我有差异，我后处理并想要弄平相等的线条。这是一个例子：
Foo
-Bar
+Bar
Baz
我想把相等的线压扁，这样它们就不会再出现在差异中了。这很简单
-(.*)\n\+\1\n
当我有多行匹配时，问题就开始了，例如：
-Foo
-Bar
+Foo
+Bar
有任何想法吗？还是我不应该做一个正则表达式并编写一个简单的解析器？还是已经存在？
如果有更好的解决方案，一些背景故事。我正在比较两个文件以查看它们是否相同。遗憾的是，输出几乎相同，但需要一些后处理，例如
-on line %d
+on line 8
所以我正在经历并将已知字符串转换为其他已知字符串，然后我试图检查差异是否为空或仍然不同。
我之前对diff输出进行了一些更简单的分析，所以我有一个 Perl 脚本，它为我提供了一个开始的基础。考虑以下两个数据文件，file.1和file.2.
Data
Foo
Bar 1
Baz
I want to squash the lines down that are equal so they don't show up in the diff anymore. This is pretty simple with
-(.*)\n\+\1\n
The problems start when I have multi-line matches like:
Foo 2
Bar 2
Etc.
Data
Foo
Bar 10
Baz
I want to squash the lines down that are equal so they don't show up in the diff anymore. This is pretty simple with
-(.*)\n\+\1\n
The problems start when I have multi-line matches like:
Foo 20
Bar 20
Etc.
原始统一diff输出为：
--- file.1 2013-03-30 18:58:35.000000000 -0700
+++ file.2 2013-03-30 18:58:48.000000000 -0700
@@ -1,7 +1,7 @@
Data
Foo
-Bar 1
+Bar 10
Baz
I want to squash the lines down that are equal so they don't show up in the diff anymore. This is pretty simple with
@@ -10,7 +10,7 @@
The problems start when I have multi-line matches like:
-Foo 2
-Bar 2
+Foo 20
+Bar 20
Etc.
现在，经过后期处理，所有的数字字符串都被替换为##，所以后期处理的文件如下所示：
--- file.1 2013-03-30 18:58:35.000000000 -0700
+++ file.2 2013-03-30 18:58:48.000000000 -0700
@@ -1,7 +1,7 @@
Data
Foo
-Bar ##
+Bar ##
Baz
I want to squash the lines down that are equal so they don't show up in the diff anymore. This is pretty simple with
@@ -10,7 +10,7 @@
The problems start when I have multi-line matches like:
-Foo ##
-Bar ##
+Foo ##
+Bar ##
Etc.
这是程序的输入，它将分析差异是否仍然真实。
为了真正有用，我们必须隔离标题行（---和+++）并保留它们。对于开始的每个差异块@@，我们需要捕获 和 的相邻行，-并且+：
+相同--行的内容相同+。-和行。+@@@@，则可以丢弃整个块。冲洗并重复。
我为此选择的编程语言是 Perl。
#!/usr/bin/env perl
use strict;
use warnings;
use constant debug => 0;
my $file1;
my $file2;
my $header = 0;
OUTER:
while (my $line = <>)
{
chomp $line;
print "[$line]\n" if debug;
if ($line =~ m/^--- /)
{
$file1 = $line;
$file2 = <>;
chomp $file2;
print "[$file2]\n" if debug;
if ($file2 !~ m/^\+\+\+ /)
{
print STDERR "Unexpected file identification lines\n";
print STDERR "$file1\n";
print STDERR "$file2\n";
next OUTER;
}
$header = 0; # Have not output file header yet
my @lines;
my $atline;
last OUTER unless defined($line = <>);
INNER:
while ($line =~ m/^@@ /)
{
chomp $line;
print "@[$line]\n" if debug;
$atline = $line;
@lines = ();
while (defined($line = <>) && $line =~ m/^[- +]/)
{
chomp $line;
print ":[$line]\n" if debug;
push @lines, $line;
}
# Got a complete @@ block of diffs
post_process($atline, @lines);
last OUTER if !defined($line);
next INNER if ($line =~ m/^@@ /);
print STDERR "Unexpected input line: [$line]\n";
last OUTER;
}
}
}
sub differences
{
my($pref, $mref) = @_;
my $pnum = scalar(@$pref);
my $mnum = scalar(@$mref);
print "-->> differences\n" if debug;
return 0 if ($pnum == 0 && $mnum == 0);
return 1 if ($pnum != $mnum);
foreach my $i (0..($pnum-1))
{
my $pline = substr(${$pref}[$i], 1);
my $mline = substr(${$mref}[$i], 1);
return 1 if ($pline ne $mline);
}
print "<<-- differences\n" if debug;
return 0;
}
sub post_process
{
my($atline, @lines) = @_;
print "-->> post_process\n" if debug;
# Work out whether there are any differences left
my @plines = (); # +lines
my @mlines = (); # -lines
my $diffs = 0;
my $ptype = ' '; # Previous line type
foreach my $line (@lines)
{
print "---- $line\n" if debug;
my ($ctype) = ($line =~ m/^(.)/);
if ($ctype eq ' ')
{
if (($ptype eq '-' || $ptype eq '+') && differences(\@plines, \@mlines))
{
$diffs = 1;
last;
}
@plines = ();
@mlines = ();
}
elsif ($ctype eq '-')
{
push @mlines, $line;
}
elsif ($ctype eq '+')
{
push @plines, $line;
}
else
{
print STDERR "Unexpected input line format: $line\n";
exit 1;
}
$ptype = $ctype;
}
$diffs = 1 if differences(\@plines, \@mlines);
if ($diffs != 0)
{
# Print the block of differences, preceded by file header if necessary
if ($header == 0)
{
print "$file1\n";
print "$file2\n";
$header = 1;
}
print "$atline\n";
foreach my $line (@lines)
{
print "$line\n";
}
}
print "<<-- post_process\n" if debug;
return;
}
使用文件进行了测试data，并使用了三个变体：
$ perl checkdiffs.pl data
$ perl checkdiffs.pl data.0
--- file.1 2013-03-30 18:58:35.000000000 -0700
+++ file.2 2013-03-30 18:58:48.000000000 -0700
@@ -1,7 +1,7 @@
Data
Foo
-Bar #0
+Bar ##
Baz
I want to squash the lines down that are equal so they don't show up in the diff anymore. This is pretty simple with
$ perl checkdiffs.pl data.1
--- file.1 2013-03-30 18:58:35.000000000 -0700
+++ file.2 2013-03-30 18:58:48.000000000 -0700
@@ -10,7 +10,7 @@
The problems start when I have multi-line matches like:
-Foo #0
-Bar ##
+Foo ##
+Bar ##
Etc.
$ perl checkdiffs.pl data.2
--- file.1 2013-03-30 18:58:35.000000000 -0700
+++ file.2 2013-03-30 18:58:48.000000000 -0700
@@ -1,7 +1,7 @@
Data
Foo
-Bar #0
+Bar ##
Baz
I want to squash the lines down that are equal so they don't show up in the diff anymore. This is pretty simple with
@@ -10,7 +10,7 @@
The problems start when I have multi-line matches like:
-Foo ##
-Bar #0
+Foo ##
+Bar ##
Etc.
$
这符合你的要求吗？
我认为这可能有效（除非您重复配对）：
sed 's/^[-+]//' filename | perl -ne 'print unless $seen{$_}++'
将开始 +/- 替换为空字符串。然后只选择独特的线条。
You could use s modifier and positive lookahead:
Here is sample matching at regexpal.
Here is C# regex sample that should be close to what you need:
var sourceString = @"-Foo
+Foo
la
-Bar
+Foo
la
-Ko
+Bar
la
+Ko
-Ena
asdsda
-Dva
+Ena
+Dva
";
Regex ItemRegex = new Regex(@"(?s)\-(.*?)\n(?=(.*?)(\+\1))", RegexOptions.Compiled);
foreach (Match ItemMatch in ItemRegex.Matches(sourceString))
{
Console.WriteLine(ItemMatch);
}