path: root/Bugzilla/Attachment.pm

# -*- Mode: perl; indent-tabs-mode: nil -*-
#
# The contents of this file are subject to the Mozilla Public
# License Version 1.1 (the "License"); you may not use this file
# except in compliance with the License. You may obtain a copy of
# the License at http://www.mozilla.org/MPL/
#
# Software distributed under the License is distributed on an "AS
# IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or
# implied. See the License for the specific language governing
# rights and limitations under the License.
#
# The Original Code is the Bugzilla Bug Tracking System.
#
# The Initial Developer of the Original Code is Netscape Communications
# Corporation. Portions created by Netscape are
# Copyright (C) 1998 Netscape Communications Corporation. All
# Rights Reserved.
#
# Contributor(s): Terry Weissman <terry@mozilla.org>
#                 Myk Melez <myk@mozilla.org>
#                 Marc Schumann <wurblzap@gmail.com>

use strict;

package Bugzilla::Attachment;

=head1 NAME

Bugzilla::Attachment - a file related to a bug that a user has uploaded
                       to the Bugzilla server

=head1 SYNOPSIS

  use Bugzilla::Attachment;

  # Get the attachment with the given ID.
  my $attachment = Bugzilla::Attachment->get($attach_id);

  # Get the attachments with the given IDs.
  my $attachments = Bugzilla::Attachment->get_list($attach_ids);

=head1 DESCRIPTION

This module defines attachment objects, which represent files related to bugs
that users upload to the Bugzilla server.

=cut

use Bugzilla::Error;
use Bugzilla::Flag;
use Bugzilla::Config qw(:locations Param);
use Bugzilla::User;
use Bugzilla::Util qw(trick_taint);

sub get {
    my $invocant = shift;
    my $id = shift;

    my $attachments = _retrieve([$id]);
    my $self = $attachments->[0];
    bless($self, ref($invocant) || $invocant) if $self;

    return $self;
}

sub get_list {
    my $invocant = shift;
    my $ids = shift;

    my $attachments = _retrieve($ids);
    foreach my $attachment (@$attachments) {
        bless($attachment, ref($invocant) || $invocant);
    }

    return $attachments;
}

sub _retrieve {
    my ($ids) = @_;

    return [] if scalar(@$ids) == 0;

    my @columns = (
        'attachments.attach_id AS id',
        'attachments.bug_id AS bug_id',
        'attachments.description AS description',
        'attachments.mimetype AS contenttype',
        'attachments.submitter_id AS _attacher_id',
        Bugzilla->dbh->sql_date_format('attachments.creation_ts',
                                       '%Y.%m.%d %H:%i') . " AS attached",
        'attachments.filename AS filename',
        'attachments.ispatch AS ispatch',
        'attachments.isurl AS isurl',
        'attachments.isobsolete AS isobsolete',
        'attachments.isprivate AS isprivate'
    );
    my $columns = join(", ", @columns);

    my $records = Bugzilla->dbh->selectall_arrayref("SELECT $columns
                                                     FROM attachments
                                                     WHERE attach_id IN (" .
                                                     join(",", @$ids) . ")
                                                     ORDER BY attach_id",
                                                    { Slice => {} });
    return $records;
}

=pod

=head2 Instance Properties

=over

=item C<id>

the unique identifier for the attachment

=back

=cut

sub id {
    my $self = shift;
    return $self->{id};
}

=over

=item C<bug_id>

the ID of the bug to which the attachment is attached

=back

=cut

# XXX Once Bug.pm slims down sufficiently this should become a reference
# to a bug object.
sub bug_id {
    my $self = shift;
    return $self->{bug_id};
}

=over

=item C<description>

user-provided text describing the attachment

=back

=cut

sub description {
    my $self = shift;
    return $self->{description};
}

=over

=item C<contenttype>

the attachment's MIME media type

=back

=cut

sub contenttype {
    my $self = shift;
    return $self->{contenttype};
}

=over

=item C<attacher>

the user who attached the attachment

=back

=cut

sub attacher {
    my $self = shift;
    return $self->{attacher} if exists $self->{attacher};
    $self->{attacher} = new Bugzilla::User($self->{_attacher_id});
    return $self->{attacher};
}

=over

=item C<attached>

the date and time on which the attacher attached the attachment

=back

=cut

sub attached {
    my $self = shift;
    return $self->{attached};
}

=over

=item C<filename>

the name of the file the attacher attached

=back

=cut

sub filename {
    my $self = shift;
    return $self->{filename};
}

=over

=item C<ispatch>

whether or not the attachment is a patch

=back

=cut

sub ispatch {
    my $self = shift;
    return $self->{ispatch};
}

=over

=item C<isurl>

whether or not the attachment is a URL

=back

=cut

sub isurl {
    my $self = shift;
    return $self->{isurl};
}

=over

=item C<isobsolete>

whether or not the attachment is obsolete

=back

=cut

sub isobsolete {
    my $self = shift;
    return $self->{isobsolete};
}

=over

=item C<isprivate>

whether or not the attachment is private

=back

=cut

sub isprivate {
    my $self = shift;
    return $self->{isprivate};
}

=over

=item C<data>

the content of the attachment

=back

=cut

sub data {
    my $self = shift;
    return $self->{data} if exists $self->{data};

    # First try to get the attachment data from the database.
    ($self->{data}) = Bugzilla->dbh->selectrow_array("SELECT thedata
                                                      FROM attach_data
                                                      WHERE id = ?",
                                                     undef,
                                                     $self->{id});

    # If there's no attachment data in the database, the attachment is stored
    # in a local file, so retrieve it from there.
    if (length($self->{data}) == 0) {
        if (open(AH, $self->_get_local_filename())) {
            local $/;
            binmode AH;
            $self->{data} = <AH>;
            close(AH);
        }
    }

    return $self->{data};
}

=over

=item C<datasize>

the length (in characters) of the attachment content

=back

=cut

# datasize is a property of the data itself, and it's unclear whether we should
# expose it at all, since you can easily derive it from the data itself: in TT,
# attachment.data.size; in Perl, length($attachment->{data}).  But perhaps
# it makes sense for performance reasons, since accessing the data forces it
# to get retrieved from the database/filesystem and loaded into memory,
# while datasize avoids loading the attachment into memory, calling SQL's
# LENGTH() function or stat()ing the file instead.  I've left it in for now.

sub datasize {
    my $self = shift;
    return $self->{datasize} if exists $self->{datasize};

    # If we have already retrieved the data, return its size.
    return length($self->{data}) if exists $self->{data};

    $self->{datasize} =
        Bugzilla->dbh->selectrow_array("SELECT LENGTH(thedata)
                                        FROM attach_data
                                        WHERE id = ?",
                                       undef, $self->{id}) || 0;

    # If there's no attachment data in the database, either the attachment
    # is stored in a local file, and so retrieve its size from the file,
    # or the attachment has been deleted.
    unless ($self->{datasize}) {
        if (open(AH, $self->_get_local_filename())) {
            binmode AH;
            $self->{datasize} = (stat(AH))[7];
            close(AH);
        }
    }

    return $self->{datasize};
}

=over

=item C<flags>

flags that have been set on the attachment

=back

=cut

sub flags {
    my $self = shift;
    return $self->{flags} if exists $self->{flags};

    $self->{flags} = Bugzilla::Flag::match({ 'attach_id' => $self->id });
    return $self->{flags};
}

# Instance methods; no POD documentation here yet because the only ones so far
# are private.

sub _get_local_filename {
    my $self = shift;
    my $hash = ($self->id % 100) + 100;
    $hash =~ s/.*(\d\d)$/group.$1/;
    return "$attachdir/$hash/attachment." . $self->id;
}

sub _validate_filename {
    my ($throw_error) = @_;
    my $cgi = Bugzilla->cgi;
    defined $cgi->upload('data')
        || ($throw_error ? ThrowUserError("file_not_specified") : return 0);

    my $filename = $cgi->upload('data');

    # Remove path info (if any) from the file name.  The browser should do this
    # for us, but some are buggy.  This may not work on Mac file names and could
    # mess up file names with slashes in them, but them's the breaks.  We only
    # use this as a hint to users downloading attachments anyway, so it's not
    # a big deal if it munges incorrectly occasionally.
    $filename =~ s/^.*[\/\\]//;

    # Truncate the filename to 100 characters, counting from the end of the
    # string to make sure we keep the filename extension.
    $filename = substr($filename, -100, 100);

    return $filename;
}

sub _validate_data {
    my ($throw_error, $hr_vars) = @_;
    my $cgi = Bugzilla->cgi;
    my $maxsize = $cgi->param('ispatch') ? Param('maxpatchsize') : Param('maxattachmentsize');
    $maxsize *= 1024; # Convert from K
    my $fh;
    # Skip uploading into a local variable if the user wants to upload huge
    # attachments into local files.
    if (!$cgi->param('bigfile')) {
        $fh = $cgi->upload('data');
    }
    my $data;

    # We could get away with reading only as much as required, except that then
    # we wouldn't have a size to print to the error handler below.
    if (!$cgi->param('bigfile')) {
        # enable 'slurp' mode
        local $/;
        $data = <$fh>;
    }

    $data
        || ($cgi->param('bigfile'))
        || ($throw_error ? ThrowUserError("zero_length_file") : return 0);

    # Windows screenshots are usually uncompressed BMP files which
    # makes for a quick way to eat up disk space. Let's compress them.
    # We do this before we check the size since the uncompressed version
    # could easily be greater than maxattachmentsize.
    if (Param('convert_uncompressed_images')
        && $cgi->param('contenttype') eq 'image/bmp') {
        require Image::Magick;
        my $img = Image::Magick->new(magick=>'bmp');
        $img->BlobToImage($data);
        $img->set(magick=>'png');
        my $imgdata = $img->ImageToBlob();
        $data = $imgdata;
        $cgi->param('contenttype', 'image/png');
        $$hr_vars->{'convertedbmp'} = 1;
    }

    # Make sure the attachment does not exceed the maximum permitted size
    my $len = $data ? length($data) : 0;
    if ($maxsize && $len > $maxsize) {
        my $vars = { filesize => sprintf("%.0f", $len/1024) };
        if ($cgi->param('ispatch')) {
            $throw_error ? ThrowUserError("patch_too_large", $vars) : return 0;
        }
        else {
            $throw_error ? ThrowUserError("file_too_large", $vars) : return 0;
        }
    }

    return $data || '';
}

=pod

=head2 Class Methods

=over

=item C<get_attachments_by_bug($bug_id)>

Description: retrieves and returns the attachments for the given bug.

Params:     C<$bug_id> - integer - the ID of the bug for which
            to retrieve and return attachments.

Returns:    a reference to an array of attachment objects.

=cut

sub get_attachments_by_bug {
    my ($class, $bug_id) = @_;
    my $attach_ids = Bugzilla->dbh->selectcol_arrayref("SELECT attach_id
                                                        FROM attachments
                                                        WHERE bug_id = ?",
                                                       undef, $bug_id);
    my $attachments = Bugzilla::Attachment->get_list($attach_ids);
    return $attachments;
}

=pod

=item C<validate_is_patch()>

Description: validates the "patch" flag passed in by CGI.

Returns:    1 on success.

=cut

sub validate_is_patch {
    my ($class, $throw_error) = @_;
    my $cgi = Bugzilla->cgi;

    # Set the ispatch flag to zero if it is undefined, since the UI uses
    # an HTML checkbox to represent this flag, and unchecked HTML checkboxes
    # do not get sent in HTML requests.
    $cgi->param('ispatch', $cgi->param('ispatch') ? 1 : 0);

    # Set the content type to text/plain if the attachment is a patch.
    $cgi->param('contenttype', 'text/plain') if $cgi->param('ispatch');

    return 1;
}

=pod

=item C<validate_description()>

Description: validates the description passed in by CGI.

Returns:    1 on success.

=cut

sub validate_description {
    my ($class, $throw_error) = @_;
    my $cgi = Bugzilla->cgi;

    $cgi->param('description')
        || ($throw_error ? ThrowUserError("missing_attachment_description") : return 0);

    return 1;
}

=pod

=item C<validate_content_type()>

Description: validates the content type passed in by CGI.

Returns:    1 on success.

=cut

sub validate_content_type {
    my ($class, $throw_error) = @_;
    my $cgi = Bugzilla->cgi;

    if (!defined $cgi->param('contenttypemethod')) {
        $throw_error ? ThrowUserError("missing_content_type_method") : return 0;
    }
    elsif ($cgi->param('contenttypemethod') eq 'autodetect') {
        my $contenttype =
            $cgi->uploadInfo($cgi->param('data'))->{'Content-Type'};
        # The user asked us to auto-detect the content type, so use the type
        # specified in the HTTP request headers.
        if ( !$contenttype ) {
            $throw_error ? ThrowUserError("missing_content_type") : return 0;
        }
        $cgi->param('contenttype', $contenttype);
    }
    elsif ($cgi->param('contenttypemethod') eq 'list') {
        # The user selected a content type from the list, so use their
        # selection.
        $cgi->param('contenttype', $cgi->param('contenttypeselection'));
    }
    elsif ($cgi->param('contenttypemethod') eq 'manual') {
        # The user entered a content type manually, so use their entry.
        $cgi->param('contenttype', $cgi->param('contenttypeentry'));
    }
    else {
        $throw_error ?
            ThrowCodeError("illegal_content_type_method",
                           { contenttypemethod => $cgi->param('contenttypemethod') }) :
            return 0;
    }

    if ( $cgi->param('contenttype') !~
           /^(application|audio|image|message|model|multipart|text|video)\/.+$/ ) {
        $throw_error ?
            ThrowUserError("invalid_content_type",
                           { contenttype => $cgi->param('contenttype') }) :
            return 0;
    }

    return 1;
}

=pod

=item C<insert_attachment_for_bug($throw_error, $bug_id, $user, $timestamp, $hr_vars)>

Description: inserts an attachment from CGI input for the given bug.

Params:     C<$bug_id> - integer - the ID of the bug for which
            to insert the attachment.
            C<$user> - Bugzilla::User object - the user we're inserting an
            attachment for.
            C<$timestamp> - scalar - timestamp of the insert as returned
            by SELECT NOW().
            C<$hr_vars> - hash reference - reference to a hash of template
            variables.

Returns:    the ID of the new attachment.

=back

=cut

sub insert_attachment_for_bug {
    my ($class, $throw_error, $bug_id, $user, $timestamp, $hr_vars) = @_;

    my $cgi = Bugzilla->cgi;
    my $dbh = Bugzilla->dbh;
    my $attachurl = $cgi->param('attachurl') || '';
    my $data;
    my $filename;
    my $contenttype;
    my $isurl;
    $class->validate_is_patch($throw_error) || return 0;
    $class->validate_description($throw_error) || return 0;

    if (($attachurl =~ /^(http|https|ftp):\/\/\S+/)
         && !(defined $cgi->upload('data'))) {
        $filename = '';
        $data = $attachurl;
        $isurl = 1;
        $contenttype = 'text/plain';
        $cgi->param('ispatch', 0);
        $cgi->delete('bigfile');
    }
    else {
        $filename = _validate_filename($throw_error) || return 0;
        # need to validate content type before data as
        # we now check the content type for image/bmp in _validate_data()
        unless ($cgi->param('ispatch')) {
            $class->validate_content_type($throw_error) || return 0;
        }
        $data = _validate_data($hr_vars, $throw_error) || return 0;
        $contenttype = $cgi->param('contenttype');

        # These are inserted using placeholders so no need to panic
        trick_taint($filename);
        trick_taint($contenttype);
        $isurl = 0;
    }

    # The order of these function calls is important, as both Flag::validate
    # and FlagType::validate assume User::match_field has ensured that the
    # values in the requestee fields are legitimate user email addresses.
    my $match_status = Bugzilla::User::match_field($cgi, {
        '^requestee(_type)?-(\d+)$' => { 'type' => 'multi' },
    }, MATCH_SKIP_CONFIRM);

    $$hr_vars->{'match_field'} = 'requestee';
    if ($match_status == USER_MATCH_FAILED) {
        $$hr_vars->{'message'} = 'user_match_failed';
    }
    elsif ($match_status == USER_MATCH_MULTIPLE) {
        $$hr_vars->{'message'} = 'user_match_multiple';
    }

    # FlagType::validate() and Flag::validate() should not detect
    # any reference to existing flags when creating a new attachment.
    # Setting the third param to -1 will force this function to check this
    # point.
    # XXX needs $throw_error treatment
    Bugzilla::Flag::validate($cgi, $bug_id, -1);
    Bugzilla::FlagType::validate($cgi, $bug_id, -1);

    # Escape characters in strings that will be used in SQL statements.
    my $description = $cgi->param('description');
    trick_taint($description);
    my $isprivate = $cgi->param('isprivate') ? 1 : 0;

    # Insert the attachment into the database.
    my $sth = $dbh->do(
        "INSERT INTO attachments
            (bug_id, creation_ts, filename, description,
             mimetype, ispatch, isurl, isprivate, submitter_id)
         VALUES (?,?,?,?,?,?,?,?,?)", undef, ($bug_id, $timestamp, $filename,
              $description, $contenttype, $cgi->param('ispatch'),
              $isurl, $isprivate, $user->id));
    # Retrieve the ID of the newly created attachment record.
    my $attachid = $dbh->bz_last_key('attachments', 'attach_id');

    # We only use $data here in this INSERT with a placeholder,
    # so it's safe.
    $sth = $dbh->prepare("INSERT INTO attach_data
                         (id, thedata) VALUES ($attachid, ?)");
    trick_taint($data);
    $sth->bind_param(1, $data, $dbh->BLOB_TYPE);
    $sth->execute();

    # If the file is to be stored locally, stream the file from the webserver
    # to the local file without reading it into a local variable.
    if ($cgi->param('bigfile')) {
        my $fh = $cgi->upload('data');
        my $hash = ($attachid % 100) + 100;
        $hash =~ s/.*(\d\d)$/group.$1/;
        mkdir "$attachdir/$hash", 0770;
        chmod 0770, "$attachdir/$hash";
        open(AH, ">$attachdir/$hash/attachment.$attachid");
        binmode AH;
        my $sizecount = 0;
        my $limit = (Param("maxlocalattachment") * 1048576);
        while (<$fh>) {
            print AH $_;
            $sizecount += length($_);
            if ($sizecount > $limit) {
                close AH;
                close $fh;
                unlink "$attachdir/$hash/attachment.$attachid";
                $throw_error ? ThrowUserError("local_file_too_large") : return 0;
            }
        }
        close AH;
        close $fh;
    }
    return $attachid;
}

1;