use strict;
use blib;
use Fuse::DBI;
+use lib '/data/WebGUI/lib/';
+use Data::Config;
-my $sql_filenames = q{
- select
- oid as id,
- namespace||'/'||name||' ['||oid||']' as filename,
- length(template) as size,
- iseditable as writable
- from template ;
-};
+=head1 NAME
-my $sql_read = q{
- select template
- from template
- where oid = ?;
-};
+webgui.pl - mount WebGUI templates as filesystem
+
+=head1 SYNOPSIS
+
+ webgui.pl /data/WebGUI/etc/webgui.conf /mnt
+
+=head1 DESCRIPTION
+
+With this script, you can utilize C<Fuse> and C<Fuse::DBI> modules to mount
+templates from WebGUI and edit them using command-line utilities (like C<vi>
+or C<ftp>).
+
+It will present templates in WebGUI as directory tree consisting of
+directories (which represent template's namespace) and files (which are
+templates from database). If template name has slash (C</>) in name, deeper
+directories will be created.
+
+Template files will have correct lengths and write permissions which are
+specified in WebGUI database.
+
+=head2 Fuse module
+
+C<Fuse::DBI> module (which is core of this utility) uses C<Fuse> perl
+bindings. Perl bindings are rather new addition to C<Fuse>, so
+you will need recent CVS version of C<Fuse>. Current stable version doesn't
+include perl binding, so you will probably have to compile C<Fuse> yourself
+(see FUSE documentation for details about compilation and installation).
+
+After compilation and installation of C<fuse> kernel module and C<Fuse> perl
+bindings for it, you will have to load C<fuse> module into kernel. For that,
+you will have to be root. If you are not administrator on particular
+machine, ask your admin to install and load C<fuse> module for you.
+
+If you used C<fusermount> command before running this script, module will be
+already loaded.
+
+=head2 unsupported operations
+
+There is no support for creation of new templates, renaming, or deleting.
+Although those operations map nicely to file system semantics there are still
+possible only using WebGUI web interface.
+
+=head2 unlink to invalidate cache
+
+Unlink command (C<rm>) is implemented on files with special function: it
+will remove in-memory cache of particular template and reload it from
+database. That enables usage of web interface to make small changes and then
+continuing editing using this script without restarting it.
+
+In-memory cache is populated with data about available templates when you
+start this script. Currently only way to refresh template list (after you
+create copy of template through web interface) is to remove directory using
+C<rmdir> or C<rm -rf>.
+
+B<Don't panic!> Destructive operations in filesystem (C<rm> and C<rmdir>)
+just invalidate in-memory cache and re-read data from database (this will
+also change ctime of file, so your editor will probably notice that file has
+changed).
+
+In-memory cache is used to speed up operations like grep on templates. If it
+wasn't there, grep wouldn't be useful at all. I think this is acceptable
+compromise.
+
+=head2 invalidating of on-disk templates
+
+Every write operation will erase all templates on disk (so that next reload
+on browser will show your changes). It would be better if just changed
+template is erased, but this works well enough. You might notice performance
+penalty of this simplification if you are running very loaded production
+site.
+
+You have to have write permission on C<uploads/temp/templates/> directory
+for your WebGUI instance for this to work. If you don't C<Fuse::DBI> will
+complain.
+
+=head2 supported databases
+
+This script have embedded SQL queries for MySQL and PostgreSQL. Other databases
+could be supported easily. Contributions are welcomed.
+
+=head2 database transactions
+
+C<Fuse::DBI> uses transactions (if your database supports them) to prevent
+accidental corruption of data by reading old version. Depending on type of
+database back-end, MySQL users might be out of luck.
+
+=head2 recovering from errors
-my $sql_update = q{
- update template
- set template = ?
- where oid = ?;
+B<Transport endpoint is not connected> is very often error when Fuse perl
+bindings exit without clean umount (through C<Fuse::DBI> C<umount> method or
+with C<fusermount -u /mnt> command).
+
+This script will automatically run C<fusermount -u /mnt> if it receives
+above error on startup. If it fails, mount point is still in use (that
+happens if you changed directory to mount point in other shell). Solution is
+simple, just change directory in other back to C<$HOME> (with just C<cd>)
+and re-run this script.
+
+=head2 missing Data::Config
+
+If this script complains about missing C<Data::Config> module, you will have
+to point path at top which points to lib directory of WebGUI installation.
+By default it points to C</data/WebGUI>:
+
+ use lib '/data/WebGUI/lib/';
+
+=cut
+
+my ($config_file,$mount) = @ARGV;
+
+unless ($config_file && $mount) {
+ print STDERR <<_USAGE_;
+usage: $0 /data/WebGUI/etc/webgui.conf /mnt
+
+For more information see perldoc webgui.pl
+_USAGE_
+ exit 1;
+}
+
+system "fusermount -u $mount" unless (-w $mount);
+
+unless (-w $mount) {
+ print STDERR "Current user doesn't have permission on mount point $mount: $!";
+ exit 1;
+}
+
+my $config = new Data::Config $config_file || "can't open config $config_file: $!";
+
+my $sql = {
+ 'pg' => {
+ 'filenames' => q{
+ select
+ oid as id,
+ namespace||'/'||name||' ['||oid||']' as filename,
+ length(template) as size,
+ iseditable as writable
+ from template ;
+ },
+ 'read' => q{
+ select template
+ from template
+ where oid = ?;
+ },
+ 'update' => q{
+ update template
+ set template = ?
+ where oid = ?;
+ },
+ },
+ 'mysql' => {
+ 'filenames' => q{
+ select
+ concat(templateid,name) as id,
+ concat(namespace,'/',name,'.html') as filename,
+ length(template) as size,
+ iseditable as writable
+ from template ;
+ },
+ 'read' => q{
+ select template
+ from template
+ where concat(templateid,name) = ?;
+ },
+ 'update' => q{
+ update template
+ set template = ?
+ where concat(templateid,name) = ?;
+ },
+ },
};
+my $dsn = $config->param('dsn');
+my $db;
+
+if ($dsn =~ m/DBI:(mysql|pg):/) {
+ $db = $1;
+} else {
+ print STDERR "can't find supported database (mysql/pg) in dsn: $dsn\n";
+ exit 1;
+}
+
+my $template_dir = $config->param('uploadsPath') . '/temp/templates/';
+
+print "using database '$db', template dir '$template_dir' and mountpoint $mount\n";
+
my $mnt = Fuse::DBI->mount({
- filenames => $sql_filenames,
- read => $sql_read,
- update => $sql_update,
- dsn => 'DBI:Pg:dbname=webgui',
- mount => '/mnt2',
+ filenames => $sql->{$db}->{'filenames'},
+ read => $sql->{$db}->{'read'},
+ update => $sql->{$db}->{'update'},
+ dsn => $config->param('dsn'),
+ user => $config->param('dbuser'),
+ password => $config->param('dbpass'),
+ mount => $mount,
+ fork => 1,
+ invalidate => sub {
+ print STDERR "invalidating content in $template_dir\n";
+ opendir(DIR, $template_dir) || die "can't opendir $template_dir: $!";
+ map { unlink "$template_dir/$_" || warn "can't remove $template_dir/$_: $!" } grep { !/^\./ && -f "$template_dir/$_" } readdir(DIR);
+ closedir DIR;
+ },
});
print "Press enter to exit...";
my $foo = <STDIN>;
$mnt->umount;
+
+=head1 SEE ALSO
+
+C<Fuse::DBI> website
+L<http://www.rot13.org/~dpavlin/fuse_dbi.html>
+
+C<FUSE (Filesystem in USErspace)> website
+L<http://sourceforge.net/projects/avf>
+
+=head1 AUTHOR
+
+Dobrica Pavlinusic, E<lt>dpavlin@rot13.orgE<gt>
+
+=head1 COPYRIGHT AND LICENSE
+
+Copyright (C) 2004 by Dobrica Pavlinusic
+
+This library is free software; you can redistribute it and/or modify
+it under the same terms as Perl itself, either Perl version 5.8.4 or,
+at your option, any later version of Perl 5 you may have available.
+
+=cut
+
projects / Fuse-DBI / commitdiff