From b958c6bedb7fcb8b71221e6895d29468a985dc43 Mon Sep 17 00:00:00 2001 From: Dominique Hazael-Massieux Date: Tue, 25 Nov 2014 11:53:13 +0100 Subject: [PATCH 1/2] allow to load ignore list from file --- bin/checklink | 36 ++++++++++++++++++++++++++++++++++-- 1 file changed, 34 insertions(+), 2 deletions(-) diff --git a/bin/checklink b/bin/checklink index f1d692f..de3a537 100755 --- a/bin/checklink +++ b/bin/checklink @@ -490,6 +490,8 @@ my %Opts = ( Suppress_All_Redirects => 0, Suppress_Broken => [], Suppress_Fragment => [], + Ignore_From_File => undef, + Ignore_List => [], Masquerade => 0, Masquerade_From => '', Masquerade_To => '', @@ -601,6 +603,16 @@ if ($Opts{Command_Line}) { # Transform the parameter into a URI $uri = &urize($uri); + + # relative fragments in suppress_fragments get prefixed with first url + for my $i (0 .. $#{$Opts{Suppress_Fragment}}) { + my $url = ${$Opts{Suppress_Fragment}}[$i]; + if ($url =~ /^#./) { + $url = URI->new_abs($url, $uri)->canonical(); + ${$Opts{Suppress_Fragment}}[$i] = $url; + } + } + $params{uri} = $uri; &check_uri(\%params, $uri, $check_num, $Opts{Depth}, undef, undef, 1); $check_num++; @@ -795,6 +807,7 @@ sub parse_arguments () 'suppress-all-redirects' => \$Opts{Suppress_All_Redirects}, 'suppress-broken=s@' => \@{$Opts{Suppress_Broken}}, 'suppress-fragment=s@' => \@{$Opts{Suppress_Fragment}}, + 'ignore-from-file=s' => \$Opts{Ignore_From_File}, 'u|user=s' => \$Opts{User}, 'p|password=s' => \$Opts{Password}, 't|timeout=i' => \$Opts{Timeout}, @@ -850,6 +863,22 @@ sub parse_arguments () $Opts{Depth} = -1 if ($Opts{Depth} == 0 && @locs); + # Open ignore files + if (defined($Opts{Ignore_From_File})) { + open(my $ignore_file, "<", $Opts{Ignore_From_File}) + or &usage(1, "Can't open ignore file: $Opts{Ignore_From_File}" ); + while (my $url = <$ignore_file>) { + chomp $url; + $url = Encode::decode_utf8($url); + if ($url !~ /#./) { + push(@{$Opts{Ignore_List}}, $url); + } else { + push(@{$Opts{Suppress_Fragment}}, $url); + } + } + } + + # Precompile/error-check regular expressions. if (defined($Opts{Exclude})) { eval { $Opts{Exclude} = qr/$Opts{Exclude}/o; }; @@ -896,7 +925,7 @@ sub parse_arguments () } } for my $sf_arg (@{$Opts{Suppress_Fragment}}) { - if ($sf_arg !~ /.#./) { + if ($sf_arg !~ /#./) { &usage(1, "Bad suppress-fragment argument, should contain \"#\": $sf_arg" ); @@ -960,6 +989,8 @@ Options: --suppress-fragment URI Do not report the given broken fragment URI. A fragment URI contains \"#\". This option may be specified multiple times. + --ignore-from-file FILE Do not report error for URIs listed in FILE (one + per line) -L, --languages LANGS Accept-Language header to send. The special value 'auto' causes autodetection from the environment. -c, --cookies FILE Use cookies, load/save them in FILE. The special @@ -1310,7 +1341,8 @@ EOF my $canon_uri = URI->new($abs_link_uri->canonical()); my $fragment = $canon_uri->fragment(undef); - if (!defined($Opts{Exclude}) || $canon_uri !~ $Opts{Exclude}) { + if ((!defined($Opts{Exclude}) || $canon_uri !~ $Opts{Exclude}) + && (!grep { $_ eq $canon_uri } @{$Opts{Ignore_List}})) { if (!exists($links{$canon_uri})) { my $hostport; $hostport = $canon_uri->host_port() From 7a11d72a2223a1cc736503db267eee2e695ded02 Mon Sep 17 00:00:00 2001 From: Dominique Hazael-Massieux Date: Wed, 26 Nov 2014 08:49:40 +0100 Subject: [PATCH 2/2] document that relative fragments are interpreted with the first URI as a base --- bin/checklink | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/bin/checklink b/bin/checklink index de3a537..205b5f4 100755 --- a/bin/checklink +++ b/bin/checklink @@ -988,9 +988,12 @@ Options: This option may be specified multiple times. --suppress-fragment URI Do not report the given broken fragment URI. A fragment URI contains \"#\". This option may be - specified multiple times. + specified multiple times. Note that you may specify + a fragment with no scheme or path (e.g., + #myfragment). In this case the fragment will be + considered internal to the first URI being checked. --ignore-from-file FILE Do not report error for URIs listed in FILE (one - per line) + per line). -L, --languages LANGS Accept-Language header to send. The special value 'auto' causes autodetection from the environment. -c, --cookies FILE Use cookies, load/save them in FILE. The special