#!/usr/bin/env perl
# -*- Mode: perl; indent-tabs-mode: nil; c-basic-offset: 2  -*-

#
#  Copyright (C) 2000, 2001 Eazel, Inc.
#  Copyright (C) 2002-2007, 2015 Apple Inc.  All rights reserved.
#  Copyright (C) 2009 Torch Mobile, Inc.
#  Copyright (C) 2009 Cameron McCormack <cam@mcc.id.au>
#
#  prepare-ChangeLog is free software; you can redistribute it and/or
#  modify it under the terms of the GNU General Public
#  License as published by the Free Software Foundation; either
#  version 2 of the License, or (at your option) any later version.
#
#  prepare-ChangeLog is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
#  General Public License for more details.
#
#  You should have received a copy of the GNU General Public
#  License along with this program; if not, write to the Free
#  Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
#


# Perl script to create a ChangeLog entry with names of files
# and functions from a diff.
#
# Darin Adler <darin@bentspoon.com>, started 20 April 2000
# Java support added by Maciej Stachowiak <mjs@eazel.com>
# Objective-C, C++ and Objective-C++ support added by Maciej Stachowiak <mjs@apple.com>
# Git support added by Adam Roben <aroben@apple.com>
# --git-index flag added by Joe Mason <joe.mason@torchmobile.com>


#
# TODO:
#   Decide what a good logical order is for the changed files
#     other than a normal text "sort" (top level first?)
#     (group directories?) (.h before .c?)
#   Handle yacc source files too (other languages?).
#   Help merge when there are ChangeLog conflicts or if there's
#     already a partly written ChangeLog entry.
#   Add command line option to put the ChangeLog into a separate file.
#   Add SVN version numbers for commit (can't do that until
#     the changes are checked in, though).
#   Work around diff stupidity where deleting a function that starts
#     with a comment makes diff think that the following function
#     has been changed (if the following function starts with a comment
#     with the same first line, such as /**)
#   Work around diff stupidity where deleting an entire function and
#     the blank lines before it makes diff think you've changed the
#     previous function.

use strict;
use warnings;

use File::Basename;
use File::Spec;
use FindBin;
use Getopt::Long;
use lib $FindBin::Bin;
use List::Util qw/max/;
use POSIX qw(strftime);
use VCSUtils;

sub actuallyGenerateFunctionLists($$$$$$);
sub attributeCommand($$);
sub changeLogDate($);
sub changeLogEmailAddressFromArgs($$);
sub changeLogNameFromArgs($$);
sub computeModifiedFunctions($$$);
sub createPatchCommand($$$$);
sub decodeEntities($);
sub determinePropertyChanges($$$);
sub diffCommand($$$$);
sub diffFromToString($$$);
sub extractLineRangeAfterChange($);
sub extractLineRangeBeforeChange($);
sub fetchBugXMLData($$);
sub fetchBugDescriptionFromBugXMLData($$$);
sub fetchRadarURLFromBugXMLData($$);
sub findChangeLogs($$);
sub findOriginalFileFromSvn($);
sub generateFileList(\%$$$);
sub generateFunctionLists($$$$$);
sub generateNewChangeLogs($$$$$$$$$$$$$$);
sub getLatestChangeLogs($);
sub get_function_line_ranges($$);
sub get_function_line_ranges_for_cpp($$);
sub delete_namespaces_from_ranges_for_cpp(\@\@);
sub is_function_in_namespace($$);
sub get_function_line_ranges_for_java($$);
sub get_function_line_ranges_for_javascript($$);
sub get_function_line_ranges_for_perl($$);
sub get_selector_line_ranges_for_css($$);
sub get_function_line_ranges_for_swift($$);
sub parseSwiftFunctionArgs($);
sub isAddedStatus($);
sub isConflictStatus($$$);
sub isModifiedStatus($);
sub isUnmodifiedStatus($);
sub main();
sub method_decl_to_selector($);
sub normalizeLineEndings($$);
sub openChangeLogs($);
sub originalFile($$$$);
sub pluralizeAndList($$@);
sub printDiff($$$$);
sub processPaths(\@);
sub propertyChangeDescription($);
sub resolveChangeLogsPath($@);
sub resolveConflictedChangeLogs($);
sub reviewerAndDescriptionForGitCommit($$);
sub statusCommand($$$$);
sub statusDescription($$$$);
sub svnUpdateCommand(@);
sub testListForChangeLog(@);

### Constant variables.
# Project time zone for Cupertino, CA, US
use constant ChangeLogTimeZone => "PST8PDT";
use constant SVN => "svn";
use constant GIT => "git";
use constant SupportedTestExtensions => {map { $_ => 1 } qw(html shtml svg xml xhtml pl php)};

my $devNull = File::Spec->devnull();
my %attributeCache;

exit(main());

sub main()
{
    my $bugDescription;
    my $bugRadarURL;
    my $bugNumber;
    my $name;
    my $emailAddress;
    my $mergeBase = 0;
    my $gitCommit = 0;
    my $gitIndex = "";
    my $gitReviewer = "";
    my $checkWebKitStyle = 0;
    my $openChangeLogs = 0;
    my $writeChangeLogs = 1;
    my $delimiters = 0;
    my $showHelp = 0;
    my $spewDiff = $ENV{"PREPARE_CHANGELOG_DIFF"};
    my $updateChangeLogs = 1;
    my $parseOptionsResult =
        GetOptions("diff|d!" => \$spewDiff,
                   "bug|b:i" => \$bugNumber,
                   "delimiters" => \$delimiters,
                   "description:s" => \$bugDescription,
                   "name:s" => \$name,
                   "email:s" => \$emailAddress,
                   "merge-base:s" => \$mergeBase,
                   "git-commit|g:s" => \$gitCommit,
                   "git-index" => \$gitIndex,
                   "git-reviewer:s" => \$gitReviewer,
                   "help|h!" => \$showHelp,
                   "style!" => \$checkWebKitStyle,
                   "open|o!" => \$openChangeLogs,
                   "write!" => \$writeChangeLogs,
                   "update!" => \$updateChangeLogs);
    if (!$parseOptionsResult || $showHelp) {
        print STDERR basename($0) . " [-b|--bug=<bugid>] [-d|--diff] [-h|--help] [-o|--open] [-g|--git-commit=<committish>] [--git-reviewer=<name>] [svndir1 [svndir2 ...]]\n";
        print STDERR "  -b|--bug        Fill in the ChangeLog bug information from the given bug.\n";
        print STDERR "  --description   One-line description that matches the bug title.\n";
        print STDERR "  -d|--diff       Spew diff to stdout when running\n";
        print STDERR "  --merge-base    Populate the ChangeLogs with the diff to this branch\n";
        print STDERR "  -g|--git-commit Populate the ChangeLogs from the specified git commit\n";
        print STDERR "  --git-index     Populate the ChangeLogs from the git index only\n";
        print STDERR "  --git-reviewer  When populating the ChangeLogs from a git commit claim that the spcified name reviewed the change.\n";
        print STDERR "                  This option is useful when the git commit lacks a Signed-Off-By: line\n";
        print STDERR "  -h|--help       Show this help message\n";
        print STDERR "  --[no-]style    Run check-webkit-style script when done (default: no-style)\n";
        print STDERR "  -o|--open       Open ChangeLogs in an editor when done\n";
        print STDERR "  --[no-]update   Update ChangeLogs from svn before adding entry (default: update)\n";
        print STDERR "  --[no-]write    Write ChangeLogs to disk (otherwise send new entries to stdout) (default: write)\n";
        print STDERR "  --delimiters    When writing to stdout, label and print a \"~\" after each entry\n";
        print STDERR "  --email=        Specify the email address to be used in the patch\n";
        return 1;
    }

    if ($checkWebKitStyle) {
        print STDERR "  Running check-webkit-style.\n  ";
        system "$FindBin::Bin/check-webkit-style";
    }

    die "--git-commit and --git-index are incompatible." if ($gitIndex && $gitCommit);

    isSVN() || isGit() || die "Couldn't determine your version control system.";

    my %paths = processPaths(@ARGV);

    # Find the list of modified files
    my ($changedFiles, $conflictFiles, $functionLists, $addedRegressionTests, $requiresTests) = generateFileList(%paths, $gitCommit, $gitIndex, $mergeBase);

    if (!@$changedFiles && !@$conflictFiles && !keys %$functionLists) {
        print STDERR "  No changes found.\n";
        return 1;
    }

    if (@$conflictFiles) {
        print STDERR "  The following files have conflicts. Run prepare-ChangeLog again after fixing the conflicts:\n";
        print STDERR join("\n", @$conflictFiles), "\n";
        return 1;
    }

    generateFunctionLists($changedFiles, $functionLists, $gitCommit, $gitIndex, $mergeBase);

    # Get some parameters for the ChangeLog we are about to write.
    $name = changeLogNameFromArgs($name, $gitCommit);
    $emailAddress = changeLogEmailAddressFromArgs($emailAddress, $gitCommit);

    print STDERR "  Change author: $name <$emailAddress>.\n";

    # Remove trailing parenthesized notes from user name (bit of hack).
    $name =~ s/\(.*?\)\s*$//g;

    my $bugURL;
    if ($bugNumber) {
        $bugURL = "https://bugs.webkit.org/show_bug.cgi?id=$bugNumber";
    }

    if ($bugNumber && !$bugDescription) {
        my $bugXMLData = fetchBugXMLData($bugURL, $bugNumber);
        $bugDescription = fetchBugDescriptionFromBugXMLData($bugURL, $bugNumber, $bugXMLData);
        $bugRadarURL = fetchRadarURLFromBugXMLData($bugNumber, $bugXMLData);
    }

    my ($filesInChangeLog, $prefixes) = findChangeLogs($functionLists, $writeChangeLogs);

    # Get the latest ChangeLog files from svn.
    my $changeLogs = getLatestChangeLogs($prefixes);

    if (@$changeLogs && $updateChangeLogs && isSVN()) {
        resolveConflictedChangeLogs($changeLogs);
    }

    generateNewChangeLogs($prefixes, $filesInChangeLog, $addedRegressionTests, $requiresTests, $functionLists, $bugURL, $bugDescription, $bugRadarURL, $name, $emailAddress, $gitReviewer, $gitCommit, $writeChangeLogs, $delimiters);

    if ($writeChangeLogs) {
        print STDERR "-- Please remember to include a detailed description in your ChangeLog entry. --\n-- See <http://webkit.org/coding/contributing.html> for more info --\n";
    }

    # Write out another diff.
    if ($spewDiff && @$changedFiles) {
        printDiff($changedFiles, $gitCommit, $gitIndex, $mergeBase);
    }

    # Open ChangeLogs.
    if ($openChangeLogs && @$changeLogs) {
        openChangeLogs($changeLogs);
    }
    return 0;
}

sub originalFile($$$$)
{
    my ($file, $gitCommit, $gitIndex, $mergeBase) = @_;

    my $command;
    if (isSVN()) {
        my $escapedPathsString = escapeSubversionPath($file);
        $command = SVN . " cat $escapedPathsString";
    } elsif (isGit()) {
        $command = GIT . " show ";
        if ($mergeBase) {
            $command .= "$mergeBase";
        } else {
            $command .= "HEAD";
        }
        $command .= ":$file";
    }

    return $command;
}

sub generateFunctionLists($$$$$)
{
    my ($changedFiles, $functionLists, $gitCommit, $gitIndex, $mergeBase) = @_;
    my %delegateHash = (
        openDiff => sub ($$$$) {
            my ($changedFiles, $gitCommit, $gitIndex, $mergeBase) = @_;
            return unless open(DIFF, "-|", diffCommand($changedFiles, $gitCommit, $gitIndex, $mergeBase));
            return \*DIFF;
        },
        openFile => sub ($) {
            my ($file) = @_;
            return unless open(SOURCE, "<", $file);
            return \*SOURCE;
        },
        openOriginalFile => sub ($) {
            my ($file, $gitCommit, $gitIndex, $mergeBase) = @_;
            return unless open(SOURCE, "-|", originalFile($file, $gitCommit, $gitIndex, $mergeBase));
            return \*SOURCE;
        },
        normalizePath => sub ($) {
            my ($path) = @_;
            return normalizePath(makeFilePathRelative($path));
        },
    );
    actuallyGenerateFunctionLists($changedFiles, $functionLists, $gitCommit, $gitIndex, $mergeBase, \%delegateHash);
}

sub actuallyGenerateFunctionLists($$$$$$)
{
    my ($changedFiles, $functionLists, $gitCommit, $gitIndex, $mergeBase, $delegateHashRef) = @_;

    my %line_ranges_after_changed;
    my %line_ranges_before_changed;
    if (@$changedFiles) {
        # For each file, build a list of modified lines.
        # Use line numbers from the "after" side of each diff.
        print STDERR "  Reviewing diff to determine which lines changed.\n";
        my $file;
        my $diffFileHandle = $delegateHashRef->{openDiff}($changedFiles, $gitCommit, $gitIndex, $mergeBase);
        if (!$diffFileHandle) {
            die "The diff failed: $!.\n";
        }
        while (<$diffFileHandle>) {
            my $filePath = parseDiffStartLine($_);
            $file = $delegateHashRef->{normalizePath}($filePath) if $filePath;
            if (defined $file) {
                my ($before_start, $before_end) = extractLineRangeBeforeChange($_);
                if ($before_start >= 1 && $before_end >= 1) {
                    push @{$line_ranges_before_changed{$file}}, [ $before_start, $before_end ];
                } elsif (/DO_NOT_COMMIT/) {
                    print STDERR "WARNING: file $file contains the string DO_NOT_COMMIT, line $.\n";
                }
                my ($after_start, $after_end) = extractLineRangeAfterChange($_);
                if ($after_start >= 1 && $after_end >= 1) {
                    push @{$line_ranges_after_changed{$file}}, [ $after_start, $after_end ];
                } elsif (/DO_NOT_COMMIT/) {
                    print STDERR "WARNING: file $file contains the string DO_NOT_COMMIT, line $.\n";
                }
            }
        }
        close($diffFileHandle);
    }

    # For each source file, convert line range to function list.
    print STDERR "  Extracting affected function names from source files.\n";
    my %filesToExamine = map { $_ => 1 } (keys(%line_ranges_before_changed), keys(%line_ranges_after_changed));
    foreach my $file (keys %filesToExamine) {
        my %saw_function;

        # Find all the functions in the file.
        my $sourceFileHandle = $delegateHashRef->{openFile}($file);
        next unless $sourceFileHandle;
        my @afterChangeFunctionRanges = get_function_line_ranges($sourceFileHandle, $file);
        close($sourceFileHandle);

        # Find modified functions in the file.
        if ($line_ranges_after_changed{$file}) {
            my @change_ranges = (@{$line_ranges_after_changed{$file}}, []);
            my @functions = computeModifiedFunctions($file, \@change_ranges, \@afterChangeFunctionRanges);

            # Format the list of functions.
            if (@functions) {
                $functionLists->{$file} = "" if !defined $functionLists->{$file};
                $functionLists->{$file} .= "\n        (" . join("):\n        (", @functions) . "):";
            }
        }
        # Find the deleted functions in the original file.
        if ($line_ranges_before_changed{$file}) {
            my $originalFileHandle = $delegateHashRef->{openOriginalFile}($file, $gitCommit, $gitIndex, $mergeBase);
            next unless $originalFileHandle;
            my @beforeChangeFunctionRanges = get_function_line_ranges($originalFileHandle, $file);
            close($originalFileHandle);

            my %existsAfterChange = map { $_->[2] => 1 } @afterChangeFunctionRanges;

            my @functions;
            my %sawFunctions;
            for my $functionRange (@beforeChangeFunctionRanges) {
                my $functionName = $functionRange->[2];
                if (!$existsAfterChange{$functionName} && !$sawFunctions{$functionName}) {
                    push @functions, $functionName;
                    $sawFunctions{$functionName} = 1;
                }
            }

            # Format the list of deleted functions.
            if (@functions) {
                $functionLists->{$file} = "" if !defined $functionLists->{$file};
                $functionLists->{$file} .= "\n        (" . join("): Deleted.\n        (", @functions) . "): Deleted.";
            }
        }
    }
}

sub computeModifiedFunctions($$$)
{
    my ($file, $changedLineRanges, $functionRanges) = @_;

    my %sawFunction;

    # Find all the modified functions.
    my @functions;
    my @change_ranges = @{$changedLineRanges};
    my @change_range = (0, 0);
    FUNCTION: foreach my $function_range_ref (@{$functionRanges}) {
        my @function_range = @{$function_range_ref};

        # FIXME: This is a hack. If the function name is empty, skip it.
        # The cpp, python, javascript, perl, css and java parsers
        # are not perfectly implemented and sometimes function names cannot be retrieved
        # correctly. As you can see in get_function_line_ranges_XXXX(), those parsers
        # are not intended to implement real parsers but intended to just retrieve function names
        # for most practical syntaxes.
        next unless $function_range[2];

        # Advance to successive change ranges.
        for (;; @change_range = @{shift @change_ranges}) {
            last FUNCTION unless @change_range;

            # If past this function, move on to the next one.
            next FUNCTION if $change_range[0] > $function_range[1];

            # If an overlap with this function range, record the function name.
            if ($change_range[1] >= $function_range[0]
                and $change_range[0] <= $function_range[1]) {
                if (!$sawFunction{$function_range[2]}) {
                    $sawFunction{$function_range[2]} = 1;
                    push @functions, $function_range[2];
                }
                next FUNCTION;
            }
        }
    }

    return @functions;
}

sub changeLogDate($)
{
    my ($timeZone) = @_;
    my $savedTimeZone = $ENV{'TZ'};
    # Set TZ temporarily so that localtime() is in that time zone
    $ENV{'TZ'} = $timeZone;
    my $date = strftime("%Y-%m-%d", localtime());
    if (defined $savedTimeZone) {
         $ENV{'TZ'} = $savedTimeZone;
    } else {
         delete $ENV{'TZ'};
    }
    return $date;
}

sub changeLogNameFromArgs($$)
{
    my ($nameFromArgs, $gitCommit) = @_;
    # Silently allow --git-commit to win, we could warn if $nameFromArgs is defined.
    my $command = GIT . ' log --max-count=1 --pretty="format:%an" "' . $gitCommit . '"';
    return `$command` if $gitCommit;

    return $nameFromArgs || changeLogName();
}

sub changeLogEmailAddressFromArgs($$)
{
    my ($emailAddressFromArgs, $gitCommit) = @_;
    # Silently allow --git-commit to win, we could warn if $emailAddressFromArgs is defined.
    my $command = GIT . ' log --max-count=1 --pretty="format:%ae" "' . $gitCommit . '"';
    return `$command` if $gitCommit;

    return $emailAddressFromArgs || changeLogEmailAddress();
}

sub fetchBugXMLData($$)
{
    my ($bugURL, $bugNumber) = @_;

    my $bugXMLURL = "$bugURL&ctype=xml&excludefield=attachmentdata";
    # Perl has no built in XML processing, so we'll fetch and parse with curl and grep
    # Pass --insecure because some cygwin installs have no certs we don't
    # care about validating that bugs.webkit.org is who it says it is here.
    my $xmlData = `curl --insecure --silent "$bugXMLURL"`;
    if ($xmlData !~ /<\?xml/) {
        # Maybe the reason the above did not work is because the curl that is installed doesn't
        # support ssl at all.
        if (`curl --version | grep ^Protocols` !~ /\bhttps\b/) {
            print STDERR "  Could not get description for bug $bugNumber.\n";
            print STDERR "  It looks like your version of curl does not support ssl.\n";
            print STDERR "  If you are using macports, this can be fixed with sudo port install curl +ssl.\n";
        }
        exit 1;
    }
    return $xmlData;
}

sub fetchBugDescriptionFromBugXMLData($$$)
{
    my ($bugURL, $bugNumber, $bugXMLData) = @_;

    if ($bugXMLData !~ /<short_desc>(.*)<\/short_desc>/) {
        print STDERR "  Bug $bugNumber has no bug description. Maybe you set wrong bug ID?\n";
        print STDERR "  The bug URL: $bugURL\n";
        exit 1;
    }

    my $bugDescription = decodeEntities($1);
    print STDERR "  Description from bug $bugNumber:\n    \"$bugDescription\".\n";
    return $bugDescription;
}

sub fetchRadarURLFromBugXMLData($$)
{
    my ($bugNumber, $bugXMLData) = @_;

    return "" if $bugXMLData !~ m|<thetext>\s*(&lt;rdar://problem/\d+&gt;)|;

    my $bugRadarURL = decodeEntities($1);
    print STDERR "  Radar URL from bug $bugNumber:\n    \"$bugRadarURL\".\n";
    return $bugRadarURL;
}

sub findChangeLogs($$)
{
    my ($functionLists, $requireChangeLogToExist) = @_;

    # Find the change logs.
    my %has_log;
    my %filesInChangeLog;
    foreach my $file (sort keys %$functionLists) {
        my $prefix = $file;
        my $has_log = 0;
        while ($prefix) {
            if ($^O eq "MSWin32") {
                $prefix =~ s-\\[^\\]+\\?$-\\- or $prefix = "";
            } else {
                $prefix =~ s-/[^/]+/?$-/- or $prefix = "";
            }
            $has_log = $has_log{$prefix};
            if (!defined $has_log) {
                $has_log = -f "${prefix}ChangeLog";
                $has_log{$prefix} = $has_log;
            }
            last if $has_log;
        }
        if (!$has_log && $requireChangeLogToExist) {
            print STDERR "No ChangeLog found for $file.\n";
        } else {
            push @{$filesInChangeLog{$prefix}}, $file;
        }
    }

    # Build the list of ChangeLog prefixes in the correct project order
    my @prefixes;
    my %prefixesSort;
    foreach my $prefix (keys %filesInChangeLog) {
        my $prefixDir = substr($prefix, 0, length($prefix) - 1); # strip trailing /
        my $sortKey = lc $prefix;
        $sortKey = "top level" unless length $sortKey;

        if ($prefixDir eq "top level") {
            $sortKey = "";
        } elsif ($prefixDir eq "Tools") {
            $sortKey = "-, just after top level";
        } elsif ($prefixDir eq "WebBrowser") {
            $sortKey = lc "WebKit, WebBrowser after";
        } elsif ($prefixDir eq "Source/WebCore") {
            $sortKey = lc "WebFoundation, WebCore after";
        } elsif ($prefixDir eq "LayoutTests") {
            $sortKey = lc "~, LayoutTests last";
        }

        $prefixesSort{$sortKey} = $prefix;
    }
    foreach my $prefixSort (sort keys %prefixesSort) {
        push @prefixes, $prefixesSort{$prefixSort};
    }
    return (\%filesInChangeLog, \@prefixes);
}

sub getLatestChangeLogs($)
{
    my ($prefixes) = @_;

    my @changeLogs = ();
    foreach my $prefix (@$prefixes) {
        push @changeLogs, File::Spec->catfile($prefix || ".", "ChangeLog");
    }
    return \@changeLogs;
}

sub svnUpdateCommand(@)
{
    my @changeLogs = shift;

    my @escapedChangeLogPaths = map(escapeSubversionPath($_), @changeLogs);
    my $escapedChangeLogPathsString = qq(") . join(qq(" "), @escapedChangeLogPaths) . qq(");
    my $command = SVN . " update $escapedChangeLogPathsString";

    return $command;
}

sub resolveChangeLogsPath($@)
{
    my ($resolveChangeLogsPath, @conflictedChangeLogs) = @_;

    my @escapedConflictedChangeLogs = map(escapeSubversionPath($_), @conflictedChangeLogs);
    my $escapedConflictedChangeLogsString = qq(") . join(qq(" "), @escapedConflictedChangeLogs) . qq(");
    my $command = "$resolveChangeLogsPath --no-warnings $escapedConflictedChangeLogsString";

    return $command;
}

sub resolveConflictedChangeLogs($)
{
    my ($changeLogs) = @_;

    print STDERR "  Running 'svn update' to update ChangeLog files.\n";
    open ERRORS, "-|", svnUpdateCommand(@$changeLogs)
        or die "The svn update of ChangeLog files failed: $!.\n";
    my @conflictedChangeLogs;
    while (my $line = <ERRORS>) {
        print STDERR "    ", $line;
        push @conflictedChangeLogs, $1 if $line =~ m/^C\s+(.+?)[\r\n]*$/;
    }
    close ERRORS;

    return if !@conflictedChangeLogs;

    print STDERR "  Attempting to merge conflicted ChangeLogs.\n";
    my $resolveChangeLogsPath = File::Spec->catfile(dirname($0), "resolve-ChangeLogs");
    open RESOLVE, "-|", resolveChangeLogsPath($resolveChangeLogsPath, @conflictedChangeLogs)
        or die "Could not open resolve-ChangeLogs script: $!.\n";
    print STDERR "    $_" while <RESOLVE>;
    close RESOLVE;
}

sub generateNewChangeLogs($$$$$$$$$$$$$$)
{
    my ($prefixes, $filesInChangeLog, $addedRegressionTests, $requiresTests, $functionLists, $bugURL, $bugDescription, $bugRadarURL, $name, $emailAddress, $gitReviewer, $gitCommit, $writeChangeLogs, $delimiters) = @_;

    my $hasWebCoreChange = 0;
    foreach my $prefix (@$prefixes) {
        if (unixPath($prefix) =~ m|/WebCore/$|) {
            $hasWebCoreChange = 1;
        }
    }

    # Generate new ChangeLog entries and (optionally) write out new ChangeLog files.
    foreach my $prefix (@$prefixes) {
        my $endl = "\n";
        my @old_change_log;

        if ($writeChangeLogs) {
            my $changeLogPath = File::Spec->catfile($prefix || ".", "ChangeLog");
            print STDERR "  Editing the ${changeLogPath} file.\n";
            open OLD_CHANGE_LOG, ${changeLogPath} or die "Could not open ${changeLogPath} file: $!.\n";
            # It's less efficient to read the whole thing into memory than it would be
            # to read it while we prepend to it later, but I like doing this part first.
            @old_change_log = <OLD_CHANGE_LOG>;
            close OLD_CHANGE_LOG;
            # We want to match the ChangeLog's line endings in case it doesn't match
            # the native line endings for this version of perl.
            if ($old_change_log[0] =~ /(\r?\n)$/g) {
                $endl = "$1";
            }
            open CHANGE_LOG, "> ${changeLogPath}" or die "Could not write ${changeLogPath}\n.";
            binmode(CHANGE_LOG);
        } else {
            open CHANGE_LOG, ">-" or die "Could not write to STDOUT\n.";
            print substr($prefix, 0, length($prefix) - 1) . ":\n\n" unless (scalar @$prefixes) == 1 && !$delimiters;
        }

        my $date = changeLogDate(ChangeLogTimeZone);
        print CHANGE_LOG normalizeLineEndings("$date  $name  <$emailAddress>\n\n", $endl);

        my ($reviewer, $description) = reviewerAndDescriptionForGitCommit($gitCommit, $gitReviewer) if $gitCommit;
        $reviewer = "NOBODY (OO" . "PS!)" if !$reviewer;

        ($bugDescription, $description) =
            ($description =~ /^(?:\s*(.*)\n)?(?:\s*\n)*((?:\n|.)*)/)
            if !$bugDescription && $description;

        $bugDescription = "Need a short description (OOPS!)." unless $bugDescription;
        $bugURL = "Need the bug URL (OOPS!)." unless $bugURL;

        print CHANGE_LOG normalizeLineEndings("        $bugDescription\n", $endl) if $bugDescription;
        print CHANGE_LOG normalizeLineEndings("        $bugURL\n", $endl) if $bugURL;
        print CHANGE_LOG normalizeLineEndings("        $bugRadarURL\n", $endl) if $bugRadarURL;
        print CHANGE_LOG normalizeLineEndings("\n", $endl);

        print CHANGE_LOG normalizeLineEndings("        Reviewed by $reviewer.\n\n", $endl);
        print CHANGE_LOG normalizeLineEndings($description . "\n", $endl) if $description;

        my $shouldMentionTests = @$requiresTests;
        $shouldMentionTests |= !$hasWebCoreChange && unixPath($prefix) =~ m|/WebKit/$|;
        $shouldMentionTests |= unixPath($prefix) =~ m|/WebCore/$|;

        if ($shouldMentionTests) {
            if (@$addedRegressionTests) {
                print CHANGE_LOG normalizeLineEndings(testListForChangeLog(sort @$addedRegressionTests), $endl);
            } else {
                print CHANGE_LOG normalizeLineEndings("        No new tests (OOPS!).\n\n", $endl);
            }
        }

        foreach my $file (sort @{$filesInChangeLog->{$prefix}}) {
            my $file_stem = substr $file, length $prefix;
            $file_stem = unixPath($file_stem);
            print CHANGE_LOG normalizeLineEndings("        * $file_stem:$functionLists->{$file}\n", $endl);
        }

        if ($writeChangeLogs) {
            print CHANGE_LOG normalizeLineEndings("\n", $endl), @old_change_log;
        } else {
            print CHANGE_LOG "\n";
            print "~\n"  if $delimiters;
        }

        close CHANGE_LOG;
    }
}

sub printDiff($$$$)
{
    my ($changedFiles, $gitCommit, $gitIndex, $mergeBase) = @_;

    print STDERR "  Running diff to help you write the ChangeLog entries.\n";
    local $/ = undef; # local slurp mode
    my $changedFilesString = "'" . join("' '", @$changedFiles) . "'";
    open DIFF, "-|", createPatchCommand($changedFilesString, $gitCommit, $gitIndex, $mergeBase) or die "The diff failed: $!.\n";
    print <DIFF>;
    close DIFF;
}

sub openChangeLogs($)
{
    my ($changeLogs) = @_;

    print STDERR "  Opening the edited ChangeLog files.\n";
    my $editor = $ENV{CHANGE_LOG_EDITOR} || $ENV{VISUAL} || $ENV{EDITOR};
    if ($editor) {
        system ((split ' ', $editor), @$changeLogs);
    } else {
        $editor = $ENV{CHANGE_LOG_EDIT_APPLICATION};
        if ($editor) {
            system "open", "-a", $editor, @$changeLogs;
        } else {
            system "open", "-e", @$changeLogs;
        }
    }
}

sub get_function_line_ranges($$)
{
    my ($file_handle, $file_name) = @_;

    # Try to determine the source language based on the file extension.

    return get_function_line_ranges_for_cpp($file_handle, $file_name) if $file_name =~ /\.(c|cpp|m|mm|h)$/;
    return get_function_line_ranges_for_java($file_handle, $file_name) if $file_name =~ /\.java$/;
    return get_function_line_ranges_for_javascript($file_handle, $file_name) if $file_name =~ /\.js$/;
    return get_selector_line_ranges_for_css($file_handle, $file_name) if $file_name =~ /\.css$/;
    return get_function_line_ranges_for_perl($file_handle, $file_name) if $file_name =~ /\.p[lm]$/;
    return get_function_line_ranges_for_python($file_handle, $file_name) if $file_name =~ /\.py$/ or $file_name =~ /master\.cfg$/;
    return get_function_line_ranges_for_swift($file_handle, $file_name) if $file_name =~ /\.swift$/;

    # Try to determine the source language based on the script interpreter.

    my $first_line = <$file_handle>;
    seek($file_handle, 0, 0);

    return () unless $first_line =~ m|^#!(?:/usr/bin/env\s+)?(\S+)|;
    my $interpreter = $1;

    return get_function_line_ranges_for_perl($file_handle, $file_name) if $interpreter =~ /perl$/;
    return get_function_line_ranges_for_python($file_handle, $file_name) if $interpreter =~ /python$/;

    return ();
}


sub method_decl_to_selector($)
{
    (my $method_decl) = @_;

    $_ = $method_decl;

    if ((my $comment_stripped) = m-([^/]*)(//|/*).*-) {
        $_ = $comment_stripped;
    }

    s/,\s*...//;

    # Strip out the return type and parameter types. The extra )? takes care of most block parameter types.
    s/\([^\)]*\)\)?//g;

    if (/:/) {
        my @components = split /:/;
        pop @components if (scalar @components > 1);
        $_ = (join ':', map {s/.*[^[:word:]]//; scalar $_;} @components) . ':';
    } else {
        s/\s*$//;
        s/.*[^[:word:]]//;
    }

    return $_;
}


# Read a file and get all the line ranges of the things that look like C functions.
# A function name is the last word before an open parenthesis before the outer
# level open brace. A function starts at the first character after the last close
# brace or semicolon before the function name and ends at the close brace.
# Comment handling is simple-minded but will work for all but pathological cases.
#
# Result is a list of triples: [ start_line, end_line, function_name ].

sub get_function_line_ranges_for_cpp($$)
{
    my ($file_handle, $file_name) = @_;

    my @ranges;

    my $in_comment = 0;
    my $in_macro = 0;
    my $in_method_declaration = 0;
    my $in_parentheses = 0;
    my $quotation_mark;
    my $in_braces = 0;
    my $in_toplevel_array_brace = 0;
    my $brace_start = 0;
    my $brace_end = 0;
    my $namespace_start = -1;
    my $skip_til_brace_or_semicolon = 0;
    my $equal_observed = 0;

    my $word = "";
    my $interface_name = "";

    my $potential_method_char = "";
    my $potential_method_spec = "";

    my $potential_start = 0;
    my $potential_name = "";

    my $start = 0;
    my $name = "";

    my $next_word_could_be_namespace = 0;
    my $potential_namespace = "";
    my @namespaces;
    my @all_namespaces;

    while (<$file_handle>) {
        # Handle continued quoted string.
        if ($quotation_mark) {
            if (!s-([^\\]|\\.)*$quotation_mark--) {
                if (!m-\\$-) {
                    warn "mismatched quotes at line $. in $file_name\n";
                    undef $quotation_mark;
                }
                next;
            }
            undef $quotation_mark;
        }

        # Handle continued multi-line comment.
        if ($in_comment) {
            next unless s-.*\*/--;
            $in_comment = 0;
        }

        # Handle continued macro.
        if ($in_macro) {
            $in_macro = 0 unless /\\$/;
            next;
        }

        # Handle start of macro (or any preprocessor directive).
        if (/^\s*\#/) {
            $in_macro = 1 if /^([^\\]|\\.)*\\$/;
            next;
        }

        # Handle comments and quoted text.
        while (m-(/\*|//|\'|\")-) { # \' and \" keep emacs perl mode happy
            my $match = $1;
            if ($match eq "/*") {
                if (!s-/\*.*?\*/--) {
                    s-/\*.*--;
                    $in_comment = 1;
                }
            } elsif ($match eq "//") {
                s-//.*--;
            } else { # ' or "
                if (!s-$match([^\\]|\\.)*?$match--) {
                    if (!s-$match.*\\$--) {
                        warn "mismatched quotes at line $. in $file_name\n";
                        s-$match.*--;
                    } else {
                        $quotation_mark = $match;
                    }
                }
            }
        }


        # continued method declaration
        if ($in_method_declaration) {
              my $original = $_;
              my $method_cont = $_;

              chomp $method_cont;
              $method_cont =~ s/[;\{].*//;
              $potential_method_spec = "${potential_method_spec} ${method_cont}";

              $_ = $original;
              if (/;/) {
                  $potential_start = 0;
                  $potential_method_spec = "";
                  $potential_method_char = "";
                  $in_method_declaration = 0;
                  s/^[^;\{]*//;
              } elsif (/{/) {
                  my $selector = method_decl_to_selector ($potential_method_spec);
                  $potential_name = "${potential_method_char}\[${interface_name} ${selector}\]";

                  $potential_method_spec = "";
                  $potential_method_char = "";
                  $in_method_declaration = 0;

                  $_ = $original;
                  s/^[^;{]*//;
              } elsif (/\@end/) {
                  $in_method_declaration = 0;
                  $interface_name = "";
                  $_ = $original;
              } else {
                  next;
              }
        }


        # start of method declaration
        if ((my $method_char, my $method_spec) = m&^([-+])([^0-9;][^;]*);?$&) {
            my $original = $_;

            if ($interface_name) {
                chomp $method_spec;
                $method_spec =~ s/\{.*//;

                $potential_method_char = $method_char;
                $potential_method_spec = $method_spec;
                $potential_start = $.;
                $in_method_declaration = 1;
            } else { 
                warn "declaring a method but don't have interface on line $. in $file_name\n";
            }
            $_ = $original;
            if (/\{/) {
              my $selector = method_decl_to_selector ($potential_method_spec);
              $potential_name = "${potential_method_char}\[${interface_name} ${selector}\]";

              $potential_method_spec = "";
              $potential_method_char = "";
              $in_method_declaration = 0;
              $_ = $original;
              s/^[^{]*//;
            } elsif (/\@end/) {
              $in_method_declaration = 0;
              $interface_name = "";
              $_ = $original;
            } else {
              next;
            }
        }


        # Find function, interface and method names.
        while (m&((?:[[:word:]]+::)*operator(?:[ \t]*\(\)|[^()]*)|[[:word:]<>:~]+|[(){}:;=])|\@(?:implementation|interface|protocol)\s+(\w+)[^{]*&g) {
            # Skip an array definition at the top level.
            # e.g. static int arr[] = { 1, 2, 3 };
            if ($1) {
                if ($1 eq "=" and !$in_parentheses and !$in_braces) {
                    $equal_observed = 1;
                } elsif ($1 eq "{" and $equal_observed) {
                    # This '{' is the beginning of an array definition, not the beginning of a method.
                    $in_toplevel_array_brace = 1;
                    $in_braces++;
                    $equal_observed = 0;
                    next;
                } elsif ($1 !~ /[ \t]/) {
                    $equal_observed = 0;
                }
            }

            # interface name
            if ($2) {
                $interface_name = $2;
                next;
            }

            # Open parenthesis.
            if ($1 eq "(") {
                $potential_name = $word unless $in_parentheses || $skip_til_brace_or_semicolon || grep { $word eq $_ } ("CF_ENUM", "CF_OPTIONS", "NS_ENUM", "NS_OPTIONS");
                $in_parentheses++;
                next;
            }

            # Close parenthesis.
            if ($1 eq ")") {
                $in_parentheses--;
                next;
            }

            if ($1 eq "const" and !$in_parentheses) {
                $potential_name .= " const";
                next;
            }

            if ($1 eq "volatile" and !$in_parentheses) {
                $potential_name .= " volatile";
                next;
            }

            # C++ auto function() -> type
            if ($1 eq ">") {
                $skip_til_brace_or_semicolon = 1 unless ($in_parentheses || $in_braces);
                next;
            }

            # C++ constructor initializers
            if ($1 eq ":") {
                $skip_til_brace_or_semicolon = 1 unless ($in_parentheses || $in_braces);
            }

            # Open brace.
            if ($1 eq "{") {
                $skip_til_brace_or_semicolon = 0;

                if (!$in_braces) {
                    if ($namespace_start >= 0 and $namespace_start < $potential_start) {
                        push @ranges, [ $namespace_start . "", $potential_start - 1, $name ];
                    }

                    if ($potential_namespace) {
                        push @namespaces, $potential_namespace;
                        push @all_namespaces, $potential_namespace;
                        $potential_namespace = "";
                        $name = $namespaces[-1];
                        $namespace_start = $. + 1;
                        next;
                    }

                    # Promote potential name to real function name at the
                    # start of the outer level set of braces (function body?).
                    if ($potential_start) {
                        $start = $potential_start;
                        $name = $potential_name;
                        if (@namespaces && $name && (length($name) < 2 || substr($name,1,1) ne "[")) {
                            $name = join ('::', @namespaces, $name);
                        }
                    }
                }

                $in_method_declaration = 0;

                $brace_start = $. if (!$in_braces);
                $in_braces++;
                next;
            }

            # Close brace.
            if ($1 eq "}") {
                if (!$in_braces && @namespaces) {
                    if ($namespace_start >= 0 and $namespace_start < $.) {
                        push @ranges, [ $namespace_start . "", $. - 1, $name ];
                    }

                    pop @namespaces;
                    if (@namespaces) {
                        $name = $namespaces[-1];
                        $namespace_start = $. + 1;
                    } else {
                        $name = "";
                        $namespace_start = -1;
                    }
                    next;
                }

                $in_braces--;
                $brace_end = $. if (!$in_braces);

                # End of an outer level set of braces.
                # This could be a function body.
                if (!$in_braces and $name) {
                    # This is the end of an array definition at the top level, not the end of a method.
                    if ($in_toplevel_array_brace) {
                        $in_toplevel_array_brace = 0;
                        next;
                    }

                    push @ranges, [ $start, $., $name ];
                    if (@namespaces) {
                        $name = $namespaces[-1];
                        $namespace_start = $. + 1;
                    } else {
                        $name = "";
                        $namespace_start = -1;
                    }
                }

                $potential_start = 0;
                $potential_name = "";
                next;
            }

            # Semicolon.
            if ($1 eq ";") {
                $skip_til_brace_or_semicolon = 0;
                $potential_start = 0;
                $potential_name = "";
                $in_method_declaration = 0;
                next;
            }

            # Ignore "const" method qualifier.
            if ($1 eq "const") {
                next;
            }

            if ($1 eq "namespace" || $1 eq "class" || $1 eq "struct") {
                $next_word_could_be_namespace = 1;
                next;
            }

            # Word.
            $word = $1;
            if (!$skip_til_brace_or_semicolon) {
                if ($next_word_could_be_namespace) {
                    $potential_namespace = $word;
                    $next_word_could_be_namespace = 0;
                } elsif ($potential_namespace) {
                    $potential_namespace = "";
                }

                if (!$in_parentheses) {
                    $potential_start = 0;
                    $potential_name = "";
                }
                if (!$potential_start) {
                    $potential_start = $.;
                    $potential_name = "";
                }
            }
        }
    }

    warn "missing close braces in $file_name (probable start at $brace_start)\n" if ($in_braces > 0);
    warn "too many close braces in $file_name (probable start at $brace_end)\n" if ($in_braces < 0);

    warn "mismatched parentheses in $file_name\n" if $in_parentheses;

    return delete_namespaces_from_ranges_for_cpp(@ranges, @all_namespaces);
}


# Take in references to an array of line ranges for C functions in a given file 
# and an array of namespaces declared in that file and return an updated
# list of line ranges with the namespaces removed.

sub delete_namespaces_from_ranges_for_cpp(\@\@)
{
    my ($ranges, $namespaces) = @_;
    return grep {!is_function_in_namespace($namespaces, $$_[2])} @$ranges;
}


sub is_function_in_namespace($$)
{
    my ($namespaces, $function_name) = @_;
    return grep {$_ eq $function_name} @$namespaces;
}


# Read a file and get all the line ranges of the things that look like Java
# classes, interfaces and methods.
#
# A class or interface name is the word that immediately follows
# `class' or `interface' when followed by an open curly brace and not
# a semicolon. It can appear at the top level, or inside another class
# or interface block, but not inside a function block
#
# A class or interface starts at the first character after the first close
# brace or after the function name and ends at the close brace.
#
# A function name is the last word before an open parenthesis before
# an open brace rather than a semicolon. It can appear at top level or
# inside a class or interface block, but not inside a function block.
#
# A function starts at the first character after the first close
# brace or after the function name and ends at the close brace.
#
# Comment handling is simple-minded but will work for all but pathological cases.
#
# Result is a list of triples: [ start_line, end_line, function_name ].

sub get_function_line_ranges_for_java($$)
{
    my ($file_handle, $file_name) = @_;

    my @current_scopes;

    my @ranges;

    my $in_comment = 0;
    my $in_macro = 0;
    my $in_parentheses = 0;
    my $in_braces = 0;
    my $in_non_block_braces = 0;
    my $class_or_interface_just_seen = 0;
    my $in_class_declaration = 0;

    my $word = "";

    my $potential_start = 0;
    my $potential_name = "";
    my $potential_name_is_class_or_interface = 0;

    my $start = 0;
    my $name = "";
    my $current_name_is_class_or_interface = 0;

    while (<$file_handle>) {
        # Handle continued multi-line comment.
        if ($in_comment) {
            next unless s-.*\*/--;
            $in_comment = 0;
        }

        # Handle continued macro.
        if ($in_macro) {
            $in_macro = 0 unless /\\$/;
            next;
        }

        # Handle start of macro (or any preprocessor directive).
        if (/^\s*\#/) {
            $in_macro = 1 if /^([^\\]|\\.)*\\$/;
            next;
        }

        # Handle comments and quoted text.
        while (m-(/\*|//|\'|\")-) { # \' and \" keep emacs perl mode happy
            my $match = $1;
            if ($match eq "/*") {
                if (!s-/\*.*?\*/--) {
                    s-/\*.*--;
                    $in_comment = 1;
                }
            } elsif ($match eq "//") {
                s-//.*--;
            } else { # ' or "
                if (!s-$match([^\\]|\\.)*?$match--) {
                    warn "mismatched quotes at line $. in $file_name\n";
                    s-$match.*--;
                }
            }
        }

        # Find function names.
        while (m-(\w+|[(){};])-g) {
            # Open parenthesis.
            if ($1 eq "(") {
                if (!$in_parentheses) {
                    $potential_name = $word;
                    $potential_name_is_class_or_interface = 0;
                }
                $in_parentheses++;
                next;
            }

            # Close parenthesis.
            if ($1 eq ")") {
                $in_parentheses--;
                next;
            }

            # Open brace.
            if ($1 eq "{") {
                $in_class_declaration = 0;

                # Promote potential name to real function name at the
                # start of the outer level set of braces (function/class/interface body?).
                if (!$in_non_block_braces
                    and (!$in_braces or $current_name_is_class_or_interface)
                    and $potential_start) {
                    if ($name) {
                          push @ranges, [ $start, ($. - 1),
                                          join ('.', @current_scopes) ];
                    }


                    $current_name_is_class_or_interface = $potential_name_is_class_or_interface;

                    $start = $potential_start;
                    $name = $potential_name;

                    push (@current_scopes, $name);
                } else {
                    $in_non_block_braces++;
                }

                $potential_name = "";
                $potential_start = 0;

                $in_braces++;
                next;
            }

            # Close brace.
            if ($1 eq "}") {
                $in_braces--;

                # End of an outer level set of braces.
                # This could be a function body.
                if (!$in_non_block_braces) {
                    if ($name) {
                        push @ranges, [ $start, $.,
                                        join ('.', @current_scopes) ];

                        pop (@current_scopes);

                        if (@current_scopes) {
                            $current_name_is_class_or_interface = 1;

                            $start = $. + 1;
                            $name =  $current_scopes[$#current_scopes-1];
                        } else {
                            $current_name_is_class_or_interface = 0;
                            $start = 0;
                            $name =  "";
                        }
                    }
                } else {
                    $in_non_block_braces-- if $in_non_block_braces;
                }

                $potential_start = 0;
                $potential_name = "";
                next;
            }

            # Semicolon.
            if ($1 eq ";") {
                $potential_start = 0;
                $potential_name = "";
                next;
            }

            if ($1 eq "class") {
                $in_class_declaration = 1;
            }
            if ($1 eq "class" or (!$in_class_declaration and $1 eq "interface")) {
                $class_or_interface_just_seen = 1;
                next;
            }

            # Word.
            $word = $1;
            if (!$in_parentheses) {
                if ($class_or_interface_just_seen) {
                    $potential_name = $word;
                    $potential_start = $.;
                    $class_or_interface_just_seen = 0;
                    $potential_name_is_class_or_interface = 1;
                    next;
                }
            }
            if (!$potential_start) {
                $potential_start = $.;
                $potential_name = "";
            }
            $class_or_interface_just_seen = 0;
        }
    }

    warn "mismatched braces in $file_name\n" if $in_braces;
    warn "mismatched parentheses in $file_name\n" if $in_parentheses;

    return @ranges;
}


# Read a file and get all the line ranges of the things that look like
# JavaScript functions or methods.
#
# A function name is the word that immediately follows `function' when
# followed by an open curly brace. It can appear at the top level,
# or inside other functions. For example:
#
#    function name() { // (name)
#        function inner() { } // (name.inner)
#    }
#
# An anonymous function name is the identifier on the left hand side of
# an assignment with the equals operator or object notation that has a
# value starting with `function' followed an open curly brace.
# For example:
#
#    namespace = {
#        name: function() {} // (namespace.name)
#    }
#    namespace.Foo = function() {} // (namespace.Foo)
#
# A getter or setter name is the word that immediately follows `get' or
# `set' when followed by params and an open curly brace. For example:
#
#    namespace = {
#      get foo() {} // (namespace.get foo)
#    }
#
# A method name is the word immediately before parenthesis, with an open
# curly brace immediately following closing parenthesis. For a class expression
# we take the assignment identifier instead of the class name for namespacing.
#
#    namespace.Foo = class DoesNotMatter extends Bar {
#        constructor() {} // (namespace.Foo)
#        static staticMethod() {} // (namespace.Foo.staticMethod)
#        instanceMethod() {} // (namespace.Foo.prototype.instanceMethod)
#        get getter() {} // (namespace.Foo.prototype.get getter)
#    }
#    class ClassName {
#        constructor() {} // (ClassName)
#        method() {} // (ClassName.prototype.method)
#    }
#
# Methods may exist in object literals, outside of classes.
#
#   Foo.prototype = {
#       method() {}, // (Foo.prototype.method)
#       otherMethod() {} // (Foo.prototype.otherMethod)
#   }
#
# Comment handling is simple-minded but will work for all but pathological cases.
#
# Result is a list of triples: [ start_line, end_line, function_name ].

sub get_function_line_ranges_for_javascript($$)
{
    my ($fileHandle, $fileName) = @_;

    my @currentScopes;
    my @currentIdentifiers;
    my @currentParsingMode = ("global");
    my @currentFunctionNames;
    my @currentFunctionDepths;
    my @currentFunctionStartLines;

    my @ranges;

    my $inComment = 0;
    my $inQuotedText = "";
    my $inExtends = 0;
    my $inMethod = 0;
    my $inAnonymousFunctionParameters = 0;
    my $parenthesesDepth = 0;
    my $globalParenthesesDepth = 0;
    my $bracesDepth = 0;

    my $classJustSeen = 0;
    my $parenthesisJustSeen = 0;
    my $functionJustSeen = 0;
    my $getterJustSeen = 0;
    my $setterJustSeen = 0;
    my $assignmentJustSeen = 0;
    my $staticOrContructorSeen = 0;

    my $currentToken = "";
    my $lastToken = "";
    my $possibleMethodName = "";
    my $word = "";

    while (<$fileHandle>) {
        # Handle continued multi-line comment.
        if ($inComment) {
            next unless s-.*\*/--;
            $inComment = 0;
        }

        # Handle continued quoted text.
        if ($inQuotedText ne "") {
            next if /\\$/;
            s-([^\\]|\\.)*?$inQuotedText--;
            $inQuotedText = "";
        }

        # Handle comments and quoted text.
        while (m-(/\*|//|\'|\")-) { # \' and \" keep emacs perl mode happy
            my $match = $1;
            if ($match eq '/*') {
                if (!s-/\*.*?\*/--) {
                    s-/\*.*--;
                    $inComment = 1;
                }
            } elsif ($match eq '//') {
                s-//.*--;
            } else { # ' or "
                if (!s-$match([^\\]|\\.)*?$match-string_appeared_here-) {
                    $inQuotedText = $match if /\\$/;
                    warn "mismatched quotes at line $. in $fileName\n" if $inQuotedText eq "";
                    s-$match.*--;
                }
            }
        }

        # Find function names.
        while (m-(\w+|[(){}=:;,.])-g) {
            # Skip everything until "{" after extends.
            if ($inExtends) {
                next if $1 ne '{';
                $inExtends = 0;
            }

            $lastToken = $currentToken;
            $currentToken = $1;

            # Open parenthesis.
            if ($1 eq '(') {
                $parenthesesDepth++;
                $globalParenthesesDepth++ if $currentParsingMode[$#currentParsingMode] eq "global";
                $possibleMethodName = join('.', @currentIdentifiers);
                $inAnonymousFunctionParameters = 1 if $functionJustSeen;
                $functionJustSeen = 0;
                next;
            }

            # Close parenthesis.
            if ($1 eq ')') {
                $parenthesesDepth--;
                $globalParenthesesDepth-- if $currentParsingMode[$#currentParsingMode] eq "global";
                @currentIdentifiers = () if $inAnonymousFunctionParameters;
                $inAnonymousFunctionParameters = 0;
                $parenthesisJustSeen = 1;
                next;
            }

            # Open brace.
            if ($1 eq '{') {
                my $methodName = "";
                my $mode = $currentParsingMode[$#currentParsingMode];

                # Method.
                if (($mode eq 'class' or $mode eq 'global') and $parenthesisJustSeen and ($staticOrContructorSeen or $possibleMethodName)) {
                    if ($mode eq 'class') {
                        $methodName = join('.', $staticOrContructorSeen ? "" : "prototype", $possibleMethodName);
                    } else {
                        $methodName = $possibleMethodName;
                    }

                    $methodName =~ s/\.{2,}/\./g; # Removes consecutive periods.
                    $methodName =~ s/\.$//; # Remove trailing period.

                    my $currentMethod = join('.', @currentScopes, $methodName);
                    $currentMethod =~ s/\.{2,}/\./g; # Removes consecutive periods.
                    $currentMethod =~ s/\.$//; # Remove trailing period.

                    push(@currentParsingMode, "method");
                    push(@currentFunctionNames, $currentMethod);
                    push(@currentFunctionDepths, $bracesDepth);
                    push(@currentFunctionStartLines, $.);
                }

                $bracesDepth++;
                $functionJustSeen = 0;

                push(@currentScopes, join('.', $methodName ? $methodName : @currentIdentifiers));
                @currentIdentifiers = ();

                $staticOrContructorSeen = 0;
                next;
            }

            # Close brace.
            if ($1 eq '}') {
                $bracesDepth--;
                $functionJustSeen = 0;

                if (@currentFunctionDepths and $bracesDepth == $currentFunctionDepths[$#currentFunctionDepths]) {
                    pop(@currentFunctionDepths);
                    pop(@currentParsingMode);

                    my $currentName = pop(@currentFunctionNames);
                    my $start = pop(@currentFunctionStartLines);

                    $currentName =~ s/^\.//g; # Removes leading periods.

                    push(@ranges, [$start, $., $currentName]);
                }

                pop(@currentScopes);
                @currentIdentifiers = ();

                next;
            }

            # Dot.
            if ($1 eq '.') {
                next;
            }

            # Semicolon or comma.
            if ($1 eq ';' or $1 eq ',') {
                @currentIdentifiers = ();
                next;
            }

            # Class.
            if ($1 eq 'class') {
                $classJustSeen = 1;
                next;
            }

            # Extends.
            if ($1 eq 'extends') {
                $inExtends = 1;
                next;
            }

            # Function.
            if ($1 eq 'function') {
                $functionJustSeen = 1;

                if ($assignmentJustSeen) {
                    my $currentFunction = join('.', (@currentScopes, @currentIdentifiers));
                    $currentFunction =~ s/\.{2,}/\./g; # Removes consecutive periods.

                    push(@currentParsingMode, "function");
                    push(@currentFunctionNames, $currentFunction);
                    push(@currentFunctionDepths, $bracesDepth);
                    push(@currentFunctionStartLines, $.);
                }

                next;
            }

            # Getter prefix.
            if ($1 eq 'get') {
                next if $lastToken eq '.'; # Avoid map.get(...).
                $getterJustSeen = 1;
                next;
            }

            # Setter prefix.
            if ($1 eq 'set') {
                next if $lastToken eq '.'; # Avoid map.set(...).
                $setterJustSeen = 1;
                next;
            }

            # Static.
            if ($1 eq 'static' or $1 eq 'constructor') {
                $staticOrContructorSeen = 1;
                next;
            }

            # Assignment operator.
            if ($1 eq '=' or $1 eq ':') {
                $assignmentJustSeen = 1;
                next;
            }

            next if $parenthesesDepth > $globalParenthesesDepth;

            # Word.
            $word = $1;

            if ($classJustSeen) {
                push(@currentIdentifiers, $word) if !$assignmentJustSeen;

                my $currentClass = join('.', (@currentScopes, @currentIdentifiers));
                $currentClass =~ s/\.{2,}/\./g; # Removes consecutive periods.

                push(@currentParsingMode, "class");
                push(@currentFunctionNames, $currentClass);
                push(@currentFunctionDepths, $bracesDepth);
                push(@currentFunctionStartLines, $.);
            } elsif ($getterJustSeen or $setterJustSeen) {
                $word = "get $word" if $getterJustSeen;
                $word = "set $word" if $setterJustSeen;

                push(@currentIdentifiers, $word);

                my $mode = $currentParsingMode[$#currentParsingMode];
                my $currentFunction = join('.', (@currentScopes, ($mode eq 'class') ? "prototype" : "", @currentIdentifiers));
                $currentFunction =~ s/\.{2,}/\./g; # Removes consecutive periods.

                push(@currentParsingMode, "function");
                push(@currentFunctionNames, $currentFunction);
                push(@currentFunctionDepths, $bracesDepth);
                push(@currentFunctionStartLines, $.);
            } elsif ($functionJustSeen and !$assignmentJustSeen) {
                push(@currentIdentifiers, $word);

                my $currentFunction = join('.', (@currentScopes, @currentIdentifiers));
                $currentFunction =~ s/\.{2,}/\./g; # Removes consecutive periods.

                push(@currentParsingMode, "function");
                push(@currentFunctionNames, $currentFunction);
                push(@currentFunctionDepths, $bracesDepth);
                push(@currentFunctionStartLines, $.);
            } elsif ($word ne 'if' and $word ne 'for' and $word ne 'do' and $word ne 'while' and $word ne 'which' and $word ne 'var') {
                push(@currentIdentifiers, $word);
            }

            $classJustSeen = 0;
            $parenthesisJustSeen = 0;
            $functionJustSeen = 0;
            $getterJustSeen = 0;
            $setterJustSeen = 0;
            $assignmentJustSeen = 0;
        }
    }

    warn "mismatched braces in $fileName\n" if $bracesDepth;
    warn "mismatched parentheses in $fileName\n" if $parenthesesDepth;

    return @ranges;
}

# Read a file and get all the line ranges of the things that look like Perl functions. Functions
# start on a line that starts with "sub ", and end on the first line starting with "}" thereafter.
#
# Result is a list of triples: [ start_line, end_line, function ].

sub get_function_line_ranges_for_perl($$)
{
    my ($fileHandle, $fileName) = @_;

    my @ranges;

    my $currentFunction = "";
    my $start = 0;
    my $hereDocumentIdentifier = "";

    while (<$fileHandle>) {
        chomp;
        if (!$hereDocumentIdentifier) {
            if (/^sub\s+([\w_][\w\d_]*)/) {
                # Skip over forward declarations, which don't contain a brace and end with a semicolon.
                next if /;\s*$/;

                if ($currentFunction) {
                    warn "nested functions found at top-level at $fileName:$.\n";
                    next;
                }
                $currentFunction = $1;
                $start = $.;
            }
            if (/<<\s*[\"\']?([\w_][\w_\d]*)/) {
                # Enter here-document.
                $hereDocumentIdentifier = $1;
            }
            if (index($_, "}") == 0) {
                next unless $start;
                push(@ranges, [$start, $., $currentFunction]);
                $currentFunction = "";
                $start = 0;
            }
        } elsif ($_ eq $hereDocumentIdentifier) {
            # Escape from here-document.
            $hereDocumentIdentifier = "";
        }
    }

    return @ranges;
}

# Read a file and get all the line ranges of the things that look like Python classes, methods, or functions.
#
# FIXME: Maybe we should use Python's ast module to do the parsing for us?
#
# Result is a list of triples: [ start_line, end_line, function ].

sub get_function_line_ranges_for_python($$)
{
    my ($fileHandle, $fileName) = @_;

    my @ranges;

    my $multilineStringLiteralSentinelRegEx = qr#(?:"""|''')#;
    my $multilineStringLiteralStartRegEx = qr#^\s*$multilineStringLiteralSentinelRegEx#;
    my $multilineStringLiteralEndRegEx = qr#$multilineStringLiteralSentinelRegEx\s*$#;

    my @scopeStack = ({ line => 0, indent => -1, name => undef });
    my $lastLine = 0;
    my $inComment = 0;
    until ($lastLine) {
        $_ = <$fileHandle>;
        unless ($_) {
            # To pop out all popped scopes, run the loop once more after
            # we encountered the end of the file.
            $_ = "pass\n";
            $.++;
            $lastLine = 1;
        }
        chomp;
        next unless /^(\s*)([^#].*)$/; # Skip non-indented lines that begin with a comment.

        my $indent = length $1;
        my $rest = $2;
        my $scope = $scopeStack[-1];

        if ($indent <= $scope->{indent}) {
            # Find all the scopes that we have just exited.
            my $i = 0;
            for (; $i < @scopeStack; ++$i) {
                last if $indent <= $scopeStack[$i]->{indent};
            }
            my @poppedScopes = splice @scopeStack, $i;

            # For each scope that was just exited, add a range that goes from the start of that
            # scope to the start of the next nested scope, or to the line just before this one for
            # the innermost scope.
            for ($i = 0; $i < @poppedScopes; ++$i) {
                my $lineAfterEnd = $i + 1 == @poppedScopes ? $. : $poppedScopes[$i + 1]->{line};
                push @ranges, [$poppedScopes[$i]->{line}, $lineAfterEnd - 1, $poppedScopes[$i]->{name}];
            }
            @scopeStack or warn "Popped off last scope at $fileName:$.\n";

            # Set the now-current scope to start at the current line. Any lines within this scope
            # before this point should already have been added to @ranges.
            $scope = $scopeStack[-1];
            $scope->{line} = $.;
        }

        # Skip multi-line string literals and docstrings.
        next if /$multilineStringLiteralStartRegEx.*$multilineStringLiteralEndRegEx/;
        if (!$inComment && /$multilineStringLiteralStartRegEx/) {
            $inComment = 1;
        } elsif ($inComment && /$multilineStringLiteralEndRegEx/) {
            $inComment = 0;
        }
        next if $inComment;

        next if /^\s*[#'"]/; # Skip indented and non-indented lines that begin with a comment or string literal (includes docstrings).

        next unless $rest =~ /(?:class|def)\s+(\w+)/;
        my $name = $1;
        my $fullName = $scope->{name} ? join('.', $scope->{name}, $name) : $name;
        push @scopeStack, { line => $., indent => $indent, name => $fullName };

        if ($scope->{indent} >= 0) {
            push @ranges, [$scope->{line}, $. - 1, $scope->{name}];
        }
    }

    return @ranges;
}

# Read a file and get all the line ranges of the things that look like CSS selectors.  A selector is
# anything before an opening brace on a line. A selector starts at the line containing the opening
# brace and ends at the closing brace.
#
# Result is a list of triples: [ start_line, end_line, selector ].

sub get_selector_line_ranges_for_css($$)
{
    my ($fileHandle, $fileName) = @_;

    my @ranges;

    my $inComment = 0;
    my $inBrace = 0;
    my @stack;
    my $context;
    my @currentParseMode = ("global");
    my $selectorBraces = 0;

    while (<$fileHandle>) {
        foreach my $token (split m-(\{|\}|/\*|\*/)-, $_) {
            if ($token eq "{") {
                if (!$inComment) {
                    $inBrace += 1;                    
                    $selectorBraces += 1 if $currentParseMode[$#currentParseMode] eq "selector";
                    warn "mismatched opening brace found in $fileName:$.\n" if $selectorBraces > 1;
                }
            } elsif ($token eq "}") {
                if (!$inComment) {
                    if (!$inBrace or $currentParseMode[$#currentParseMode] eq "global") {
                        warn "mismatched closing brace found in $fileName:$.\n";
                        next;
                    }

                    $inBrace -= 1;

                    pop(@currentParseMode);
                    my $name = pop(@stack);
                    my $startLine = pop(@stack);
                    my $endLine = $.;
                    push(@ranges, [$startLine, $endLine, $name]);
                    $selectorBraces = 0;
                }
            } elsif ($token eq "/*") {
                $inComment = 1;
            } elsif ($token eq "*/") {
                warn "mismatched comment found in $fileName:$.\n" if !$inComment;
                $inComment = 0;
            } else {
                if (!$inComment and $currentParseMode[$#currentParseMode] ne "selector" and $token !~ /^[\s\t]*$/) {
                    $token =~ s/^[\s\t]*|[\s\t]*$//g;
                    my $startLine = $.;
                    if ($token =~ /^\@media/) {
                        push(@currentParseMode, "media");
                        push(@stack, ($startLine, $token));
                    } else {
                        push(@currentParseMode, "selector");
                        push(@stack, ($startLine, $token));
                    }
                }
            }
        }
    }

    # Sort by start line.
    return sort {$a->[0] <=> $b->[0]} @ranges;
}

# Read a file and get all the line ranges of the things that look like Swift classes, methods,
# or functions.
#
# Result is a list of triples: [ start_line, end_line, function ].

sub get_function_line_ranges_for_swift($$)
{
    my ($fileHandle, $fileName) = @_;

    my @ranges;

    my $currentFunction = "";
    my $currentClass = "";
    my $functionStart = 0;
    my $classStart = 0;
    my $functionScopeDepth = 0;
    my $classScopeDepth = 0;
    my $scopeDepth = 0;

    while (<$fileHandle>) {
        chomp;
        next if (/^\s*\/\/.*/);
        if (/func\s+([\w_][\w\d_]*)\((.*)\)/ || /var\s+([\w_][\w\d_]*):\s+/) {
            $functionScopeDepth = $scopeDepth;
            $currentFunction = $1;
            if ($2) {
                $currentFunction = "$currentFunction(". parseSwiftFunctionArgs($2) . ")";
            }
            if ($currentClass) {
                $currentFunction = "$currentClass.$currentFunction";
            }
            $functionStart = $.;
        } elsif (/class\s+([\w_][\w\d_]*)/) {
            $classScopeDepth = $scopeDepth;
            $currentClass = $1;
            $classStart = $.;
        }
        if (index($_, "{") > -1) {
            $scopeDepth++;
        }
        if (index($_, "}") > -1) {
            $scopeDepth--;
        }
        if ($scopeDepth == $functionScopeDepth) {
            next unless $functionStart;
            push(@ranges, [$functionStart, $., $currentFunction]);
            $currentFunction = "";
            $functionStart = 0;
        } elsif ($scopeDepth == $classScopeDepth) {
            next unless $classStart;
            $currentClass = "";
            $classStart = 0;
        }
    }

    return @ranges;
}

sub parseSwiftFunctionArgs($)
{
    my ($functionArgs) = @_;
    my @words = split /, /, $functionArgs;
    my $argCount = scalar(@words);
    if ($argCount == 0) {
        return "";
    } elsif ($argCount > 0) {
        # If the first argument is unnamed, give it the name "_"
        $words[0] =~ s/^(\w+: .*)/_ $1/;
        return join("", map { $_ =~ s/^(\w+).*/$1/; "$_:" } @words);
    } else {
        warn "Unknown argument count.\n";
    }
}

sub processPaths(\@)
{
    my ($paths) = @_;
    return ("." => 1) if (!@{$paths});

    my %result = ();

    for my $file (@{$paths}) {
        die "can't handle absolute paths like \"$file\"\n" if File::Spec->file_name_is_absolute($file);
        die "can't handle empty string path\n" if $file eq "";
        die "can't handle path with single quote in the name like \"$file\"\n" if $file =~ /'/; # ' (keep Xcode syntax highlighting happy)

        my $untouchedFile = $file;

        $file = canonicalizePath($file);

        die "can't handle paths with .. like \"$untouchedFile\"\n" if $file =~ m|/\.\./|;

        $result{$file} = 1;
    }

    return ("." => 1) if ($result{"."});

    # Remove any paths that also have a parent listed.
    for my $path (keys %result) {
        for (my $parent = dirname($path); $parent ne '.'; $parent = dirname($parent)) {
            if ($result{$parent}) {
                delete $result{$path};
                last;
            }
        }
    }

    return %result;
}

sub diffFromToString($$$)
{
    my ($gitCommit, $gitIndex, $mergeBase) = @_;

    return "" if isSVN();
    return $gitCommit if $gitCommit =~ m/.+\.\..+/;
    return "\"$gitCommit^\" \"$gitCommit\"" if $gitCommit;
    return "--cached" if $gitIndex;
    return $mergeBase if $mergeBase;
    return "HEAD" if isGit();
}

sub diffCommand($$$$)
{
    my ($paths, $gitCommit, $gitIndex, $mergeBase) = @_;

    # The function overlap detection logic in computeModifiedFunctions() assumes that its line
    # ranges were from a unified diff without any context lines.
    my $command;
    if (isSVN()) {
        my @escapedPaths = map(escapeSubversionPath($_), @$paths);
        my $escapedPathsString = qq(") . join(qq(" "), @escapedPaths) . qq(");
        $command = SVN . " diff --diff-cmd diff -x -U0 $escapedPathsString";
    } elsif (isGit()) {
        my $pathsString = "'" . join("' '", @$paths) . "'"; 
        $command = GIT . " diff --no-ext-diff -U0 " . diffFromToString($gitCommit, $gitIndex, $mergeBase);
        $command .= " -- $pathsString" unless $gitCommit or $mergeBase;
    }

    return $command;
}

sub statusCommand($$$$)
{
    my ($paths, $gitCommit, $gitIndex, $mergeBase) = @_;

    my $command;
    if (isSVN()) {
        my @escapedFiles = map(escapeSubversionPath($_), keys %$paths);
        my $escapedFilesString = qq(") . join(qq(" "), @escapedFiles) . qq(");
        $command = SVN . " stat $escapedFilesString";
    } elsif (isGit()) {
        my $filesString = '"' . join('" "', keys %$paths) . '"';
        $command = GIT . " diff -r --name-status -M -C " . diffFromToString($gitCommit, $gitIndex, $mergeBase);
        $command .= " -- $filesString" unless $gitCommit;
    }

    return "$command 2>&1";
}

sub attributeCommand($$)
{
    my ($file, $attr) = @_;

    my $result;
    if (isSVN()) {
        my $foundAttribute = 0;
        my $subPath = ".";
        my (@directoryParts) = File::Spec->splitdir($file);
        foreach my $part (@directoryParts) {
            if ($part eq ".") {
                next;
            }
            $subPath = File::Spec->join($subPath, $part);
            $subPath =~ s/^\.\///;
            if ($foundAttribute || exists $attributeCache{$attr}{$subPath} && $attributeCache{$attr}{$subPath} eq "1") {
                $attributeCache{$attr}{$subPath} = "1";
                $foundAttribute = 1;
                next;
            }
            my $command = SVN . " propget $attr '$subPath'";
            my $attrib = $attributeCache{$attr}{$subPath} || `$command 2> $devNull`;
            chomp $attrib;
            if ($attrib eq "1") {
                $foundAttribute = 1;
            }
            $attributeCache{$attr}{$subPath} = $attrib || "0";
        }
        $result = $attributeCache{$attr}{$file};
    } elsif (isGit()) {
        my $command = GIT . " check-attr $attr -- $file";
        $result = `$command`;
        chomp $result;
        $result =~ s/.*\W(\w)/$1/;
    }

    $result =~ s/\D//g;
    return int($result || 0);
}

sub createPatchCommand($$$$)
{
    my ($changedFilesString, $gitCommit, $gitIndex, $mergeBase) = @_;

    my $command;
    if (isSVN()) {
        $command = "'$FindBin::Bin/svn-create-patch --no-style' $changedFilesString";
    } elsif (isGit()) {
        $command = GIT . " diff -M -C " . diffFromToString($gitCommit, $gitIndex, $mergeBase);
        $command .= " -- $changedFilesString" unless $gitCommit;
    }

    return $command;
}

sub findOriginalFileFromSvn($)
{
    my ($file) = @_;
    my $baseUrl;
    open INFO, SVN . " info . |" or die;
    while (<INFO>) {
        if (/^URL: (.+?)[\r\n]*$/) {
            $baseUrl = $1;
        }
    }
    close INFO;
    my $sourceFile;
    my $escapedFile = escapeSubversionPath($file);
    open INFO, SVN . " info '$escapedFile' |" or die;
    while (<INFO>) {
        if (/^Copied From URL: (.+?)[\r\n]*$/) {
            $sourceFile = File::Spec->abs2rel($1, $baseUrl);
        }
    }
    close INFO;
    return $sourceFile;
}

sub determinePropertyChanges($$$)
{
    my ($file, $isAdd, $original) = @_;

    my $escapedFile = escapeSubversionPath($file);
    my %changes;
    if ($isAdd) {
        my %addedProperties;
        my %removedProperties;
        open PROPLIST, SVN . " proplist '$escapedFile' |" or die;
        while (<PROPLIST>) {
            $addedProperties{$1} = 1 if /^  (.+?)[\r\n]*$/ && $1 ne 'svn:mergeinfo';
        }
        close PROPLIST;
        if ($original) {
            my $escapedOriginal = escapeSubversionPath($original);
            open PROPLIST, SVN . " proplist '$escapedOriginal' |" or die;
            while (<PROPLIST>) {
                next unless /^  (.+?)[\r\n]*$/;
                my $property = $1;
                if (exists $addedProperties{$property}) {
                    delete $addedProperties{$1};
                } else {
                    $removedProperties{$1} = 1;
                }
            }
        }
        $changes{"A"} = [sort keys %addedProperties] if %addedProperties;
        $changes{"D"} = [sort keys %removedProperties] if %removedProperties;
    } else {
        open DIFF, SVN . " diff '$escapedFile' |" or die;
        while (<DIFF>) {
            if (/^Property changes on:/) {
                while (<DIFF>) {
                    my $operation;
                    my $property;
                    if (/^Added: (\S*)/) {
                        $operation = "A";
                        $property = $1;
                    } elsif (/^Modified: (\S*)/) {
                        $operation = "M";
                        $property = $1;
                    } elsif (/^Deleted: (\S*)/) {
                        $operation = "D";
                        $property = $1;
                    } elsif (/^Name: (\S*)/) {
                        # Older versions of svn just say "Name" instead of the type
                        # of property change.
                        $operation = "C";
                        $property = $1;
                    }
                    if ($operation) {
                        $changes{$operation} = [] unless exists $changes{$operation};
                        push @{$changes{$operation}}, $property;
                    }
                }
            }
        }
        close DIFF;
    }
    return \%changes;
}

sub pluralizeAndList($$@)
{
    my ($singular, $plural, @items) = @_;

    return if @items == 0;
    return "$singular $items[0]" if @items == 1;
    return "$plural " . join(", ", @items[0 .. $#items - 1]) . " and " . $items[-1];
}

sub generateFileList(\%$$$)
{
    my ($paths, $gitCommit, $gitIndex, $mergeBase) = @_;

    my @changedFiles;
    my @conflictFiles;
    my %functionLists;
    my @addedRegressionTests;
    my @requiresTests;
    print STDERR "  Running status to find changed, added, or removed files.\n";
    open STAT, "-|", statusCommand($paths, $gitCommit, $gitIndex, $mergeBase) or die "The status failed: $!.\n";
    while (<STAT>) {
        my $status;
        my $propertyStatus;
        my $propertyChanges;
        my $original;
        my $file;

        if (isSVN()) {
            my $matches;
            if (isSVNVersion16OrNewer()) {
                $matches = /^([ ACDMR])([ CM]).{5} (.+?)[\r\n]*$/;
                $status = $1;
                $propertyStatus = $2;
                $file = $3;
            } else {
                $matches = /^([ ACDMR])([ CM]).{4} (.+?)[\r\n]*$/;
                $status = $1;
                $propertyStatus = $2;
                $file = $3;
            }
            if ($matches) {
                $file = normalizePath($file);
                $original = findOriginalFileFromSvn($file) if substr($_, 3, 1) eq "+";
                my $isAdd = isAddedStatus($status);
                $propertyChanges = determinePropertyChanges($file, $isAdd, $original) if isModifiedStatus($propertyStatus) || $isAdd;
            } else {
                print;  # error output from svn stat
            }
        } elsif (isGit()) {
            if (/^([ADM])\t(.+)$/) {
                $status = $1;
                $propertyStatus = " ";  # git doesn't have properties
                $file = normalizePath($2);
            } elsif (/^([CR])[0-9]{1,3}\t([^\t]+)\t([^\t\n]+)$/) { # for example: R90%    newfile    oldfile
                $status = $1;
                $propertyStatus = " ";
                $original = normalizePath($2);
                $file = normalizePath($3);
            } else {
                print;  # error output from git diff
            }
        }

        next if !$status || isUnmodifiedStatus($status) && isUnmodifiedStatus($propertyStatus);

        $file = makeFilePathRelative($file);

        if (isModifiedStatus($status) || isAddedStatus($status) || isModifiedStatus($propertyStatus)) {
            my @components = File::Spec->splitdir($file);
            if ($components[0] eq "LayoutTests") {
                push @addedRegressionTests, $file
                    if isAddedStatus($status)
                       && $file =~ /\.([a-zA-Z]+)$/
                       && SupportedTestExtensions->{lc($1)}
                       && $file !~ /-expected(-mismatch)?\.html$/
                       && !scalar(grep(/^resources$/i, @components))
                       && !scalar(grep(/^script-tests$/i, @components));
            } elsif (attributeCommand($file, "test")) {
                push @addedRegressionTests, $file;
            } elsif (attributeCommand($file, "requiresTests")) {
                push @requiresTests, $file
            }
            push @changedFiles, $file if $components[$#components] ne "ChangeLog";
        } elsif (isConflictStatus($status, $gitCommit, $gitIndex) || isConflictStatus($propertyStatus, $gitCommit, $gitIndex)) {
            push @conflictFiles, $file;
        }
        if (basename($file) ne "ChangeLog") {
            my $description = statusDescription($status, $propertyStatus, $original, $propertyChanges);
            $functionLists{$file} = $description if defined $description;
        }
    }
    close STAT;
    return (\@changedFiles, \@conflictFiles, \%functionLists, \@addedRegressionTests, \@requiresTests);
}

sub isUnmodifiedStatus($)
{
    my ($status) = @_;

    my %statusCodes = (
        " " => 1,
    );

    return $statusCodes{$status};
}

sub isModifiedStatus($)
{
    my ($status) = @_;

    my %statusCodes = (
        "M" => 1,
    );

    return $statusCodes{$status};
}

sub isAddedStatus($)
{
    my ($status) = @_;

    my %statusCodes = (
        "A" => 1,
        "C" => isGit(),
        "R" => 1,
    );

    return $statusCodes{$status};
}

sub isConflictStatus($$$)
{
    my ($status, $gitCommit, $gitIndex) = @_;

    my %svn = (
        "C" => 1,
    );

    my %git = (
        "U" => 1,
    );

    return 0 if ($gitCommit || $gitIndex); # an existing commit or staged change cannot have conflicts
    return $svn{$status} if isSVN();
    return $git{$status} if isGit();
}

sub statusDescription($$$$)
{
    my ($status, $propertyStatus, $original, $propertyChanges) = @_;

    my $propertyDescription = defined $propertyChanges ? propertyChangeDescription($propertyChanges) : "";

    my %svn = (
        "A" => defined $original ? sprintf(" Copied from \%s.", $original) : " Added.",
        "D" => " Removed.",
        "M" => "",
        "R" => defined $original ? sprintf(" Replaced with \%s.", $original) : " Replaced.",
        " " => "",
    );

    my %git = %svn;
    $git{"A"} = " Added.";
    if (defined $original) {
        $git{"C"} = sprintf(" Copied from \%s.", $original);
        $git{"R"} = sprintf(" Renamed from \%s.", $original);
    }

    my $description;
    $description = $svn{$status} if isSVN() && exists $svn{$status};
    $description = $git{$status} if isGit() && exists $git{$status};
    return unless defined $description;

    $description .= $propertyDescription unless isAddedStatus($status);
    return $description;
}

sub propertyChangeDescription($)
{
    my ($propertyChanges) = @_;

    my %operations = (
        "A" => "Added",
        "M" => "Modified",
        "D" => "Removed",
        "C" => "Changed",
    );

    my $description = "";
    while (my ($operation, $properties) = each %$propertyChanges) {
        my $word = $operations{$operation};
        my $list = pluralizeAndList("property", "properties", @$properties);
        $description .= " $word $list.";
    }
    return $description;
}

sub extractLineRangeAfterChange($)
{
    my ($string) = @_;
    my $chunkRange = parseChunkRange($string);
    if (!$chunkRange) {
        return (-1, -1); # Malformed
    }
    if (!$chunkRange->{newStartingLine} || !$chunkRange->{newLineCount}) {
         # Deletion; no lines exist after change.
        return ($chunkRange->{newStartingLine}, $chunkRange->{newStartingLine});
    }
    return ($chunkRange->{newStartingLine}, $chunkRange->{newStartingLine} + $chunkRange->{newLineCount} - 1);
}

sub extractLineRangeBeforeChange($)
{
    my ($string) = @_;
    my $chunkRange = parseChunkRange($string);
    if (!$chunkRange) {
        return (-1, -1); # Malformed
    }
    if (!$chunkRange->{startingLine} || !$chunkRange->{lineCount}) {
        # Addition; no lines existed before change.
        return ($chunkRange->{startingLine}, $chunkRange->{startingLine});
    }
    return ($chunkRange->{startingLine}, $chunkRange->{startingLine} + $chunkRange->{lineCount} - 1);
}

sub testListForChangeLog(@)
{
    my (@tests) = @_;

    return "" unless @tests;

    my $leadString = "        Test" . (@tests == 1 ? "" : "s") . ": ";
    my $list = $leadString;
    foreach my $i (0..$#tests) {
        $list .= " " x length($leadString) if $i;
        my $test = $tests[$i];
        $test =~ s/^LayoutTests\///;
        $list .= "$test\n";
    }
    $list .= "\n";

    return $list;
}

sub reviewerAndDescriptionForGitCommit($$)
{
    my ($commit, $gitReviewer) = @_;

    my $description = '';
    my $reviewer;

    my @args = qw(rev-list --pretty);
    push @args, '-1' if $commit !~ m/.+\.\..+/;
    my $gitLog;
    {
        local $/ = undef;
        open(GITLOG, "-|", GIT, @args, $commit) || die;
        $gitLog = <GITLOG>;
        close(GITLOG);
    }

    my @commitLogs = split(/^[Cc]ommit [a-f0-9]{40}/m, $gitLog);
    shift @commitLogs; # Remove initial blank commit log
    my $commitLogCount = 0;
    foreach my $commitLog (@commitLogs) {
        $description .= "\n" if $commitLogCount;
        $commitLogCount++;
        my $inHeader = 1;
        my $commitLogIndent; 
        my @lines = split(/\n/, $commitLog);
        shift @lines; # Remove initial blank line
        foreach my $line (@lines) {
            if ($inHeader) {
                if (!$line) {
                    $inHeader = 0;
                }
                next;
            } elsif ($line =~ /[Ss]igned-[Oo]ff-[Bb]y: (.+)/) {
                if (!$reviewer) {
                    $reviewer = $1;
                } else {
                    $reviewer .= ", " . $1;
                }
            } elsif ($line =~ /^\s*$/) {
                $description = $description . "\n";
            } else {
                if (!defined($commitLogIndent)) {
                    # Let the first line with non-white space determine
                    # the global indent.
                    $line =~ /^(\s*)\S/;
                    $commitLogIndent = length($1);
                }
                # Strip at most the indent to preserve relative indents.
                $line =~ s/^\s{0,$commitLogIndent}//;
                $description = $description . (" " x 8) . $line . "\n";
            }
        }
    }
    if (!$reviewer) {
      $reviewer = $gitReviewer;
    }

    return ($reviewer, $description);
}

sub normalizeLineEndings($$)
{
    my ($string, $endl) = @_;
    $string =~ s/\r?\n/$endl/g;
    return $string;
}

sub decodeEntities($)
{
    my ($text) = @_;
    $text =~ s/\&lt;/</g;
    $text =~ s/\&gt;/>/g;
    $text =~ s/\&quot;/\"/g;
    $text =~ s/\&apos;/\'/g;
    $text =~ s/\&amp;/\&/g;
    return $text;
}