os/doc
3
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Exclude all files in htdocs/releases, and better check and sanitize the

Browse source 
exclude path used.
This commit is contained in:
Warren Block 2017-09-09 13:35:17 +00:00
parent 2e832d5849
commit 90c78f7750
Notes: svn2git 2020-12-08 03:00:23 +00:00 
svn path=/head/; revision=50814
1 changed files with 15 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

27
share/tools/convert2utf8/convert2utf8
View file

@ -48,7 +48,7 @@ my $svn = '/usr/local/bin/svn';
my $xargs = '/usr/bin/xargs'; my $xargs = '/usr/bin/xargs';
my $docdir = '/usr/doc/en_US.ISO8859-1'; my $docdir = '/usr/doc/en_US.ISO8859-1';
my $exclpath = 'htdocs/releases/*/*.html'; my $exclpath = 'htdocs/releases/*/*';
my $verbose = 0; my $verbose = 0;
sub usage { sub usage {
@ -64,8 +64,9 @@ sub usage {
print "encoding directories like en_US.ISO8859-1 to UTF-8. The\n"; print "encoding directories like en_US.ISO8859-1 to UTF-8. The\n";
print "documentation directory must be a Subversion checkout. After\n"; print "documentation directory must be a Subversion checkout. After\n";
print "files are converted, the directory is renamed to *.UTF-8.\n\n"; print "files are converted, the directory is renamed to *.UTF-8.\n\n";
print "The default exclude path prevents conversion of statically-\n"; print "The default exclude path prevents conversion of all files\n";
print "generated HTML files in htdocs/releases/*/*.html.\n"; print "htdocs/releases/*/*. These files are typically statically-\n";
print "generated historical files.\n";
exit 0; exit 0;
} }
@ -94,10 +95,17 @@ sub make_clean {
} }
sub find_files { sub find_files {
my ($dn) = @_; my ($dn,$exclpath) = @_;
my $fullexclpath = "$dn/$exclpath";
# sanitize exclude path, find is very picky about matches
# convert multiple slashes to single
$fullexclpath =~ s%/{2,}%/%;
die "** exclude path '$exclpath' must be relative (under '$docdir')\n" if $exclpath =~ m%^/%;
print "finding files to be converted in '$dn'\n" if $verbose; print "finding files to be converted in '$dn'\n" if $verbose;
print "excluding files matching \"$dn$exclpath/*\"\n" if $verbose; print "excluding files matching \"$fullexclpath/*\"\n" if $verbose;
return map(/^(.*):/, `$find $dn -not -path \"$dn$exclpath\" -type f -print0 | $xargs -0 $file | $grep 'XML\\|SGML\\|BSD'`); return map(/^(.*):/, `$find $dn -not -path \"$fullexclpath\" -type f -print0 | $xargs -0 $file | $grep 'XML\\|SGML\\|BSD'`);
} }
sub convert_file { sub convert_file {
@ -149,11 +157,6 @@ sub main {
$docdir = $opt_d if $opt_d; $docdir = $opt_d if $opt_d;
$exclpath = $opt_e if $opt_e; $exclpath = $opt_e if $opt_e;
# sanitize exclude path, find is very picky about matches
# convert multiple slashes to single
$exclpath =~ s%/{2,}%/%;
die "** exclude path '$exclpath' must be relative (under '$docdir')\n" if $exclpath =~ m%^/%;
my $basedir = basename($docdir); my $basedir = basename($docdir);
die "** '$docdir' does not have a standard ISO xy_AB directory name\n" unless $basedir =~ /^([a-z]{2}_[A-Z]{1,3})\./; die "** '$docdir' does not have a standard ISO xy_AB directory name\n" unless $basedir =~ /^([a-z]{2}_[A-Z]{1,3})\./;
my $isolang = $1; my $isolang = $1;
@ -179,7 +182,7 @@ sub main {
make_clean($docdir); make_clean($docdir);
my @files = find_files($docdir); my @files = find_files($docdir, $exclpath);
for my $f (@files) { for my $f (@files) {
convert_file($f, $basedir, $isolang, $fromcode, $tocode); convert_file($f, $basedir, $isolang, $fromcode, $tocode);