diff options
-rwxr-xr-x | scripts/cleanfile | 126 |
1 files changed, 126 insertions, 0 deletions
diff --git a/scripts/cleanfile b/scripts/cleanfile new file mode 100755 index 000000000000..f1ba8aa58a40 --- /dev/null +++ b/scripts/cleanfile | |||
@@ -0,0 +1,126 @@ | |||
1 | #!/usr/bin/perl -w | ||
2 | # | ||
3 | # Clean a text file -- or directory of text files -- of stealth whitespace. | ||
4 | # WARNING: this can be a highly destructive operation. Use with caution. | ||
5 | # | ||
6 | |||
7 | use bytes; | ||
8 | use File::Basename; | ||
9 | |||
10 | # | ||
11 | # Clean up space-tab sequences, either by removing spaces or | ||
12 | # replacing them with tabs. | ||
13 | sub clean_space_tabs($) | ||
14 | { | ||
15 | no bytes; # Tab alignment depends on characters | ||
16 | |||
17 | my($li) = @_; | ||
18 | my($lo) = ''; | ||
19 | my $pos = 0; | ||
20 | my $nsp = 0; | ||
21 | my($i, $c); | ||
22 | |||
23 | for ($i = 0; $i < length($li); $i++) { | ||
24 | $c = substr($li, $i, 1); | ||
25 | if ($c eq "\t") { | ||
26 | my $npos = ($pos+$nsp+8) & ~7; | ||
27 | my $ntab = ($npos >> 3) - ($pos >> 3); | ||
28 | $lo .= "\t" x $ntab; | ||
29 | $pos = $npos; | ||
30 | $nsp = 0; | ||
31 | } elsif ($c eq "\n" || $c eq "\r") { | ||
32 | $lo .= " " x $nsp; | ||
33 | $pos += $nsp; | ||
34 | $nsp = 0; | ||
35 | $lo .= $c; | ||
36 | $pos = 0; | ||
37 | } elsif ($c eq " ") { | ||
38 | $nsp++; | ||
39 | } else { | ||
40 | $lo .= " " x $nsp; | ||
41 | $pos += $nsp; | ||
42 | $nsp = 0; | ||
43 | $lo .= $c; | ||
44 | $pos++; | ||
45 | } | ||
46 | } | ||
47 | $lo .= " " x $nsp; | ||
48 | return $lo; | ||
49 | } | ||
50 | |||
51 | $name = basename($0); | ||
52 | |||
53 | foreach $f ( @ARGV ) { | ||
54 | print STDERR "$name: $f\n"; | ||
55 | |||
56 | if (! -f $f) { | ||
57 | print STDERR "$f: not a file\n"; | ||
58 | next; | ||
59 | } | ||
60 | |||
61 | if (!open(FILE, '+<', $f)) { | ||
62 | print STDERR "$name: Cannot open file: $f: $!\n"; | ||
63 | next; | ||
64 | } | ||
65 | |||
66 | binmode FILE; | ||
67 | |||
68 | # First, verify that it is not a binary file; consider any file | ||
69 | # with a zero byte to be a binary file. Is there any better, or | ||
70 | # additional, heuristic that should be applied? | ||
71 | $is_binary = 0; | ||
72 | |||
73 | while (read(FILE, $data, 65536) > 0) { | ||
74 | if ($data =~ /\0/) { | ||
75 | $is_binary = 1; | ||
76 | last; | ||
77 | } | ||
78 | } | ||
79 | |||
80 | if ($is_binary) { | ||
81 | print STDERR "$name: $f: binary file\n"; | ||
82 | next; | ||
83 | } | ||
84 | |||
85 | seek(FILE, 0, 0); | ||
86 | |||
87 | $in_bytes = 0; | ||
88 | $out_bytes = 0; | ||
89 | $blank_bytes = 0; | ||
90 | |||
91 | @blanks = (); | ||
92 | @lines = (); | ||
93 | |||
94 | while ( defined($line = <FILE>) ) { | ||
95 | $in_bytes += length($line); | ||
96 | $line =~ s/[ \t\r]*$//; # Remove trailing spaces | ||
97 | $line = clean_space_tabs($line); | ||
98 | |||
99 | if ( $line eq "\n" ) { | ||
100 | push(@blanks, $line); | ||
101 | $blank_bytes += length($line); | ||
102 | } else { | ||
103 | push(@lines, @blanks); | ||
104 | $out_bytes += $blank_bytes; | ||
105 | push(@lines, $line); | ||
106 | $out_bytes += length($line); | ||
107 | @blanks = (); | ||
108 | $blank_bytes = 0; | ||
109 | } | ||
110 | } | ||
111 | |||
112 | # Any blanks at the end of the file are discarded | ||
113 | |||
114 | if ($in_bytes != $out_bytes) { | ||
115 | # Only write to the file if changed | ||
116 | seek(FILE, 0, 0); | ||
117 | print FILE @lines; | ||
118 | |||
119 | if ( !defined($where = tell(FILE)) || | ||
120 | !truncate(FILE, $where) ) { | ||
121 | die "$name: Failed to truncate modified file: $f: $!\n"; | ||
122 | } | ||
123 | } | ||
124 | |||
125 | close(FILE); | ||
126 | } | ||