re2 1.19.2 → 1.20.0

package/vendor/re2/app/app.ts

@@ -0,0 +1,111 @@
+// Copyright 2022 The RE2 Authors.  All Rights Reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+import {css, html, LitElement, render} from 'lit';
+import {customElement} from 'lit/decorators.js';
+
+import /*default*/ loadModule from './_re2';
+import {Info, MainModule} from './_re2';
+
+var _re2: MainModule;
+loadModule().then((module: MainModule) => {
+  _re2 = module;
+  render(html`<title>re2-dev</title><re2-dev></re2-dev>`, document.body);
+});
+
+@customElement('re2-dev')
+export class RE2Dev extends LitElement {
+  private _pattern: string = '';
+  private _info: Info|null = null;
+
+  constructor() {
+    super();
+    this._pattern = decodeURIComponent(window.location.hash.slice(1));
+    this._info = this._pattern ? _re2.getInfo(this._pattern) : null;
+    this.requestUpdate();
+  }
+
+  private _onChange = (e: Event) => {
+    this._pattern = (e.target as HTMLInputElement).value;
+    this._info = this._pattern ? _re2.getInfo(this._pattern) : null;
+    this.requestUpdate();
+    window.location.hash = '#' + encodeURIComponent(this._pattern);
+  };
+
+  static override styles = css`
+.code {
+  font-family: monospace;
+  white-space: pre-line;
+}
+`;
+
+  override render() {
+    var fragments = [];
+    fragments.push(html`
+<div>
+  <input type="text" size="48" @change=${this._onChange} .value=${this._pattern}>
+</div>
+`);
+
+    if (this._info === null) {
+      return html`${fragments}`;
+    }
+
+    if (this._info.error) {
+      fragments.push(html`
+<br>
+<div>
+  error:
+  <span class="code">${this._info.error}</span>
+</div>
+`);
+      return html`${fragments}`;
+    }
+
+    fragments.push(html`
+<br>
+<div>
+  pattern:
+  <span class="code">${this._info.pattern}</span>
+  <br>
+  prefix:
+  <span class="code">${this._info.prefix}</span>
+  ·
+  _foldcase:
+  <span class="code">${this._info.prefix_foldcase}</span>
+  <br>
+  accel_prefix:
+  <span class="code">${this._info.accel_prefix}</span>
+  ·
+  _foldcase:
+  <span class="code">${this._info.accel_prefix_foldcase}</span>
+  <br>
+  num_captures:
+  <span class="code">${this._info.num_captures}</span>
+  <br>
+  is_one_pass:
+  <span class="code">${this._info.is_one_pass}</span>
+  <br>
+  can_bit_state:
+  <span class="code">${this._info.can_bit_state}</span>
+  <br>
+  <br>
+  bytecode:
+  <br>
+  <span class="code">${this._info.bytecode}</span>
+  <br>
+  bytemap:
+  <br>
+  <span class="code">${this._info.bytemap}</span>
+</div>
+`);
+    return html`${fragments}`;
+  }
+}
+
+declare global {
+  interface HTMLElementTagNameMap {
+    're2-dev': RE2Dev;
+  }
+}

package/vendor/re2/app/build.sh

@@ -0,0 +1,50 @@
+#!/bin/bash
+set -eux
+
+SRCDIR=$(readlink --canonicalize $(dirname $0))
+DSTDIR=$(mktemp --directory --tmpdir $(basename $0).XXXXXXXXXX)
+
+BAZEL=/tmp/bazel
+BAZELISK_RELEASE=v1.16.0
+
+if [[ ${UID} -ne 0 ]]; then
+  if [[ -d deploy ]]; then
+    echo -e '\033[1;31m' "** The ${PWD}/deploy directory exists! Refusing to clobber it! **" '\033[0m'
+    exit 1
+  fi
+  mkdir deploy
+  sudo docker run -i -t --pull always --rm -v ${SRCDIR}/..:/src -v ${PWD}:/dst emscripten/emsdk /src/app/$(basename $0)
+  ls -l deploy
+else
+  wget -O ${BAZEL} https://github.com/bazelbuild/bazelisk/releases/download/${BAZELISK_RELEASE}/bazelisk-linux-amd64
+  chmod +x ${BAZEL}
+
+  cd ${SRCDIR}
+  # Emscripten doesn't support `-fstack-protector`.
+  AR=emar CC=emcc \
+    ${BAZEL} build --compilation_mode=opt \
+    --copt=-fno-stack-protector \
+    -- :all
+  # Bazel doesn't retain the `_re2.wasm` artifact;
+  # we have to redo the link command to obtain it.
+  pushd ..
+  emcc @bazel-bin/app/_re2.js-2.params
+  cd bazel-bin/app
+  cp _re2.js _re2.wasm ${DSTDIR}
+  popd
+  # Clean up the sundry Bazel output directories.
+  ${BAZEL} clean --expunge
+  cp app.ts index.html _re2.d.ts ${DSTDIR}
+  cp package.json rollup.config.js tsconfig.json ${DSTDIR}
+
+  cd ${DSTDIR}
+  npm install
+  npx tsc
+  npx rollup -c rollup.config.js -d deploy
+  mv deploy/* /dst/deploy
+fi
+
+cd ${SRCDIR}
+rm -rf ${DSTDIR}
+
+exit 0

package/vendor/re2/app/index.html

@@ -0,0 +1,5 @@
+<!DOCTYPE html>
+<meta charset="utf-8">
+<meta name="viewport" content="width=device-width, initial-scale=1">
+<style>:root { color-scheme: dark light; }</style>
+<script type="module" src="app.js"></script>

package/vendor/re2/app/package.json

@@ -0,0 +1,14 @@
+{
+  "dependencies": {
+    "lit": "*"
+  },
+  "devDependencies": {
+    "@rollup/plugin-node-resolve": "*",
+    "@rollup/plugin-terser": "*",
+    "@web/rollup-plugin-html": "*",
+    "@web/rollup-plugin-import-meta-assets": "*",
+    "rollup": "~2",
+    "tslib": "*",
+    "typescript": "*"
+  }
+}

package/vendor/re2/app/rollup.config.js

@@ -0,0 +1,28 @@
+// Copyright 2022 The RE2 Authors.  All Rights Reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+import nodeResolve from '@rollup/plugin-node-resolve';
+import terser from '@rollup/plugin-terser';
+import html from '@web/rollup-plugin-html';
+import {importMetaAssets} from '@web/rollup-plugin-import-meta-assets';
+
+export default {
+  input: 'index.html',
+  output: {
+    entryFileNames: '[hash].js',
+    chunkFileNames: '[hash].js',
+    assetFileNames: '[hash][extname]',
+    format: 'es',
+  },
+  preserveEntrySignatures: false,
+  plugins:
+      [
+        html({
+          minify: true,
+        }),
+        nodeResolve(),
+        terser(),
+        importMetaAssets(),
+      ],
+};

package/vendor/re2/app/tsconfig.json

@@ -0,0 +1,17 @@
+{
+  "compilerOptions": {
+    "target": "esnext",
+    "module": "esnext",
+    "moduleResolution": "node",
+    "noEmitOnError": true,
+    "lib": ["esnext", "dom"],
+    "strict": true,
+    "esModuleInterop": false,
+    "allowSyntheticDefaultImports": true,
+    "experimentalDecorators": true,
+    "importHelpers": true,
+    "sourceMap": true,
+    "inlineSources": true,
+    "incremental": true
+  }
+}

package/vendor/re2/benchlog/benchplot.py

@@ -0,0 +1,98 @@
+#!/usr/bin/env python
+
+import argparse     # for ArgumentParser
+import subprocess   # for Popen
+import tempfile     # for NamedTemporaryFile
+import os           # for remove
+
+class gnuplot(object):
+
+    output = "result.png"
+
+    script = """
+             set terminal png size 1024, 768
+             set output "{}.png"
+             set title "re2 benchlog"
+             set datafile separator ";"
+             set grid x y
+             set ylabel "MB/s"
+             set autoscale
+             plot """
+
+    template = """'{}' using 1:5:xticlabels(2) with linespoints linewidth 3 title "{}",\\\n"""
+
+    benchdata = dict()
+    tempfiles = []
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, type, value, traceback):
+        """
+        remove all temporary files
+        """
+
+        for filename in self.tempfiles:
+            os.remove(filename)
+
+    def parse_re2_benchlog(self, filename):
+        """
+        parse the input benchlog and return a dictionary contain bench data
+        """
+
+        benchdata = self.benchdata
+
+        with open(filename) as f:
+
+            for raw in f.readlines():
+
+                data = raw.split('\t')
+
+                if len(data) == 4:
+
+                    data = data[0].split('/') + data[1:]
+                    data = list(map(str.strip, data))
+
+                    if not benchdata.get(data[0]):
+                        benchdata[data[0]] = [ data[1:] ]
+                    else:
+                        benchdata[data[0]].append(data[1:])
+
+    def gen_csv(self):
+        """
+        generate temporary csv files
+        """
+
+        for name, data in self.benchdata.items():
+
+            with tempfile.NamedTemporaryFile(delete=False) as f:
+
+                for index, line in enumerate(data):
+                    f.write('{};{}\n'.format(index, ';'.join(line)).encode())
+
+                self.tempfiles.append(f.name)
+                self.script = self.script + self.template.format(f.name, name)
+
+    def run(self):
+        self.gen_csv()
+        script = self.script[:-3].format(self.output)
+        command = subprocess.Popen(['gnuplot'], stdin=subprocess.PIPE)
+        command.communicate(script.encode())
+
+
+if __name__ == '__main__':
+
+    parser = argparse.ArgumentParser(description='generate plots for benchlog')
+    parser.add_argument('benchlog', type=str, help='benchlog generated by re2')
+    args = parser.parse_args()
+
+    try:
+        subprocess.Popen(['gnuplot'], stdin=subprocess.PIPE)
+    except FileNotFoundError:
+        print('you can install "gnuplot" to generate plots automatically')
+        exit(1)
+
+    with gnuplot() as plot:
+        plot.output = args.benchlog
+        plot.parse_re2_benchlog(args.benchlog)
+        plot.run()

package/vendor/re2/benchlog/mktable

@@ -0,0 +1,155 @@
+#!/usr/bin/perl
+# XXX
+
+sub table() {
+	my ($name) = @_;
+	print <<'EOF';
+<table border=0>
+<tr><th>System</th><th>PCRE</th><th>RE2</th></tr>
+EOF
+	foreach my $sys (@sys) {
+		my $ns_pcre = $data{$sys}->{sprintf($name, "PCRE")}->{'ns/op'};
+		my $ns_re2 = $data{$sys}->{sprintf($name, "RE2")}->{'ns/op'};
+		printf "<tr><td>%s</td><td>%.1f µs</td><td>%.1f µs</td></tr>\n", $sysname{$sys}, $ns_pcre/1000., $ns_re2/1000.;
+	}
+	print <<'EOF';
+<tr height=5><td colspan=3></td></tr>
+</table>
+EOF
+}
+
+@sizes = (
+	"8", "16", "32", "64", "128", "256", "512",
+	"1K", "2K", "4K", "8K", "16K", "32K", "64K", "128K", "256K", "512K",
+	"1M", "2M", "4M", "8M", "16M"
+);
+
+%color = (
+	"PCRE" => "0.7 0 0",
+	"RE2" => "0 0 1",
+);
+
+$ngraph = 0;
+
+sub graph() {
+	my ($name) = @_;
+	
+	my $sys = "wreck";
+	my $base = sprintf("regexp3g%d", ++$ngraph);
+
+	open(JGR, ">$base.jgr") || die "open >$base.jgr: $!";
+	printf JGR "bbox -20 -12 392 95\n";
+	printf JGR "newgraph clip x_translate 0.25 y_translate 0.25\n";
+	$ymax = 0;
+	%lastx = ();
+	%lasty = ();
+	foreach my $who ("PCRE", "RE2") {
+		printf JGR "newcurve pts\n";
+		for(my $i=0; $i<@sizes; $i++) {
+			my $key = sprintf("%s%s/%s", $name, $who, $sizes[$i]);
+			my $val = $data{$sys}->{$key}->{'MB/s'};
+			next if !defined($val);
+			if($val > $ymax) {
+				$ymax = $val;
+			}
+			$lastx{$who} = $i;
+			$lasty{$who} = $val;
+			printf JGR "$i %f (* %s *)\n", $val, $key;
+		}
+		my $color = $color{$who};
+		printf JGR "marktype none color $color linethickness 2 linetype solid label : $who\n";
+	}
+	my $n = @sizes;
+	printf JGR "xaxis min -1 max $n size 5 label : text size (bytes)\n";
+	printf JGR "  no_auto_hash_marks hash_labels fontsize 9\n";
+	for($i=0; $i<@sizes; $i+=3) {
+		printf JGR "  hash_at $i hash_label at $i : $sizes[$i]\n";
+	}
+	my $y = 1;
+	while(10*$y <= $ymax) {
+		$y = 10*$y;
+	}
+	for($i=2; $i<=10; $i++) {
+		if($i*$y > $ymax) {
+			$y = $i*$y;
+			last;
+		}
+	}
+	foreach my $who ("PCRE", "RE2") {
+		$x1 = $lastx{$who};
+		$y1 = $lasty{$who};
+		$x1 *= 1.01;
+		my $v = "vjc";
+		if($y1 < 0.05 * $y) {
+			$v = "vjb";
+			$y1 = 0.05 * $y;
+		}
+		printf JGR "newstring x $x1 y $y1 hjl $v : $who\n";
+	}
+	printf JGR "yaxis min 0 max $y size 1 label : speed (MB/s)\n";
+	printf JGR "  hash_labels fontsize 9\n";
+	# printf JGR "legend defaults font Times-Roman fontsize 10 x 0 y $y hjl vjt\n";
+
+	system("jgraph $base.jgr >$base.eps"); # die "system: $!";
+	system("gs -dTextAlphaBits=4 -dGraphicsAlphaBits=4 -dEPSCrop -sDEVICE=png16m -r100 -sOutputFile=$base.png -dBATCH -dQUIT -dQUIET -dNOPAUSE $base.eps");
+	
+	printf "<img src=$base.png>\n"
+	
+}
+
+sub skip() {
+	while(<>) {
+		if(/^<!-- -->/) {
+			print;
+			last;
+		}
+	}
+}
+
+@sys = ("r70", "c2", "wreck", "mini");
+%sysname = (
+	"r70" => "AMD Opteron 8214 HE, 2.2 GHz",
+	"c2" => "Intel Core2 Duo E7200, 2.53 GHz",
+	"wreck" => "Intel Xeon 5150, 2.66 GHz (Mac Pro)",
+	"mini" => "Intel Core2 T5600, 1.83 GHz (Mac Mini)",
+);
+
+%func = (
+	"table" => \&table,
+	"graph" => \&graph,
+	
+);
+
+foreach my $sys (@sys) {
+	open(F, "benchlog.$sys") || die "open benchlog.$sys: $!";
+	my %sysdat;
+	while(<F>) {
+		if(/^([A-Za-z0-9_\/]+)\s+(\d+)\s+(\d+) ns\/op/) {
+			my %row;
+			$row{"name"} = $1;
+			$row{"iter"} = $2;
+			$row{"ns/op"} = $3;
+			if(/([\d.]+) MB\/s/){
+				$row{"MB/s"} = $1;
+			}
+			$sysdat{$row{"name"}} = \%row;
+		}
+	}
+	close F;	
+	$data{$sys} = \%sysdat;
+}
+
+while(<>) {
+	print;
+	if(/^<!-- benchlog (\w+) -->/) {
+		$func{$1}();
+		skip();
+		next;
+	}
+	if(/^<!-- benchlog (\w+) ([%\w]+) -->/) {
+		$func{$1}($2);
+		skip();
+		next;
+	}
+}
+

package/vendor/re2/doc/mksyntaxgo

@@ -0,0 +1,42 @@
+#!/bin/sh
+
+set -e
+out=$GOROOT/src/regexp/syntax/doc.go
+cp syntax.txt $out
+sam -d $out <<'!'
+,x g/NOT SUPPORTED/d
+/^Unicode character class/,$d
+,s/[«»]//g
+,x g/^Possessive repetitions:/d
+,x g/\\C/d
+,x g/Flag syntax/d
+,s/.=(true|false)/flag &/g
+,s/^Flags:/  Flag syntax is xyz (set) or -xyz (clear) or xy-z (set xy, clear z). The flags are:\n/
+,s/\n\n\n+/\n\n/g
+,x/(^.*	.*\n)+/ | awk -F'	' '{printf("  %-14s %s\n", $1, $2)}'
+1,2c
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// DO NOT EDIT. This file is generated by mksyntaxgo from the RE2 distribution.
+
+/*
+Package syntax parses regular expressions into parse trees and compiles
+parse trees into programs. Most clients of regular expressions will use the
+facilities of package regexp (such as Compile and Match) instead of this package.
+
+Syntax
+
+The regular expression syntax understood by this package when parsing with the Perl flag is as follows.
+Parts of the syntax can be disabled by passing alternate flags to Parse.
+
+.
+$a
+Unicode character classes are those in unicode.Categories and unicode.Scripts.
+*/
+package syntax
+.
+w
+q
+!

package/vendor/re2/doc/mksyntaxhtml

@@ -0,0 +1,42 @@
+#!/bin/sh
+
+cp syntax.txt syntax.html
+sam -d syntax.html <<'!'
+,s/\&/\&amp;/g
+,s/</\&lt;/g
+,s/>/\&gt;/g
+,s!== (([^()]|\([^()]*\))*)!≡ <code>\1</code>!g
+,s!«!<code>!g
+,s!»!</code>!g
+,s! vim$! <font size=-2>VIM</font>!g
+,s! pcre$! <font size=-2>PCRE</font>!g
+,s! perl$! <font size=-2>PERL</font>!g
+,x g/NOT SUPPORTED/ s!^[^	]+!<font color=#808080>&</font>!
+,s!NOT SUPPORTED!!g
+,s!(^[^	]+)	(.*)\n!<tr><td><code>\1</code></td><td>\2</td></tr>\n!g
+,s!.*:$!<b>&</b>!g
+,s!^$!<tr><td></td></tr>!g
+,x v/<tr>/ s!.*!<tr><td colspan=2>&</td></tr>!
+1,2c
+<html>
+<!-- AUTOMATICALLY GENERATED by mksyntaxhtml -->
+<head>
+<meta http-equiv="Content-Type" content="text/html; charset=UTF-8"/>
+<title>RE2 regular expression syntax reference</title>
+</head>
+<body>
+<h1>RE2 regular expression syntax reference</h1>
+
+<table border=0 cellpadding=2 cellspacing=2>
+<tr><td colspan=2>This page lists the regular expression syntax accepted by RE2.</td></tr>
+<tr><td colspan=2>It also lists syntax accepted by PCRE, PERL, and VIM.</td></tr>
+<tr><td colspan=2>Grayed out expressions are not supported by RE2.</td></tr>
+.
+$a
+</table>
+</body>
+</html>
+.
+w
+q
+!

package/vendor/re2/doc/mksyntaxwiki

@@ -0,0 +1,36 @@
+#!/bin/sh
+
+cp syntax.txt syntax.wiki
+sam -d syntax.wiki <<'!'
+,s!`!`````!g
+,s!== (([^()]|\([^()]*\))*)!≡ `\1`!g
+,s!«!`!g
+,s!»!`!g
+,s! vim$! <font size="1">VIM</font>!g
+,s! pcre$! <font size="1">PCRE</font>!g
+,s! perl$! <font size="1">PERL</font>!g
+,s!(^[^	]+)	(.*)\n!`\1`	\2\n!g
+,x g/NOT SUPPORTED/ s!^[^	]+!<font color="#808080">&</font>!
+,s!NOT SUPPORTED!<font size="1">(&)</font>!g
+,s!(^[^	]+)	(.*)\n!<tr><td>\1</td><td>\2</td></tr>\n!g
+,s!.*:$!<b>&</b>!g
+,s!^$!<tr><td></td></tr>!g
+,x v/<tr>/ s!.*!<tr><td colspan="2">&</td></tr>!
+1,2c
+#summary I define UNIX as “30 definitions of regular expressions living under one roof.” —Don Knuth
+
+<wiki:comment>
+GENERATED BY mksyntaxwiki.  DO NOT EDIT
+</wiki:comment>
+
+<table border="0" cellpadding="2" cellspacing="2">
+<tr><td colspan="2">This page lists the regular expression syntax accepted by RE2.</td></tr>
+<tr><td colspan="2">It also lists syntax accepted by PCRE, PERL, and VIM.</td></tr>
+<tr><td colspan="2">Grayed out expressions are not supported by RE2.</td></tr>
+.
+$a
+</table>
+.
+w
+q
+!