was transformed to the star normal form.
104        errs() << "Star_Normal_Form:\n" << Printer_RE::PrintRE(r) << '\n';
105    }
106    r = re::resolveNames(r);
107    if (PrintOptions.isSet(ShowAllREs)) {
108        errs() << "resolveNames:\n" << Printer_RE::PrintRE(r) << '\n';
109    }
110
111    r = RE_Simplifier::simplify(r);
112    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowSimplifiedREs)) {
113        //Print to the terminal the AST that was generated by the simplifier.
114        errs() << "Simplifier:\n" << Printer_RE::PrintRE(r) << '\n';
115    }
116    return r;
117}
118
119RE * regular_expression_passes(RE * r)  {
120
121    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowREs)) {
122        errs() << "Parser:\n" << Printer_RE::PrintRE(r) << '\n';
123    }
124
125    //Optimization passes to simplify the AST.
126    r = RE_Nullable::removeNullablePrefix(r);
127    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowStrippedREs)) {
128        errs() << "RemoveNullablePrefix:\n" << Printer_RE::PrintRE(r) << '\n';
129    }
130    r = RE_Nullable::removeNullableSuffix(r);
131    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowStrippedREs)) {
132        errs() << "RemoveNullableSuffix:\n" << Printer_RE::PrintRE(r) << '\n';
133    }
134    r = RE_Simplifier::simplify(r);
135
136    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowSimplifiedREs)) {
137        //Print to the terminal the AST that was generated by the simplifier.
138        errs() << "Simplifier:\n" << Printer_RE::PrintRE(r) << '\n';
139    }
140   
141//    r = RE_Minimizer::minimize(r);
142
143//    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowSimplifiedREs)) {
144//        //Print to the terminal the AST that was generated by the simplifier.
145//        errs() << "Minimizer:\n" << Printer_RE::PrintRE(r) << '\n';
146//    }
147
148    r = RE_Star_Normal::star_normal(r);
149
150    if (PrintOptions.isSet(ShowAllREs) || PrintOptions.isSet(ShowSimplifiedREs)) {
151        //Print to the terminal the AST that was transformed to the star normal form.
152        errs() << "Star_Normal_Form:\n" << Printer_RE::PrintRE(r) << '\n';
153    }
154
155    return r;
156}
157   
158PabloAST * re2pablo_compiler(PabloKernel * kernel, RE * re_ast) {
159    Var * const basis = kernel->getInputStreamVar("basis");
160    cc::CC_Compiler cc_compiler(kernel, basis);
161    RE_Compiler re_compiler(kernel, cc_compiler);
162    return re_compiler.compile(re_ast);
163}
164
165}
Note: See TracBrowser for help on using the repository browser.