--- rat/lonuserstate.pm 2000/10/02 16:16:06 1.10 +++ rat/lonuserstate.pm 2020/01/16 21:49:07 1.149.2.2.2.1 @@ -1,241 +1,983 @@ # The LearningOnline Network with CAPA # Construct and maintain state and binary representation of course for user # -# (Server for RAT Maps +# $Id: lonuserstate.pm,v 1.149.2.2.2.1 2020/01/16 21:49:07 raeburn Exp $ # -# (Edit Handler for RAT Maps -# (TeX Content Handler +# Copyright Michigan State University Board of Trustees # -# 05/29/00,05/30 Gerd Kortemeyer) -# 7/1 Gerd Kortemeyer) -# 7/1,7/3,7/4,7/7,7/8,7/10 Gerd Kortemeyer) +# This file is part of the LearningOnline Network with CAPA (LON-CAPA). # -# 7/15,7/17,7/18,8/1,8/2,8/4,8/5,8/21,8/22,8/23,8/30, -# 9/2,9/4,9/29,9/30,10/2 Gerd Kortemeyer +# LON-CAPA is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# LON-CAPA is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with LON-CAPA; if not, write to the Free Software +# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA +# +# /home/httpd/html/adm/gpl.txt +# +# http://www.lon-capa.org/ +# +### package Apache::lonuserstate; +# ------------------------------------------------- modules used by this module use strict; -use Apache::Constants qw(:common :http); -use Apache::File; use HTML::TokeParser; -use Apache::lonnet(); +use Apache::lonnet; +use Apache::lonlocal; +use Apache::loncommon(); use GDBM_File; +use Apache::lonmsg; +use Safe; +use Safe::Hole; +use Opcode; +use Apache::lonenc; +use Fcntl qw(:flock); +use LONCAPA; +use File::Basename; + + # ---------------------------------------------------- Globals for this package -my $pc; # Package counter +my $pc; # Package counter is this what 'Guts' calls the map counter? my %hash; # The big tied hash +my %parmhash;# The hash with the parameters my @cond; # Array with all of the conditions my $errtext; # variable with all errors +my $retfrid; # variable with the very first RID in the course +my $retfurl; # first URL +my %randompick; # randomly picked resources +my %randompickseed; # optional seed for randomly picking resources +my %randomorder; # maps to order contents randomly +my %randomizationcode; # code used to grade folder for bubblesheet exam +my %encurl; # URLs in this folder are supposed to be encrypted +my %hiddenurl; # this URL (or complete folder) is supposed to be hidden +my %rescount; # count of unhidden items in each map +my %mapcount; # count of unhidden maps in each map + +# ----------------------------------- Remove version from URL and store in hash + +sub versionerror { + my ($uri,$usedversion,$unusedversion)=@_; + return '
'.&mt('Version discrepancy: resource [_1] included in both version [_2] and version [_3]. Using version [_2].', + $uri,$usedversion,$unusedversion).'
'; +} + +# Removes the version number from a URI and returns the resulting +# URI (e.g. mumbly.version.stuff => mumbly.stuff). +# +# If the URI has not been seen with a versio before the +# hash{'version_'.resultingURI} is set to the version number. +# If the URI has been seen and the version does not match and error +# is added to the error string. +# +# Parameters: +# URI potentially with a version. +# Returns: +# URI with the version cut out. +# See above for side effects. +# + +sub versiontrack { + my $uri=shift; + if ($uri=~/\.(\d+)\.\w+$/) { + my $version=$1; + $uri=~s/\.\d+\.(\w+)$/\.$1/; + unless ($hash{'version_'.$uri}) { + $hash{'version_'.$uri}=$version; + } elsif ($version!=$hash{'version_'.$uri}) { + $errtext.=&versionerror($uri,$hash{'version_'.$uri},$version); + } + } + return $uri; +} -# --------------------------------------------------------- Loads map from disk +# -------------------------------------------------------------- Put in version -sub loadmap { +sub putinversion { my $uri=shift; - if ($hash{'map_pc_'.$uri}) { return OK; } + my $key=$env{'request.course.id'}.'_'.&Apache::lonnet::clutter($uri); + if ($hash{'version_'.$uri}) { + my $version=$hash{'version_'.$uri}; + if ($version eq 'mostrecent') { return $uri; } + if ($version eq &Apache::lonnet::getversion( + &Apache::lonnet::filelocation('',$uri))) + { return $uri; } + $uri=~s/\.(\w+)$/\.$version\.$1/; + } + &Apache::lonnet::do_cache_new('courseresversion',$key,&Apache::lonnet::declutter($uri),600); + return $uri; +} + +# ----------------------------------------- Processing versions file for course + +sub processversionfile { + my %cenv=@_; + my %versions=&Apache::lonnet::dump('resourceversions', + $cenv{'domain'}, + $cenv{'num'}); + foreach my $ver (keys(%versions)) { + if ($ver=~/^error\:/) { return; } + $hash{'version_'.$ver}=$versions{$ver}; + } +} + +# --------------------------------------------------------- Loads from disk + + +# +# Loads a map file. +# Note that this may implicitly recurse via parse_resource if one of the resources +# is itself composed. +# +# Parameters: +# uri - URI of the map file. +# parent_rid - Resource id in the map of the parent resource (0.0 for the top level map) +# courseid - Course id for the course for which the map is being loaded +# +sub loadmap { + my ($uri,$parent_rid,$courseid)=@_; + + # Is the map already included? + + if ($hash{'map_pc_'.$uri}) { + $errtext.='

'. + &mt('Multiple use of sequence/page [_1]! The course will not function properly.',''.$uri.''). + '

'; + return; + } + # Register the resource in it's map_pc_ [for the URL] + # map_id.nnn is the nesting level -> to the URI. $pc++; my $lpc=$pc; $hash{'map_pc_'.$uri}=$lpc; $hash{'map_id_'.$lpc}=$uri; - my $fn='/home/httpd/html'.$uri; + # If the parent is of the form n.m hang this map underneath it in the + # map hierarchy. - unless (($fn=~/\.sequence$/) || - ($fn=~/\.page$/)) { - $errtext.="Invalid map: $fn\n"; - return OK; - } - - unless (-e $fn) { - my $returned=Apache::lonnet::repcopy($fn); - unless ($returned eq OK) { - $errtext.="Could not import: $fn - "; - if ($returned eq HTTP_SERVICE_UNAVAILABLE) { - $errtext.="Server unavailable\n"; - } - if ($returned eq HTTP_NOT_FOUND) { - $errtext.="File not found\n"; - } - if ($returned eq FORBIDDEN) { - $errtext.="Access forbidden\n"; - } - return OK; - } + if ($parent_rid =~ /^(\d+)\.\d+$/) { + my $parent_pc = $1; + if (defined($hash{'map_hierarchy_'.$parent_pc})) { + $hash{'map_hierarchy_'.$lpc}=$hash{'map_hierarchy_'.$parent_pc}.','. + $parent_pc; + } else { + $hash{'map_hierarchy_'.$lpc}=$parent_pc; + } } - if (-e $fn) { - my @content; - { - my $fh=Apache::File->new($fn); - @content=<$fh>; - } - my $instr=join('',@content); - my $parser = HTML::TokeParser->new(\$instr); - my $token; +# Determine and check filename of the sequence we need to read: - my $linkpc=0; + my $fn=&Apache::lonnet::filelocation('',&putinversion($uri)); - $fn=~/\.(\w+)$/; + my $ispage=($fn=~/\.page$/); - $hash{'map_type_'.$lpc}=$1; + # We can only nest sequences or pages. Anything else is an illegal nest. - while ($token = $parser->get_token) { - if ($token->[0] eq 'S') { - if ($token->[1] eq 'resource') { -# -------------------------------------------------------------------- Resource + unless (($fn=~/\.sequence$/) || $ispage) { + $errtext.='
'.&mt('Invalid map: [_1]',"$fn"); + return; + } - my $rid=$lpc.'.'.$token->[2]->{'id'}; + # Read the XML that constitutes the file. - $hash{'kind_'.$rid}='res'; - $hash{'title_'.$rid}=$token->[2]->{'title'}; - my $turi=$token->[2]->{'src'}; - $hash{'src_'.$rid}=$turi; + my $instr=&Apache::lonnet::getfile($fn); - if (defined($hash{'ids_'.$turi})) { - $hash{'ids_'.$turi}.=','.$rid; - } else { - $hash{'ids_'.$turi}=''.$rid; - } + if ($instr eq -1) { + $errtext.= '
' + .&mt('Map not loaded: The file [_1] does not exist.', + "$fn"); + return; + } - if ($token->[2]->{'src'}=~/\/\//) { - $hash{'ext_'.$rid}='true:'; - } else { - $hash{'ext_'.$rid}='false:'; - } - if ($token->[2]->{'type'}) { - $hash{'type_'.$rid}=$token->[2]->{'type'}; - if ($token->[2]->{'type'} eq 'start') { - $hash{'map_start_'.$uri}="$rid"; - } - if ($token->[2]->{'type'} eq 'finish') { - $hash{'map_finish_'.$uri}="$rid"; - } - } else { - $hash{'type_'.$rid}='normal'; - } + # Successfully got file, parse it - if (($turi=~/\.sequence$/) || - ($turi=~/\.page$/)) { - $hash{'is_map_'.$rid}=1; - &loadmap($turi); - } - - } elsif ($token->[1] eq 'condition') { -# ------------------------------------------------------------------- Condition + # parse for parameter processing. + # Note that these are tags + # so we only care about 'S' (tag start) nodes. - my $rid=$lpc.'.'.$token->[2]->{'id'}; - $hash{'kind_'.$rid}='cond'; - $cond[$#cond+1]=$token->[2]->{'value'}; - $hash{'condid_'.$rid}=$#cond; - if ($token->[2]->{'type'}) { - $cond[$#cond].=':'.$token->[2]->{'type'}; - } else { - $cond[$#cond].=':normal'; - } + my $parser = HTML::TokeParser->new(\$instr); + $parser->attr_encoded(1); - } elsif ($token->[1] eq 'link') { -# ----------------------------------------------------------------------- Links + # first get all parameters - $linkpc++; - my $linkid=$lpc.'.'.$linkpc; - my $goesto=$lpc.'.'.$token->[2]->{'to'}; - my $comesfrom=$lpc.'.'.$token->[2]->{'from'}; - my $undercond=0; + while (my $token = $parser->get_token) { + next if ($token->[0] ne 'S'); + if ($token->[1] eq 'param') { + &parse_param($token,$lpc); + } + } - if ($token->[2]->{'condition'}) { - $undercond=$lpc.'.'.$token->[2]->{'condition'}; - } + # Get set to take another pass through the XML: + # for resources and links. - $hash{'goesto_'.$linkid}=$goesto; - $hash{'comesfrom_'.$linkid}=$comesfrom; - $hash{'undercond_'.$linkid}=$undercond; + $parser = HTML::TokeParser->new(\$instr); + $parser->attr_encoded(1); - if (defined($hash{'to_'.$comesfrom})) { - $hash{'to_'.$comesfrom}.=','.$linkid; - } else { - $hash{'to_'.$comesfrom}=''.$linkid; + my $linkpc=0; + + $fn=~/\.(\w+)$/; + + $hash{'map_type_'.$lpc}=$1; + + my $randomize = ($randomorder{$parent_rid} =~ /^yes$/i); + + # Parse the resources, link and condition tags. + # Note that if randomorder or random select is chosen the links and + # conditions are meaningless but are determined by the randomization. + # This is handled in the next chunk of code. + + my @map_ids; + my $codechecked; + $rescount{$lpc} = 0; + $mapcount{$lpc} = 0; + while (my $token = $parser->get_token) { + next if ($token->[0] ne 'S'); + + # Resource + + if ($token->[1] eq 'resource') { + my $resource_id = &parse_resource($token,$lpc,$ispage,$uri,$courseid); + if (defined $resource_id) { + push(@map_ids, $resource_id); + if ($hash{'src_'.$lpc.'.'.$resource_id}) { + $rescount{$lpc} ++; + if (($hash{'src_'.$lpc.'.'.$resource_id}=~/\.sequence$/) || + ($hash{'src_'.$lpc.'.'.$resource_id}=~/\.page$/)) { + $mapcount{$lpc} ++; } - if (defined($hash{'from_'.$goesto})) { - $hash{'from_'.$goesto}.=','.$linkid; - } else { - $hash{'from_'.$goesto}=''.$linkid; + } + unless ($codechecked) { + my $startsymb = + &Apache::lonnet::encode_symb($hash{'map_id_'.$lpc},$resource_id, + $hash{'src_'."$lpc.$resource_id"}); + my $code = + &Apache::lonnet::EXT('resource.0.examcode',$startsymb,undef,undef, + undef,undef,$courseid); + if ($code) { + $randomizationcode{$parent_rid} = $code; } - } + $codechecked = 1; + } + } - } + # Link + + } elsif ($token->[1] eq 'link' && !$randomize) { + &make_link(++$linkpc,$lpc,$token->[2]->{'to'}, + $token->[2]->{'from'}, + $token->[2]->{'condition'}); # note ..condition may be undefined. + + # condition + + } elsif ($token->[1] eq 'condition' && !$randomize) { + &parse_condition($token,$lpc); + } + } + undef($codechecked); + + # Handle randomization and random selection + + if ($randomize) { + my $advanced; + if ($env{'request.course.id'}) { + $advanced = (&Apache::lonnet::allowed('adv') eq 'F'); + } else { + $env{'request.course.id'} = $courseid; + $advanced = (&Apache::lonnet::allowed('adv') eq 'F'); + $env{'request.course.id'} = ''; } + unless ($advanced) { + # Order of resources is not randomized if user has and advanced role in the course. + my $seed; + + # If the map's random seed parameter has been specified + # it is used as the basis for computing the seed ... + + if (defined($randompickseed{$parent_rid})) { + $seed = $randompickseed{$parent_rid}; + } else { + + # Otherwise the parent's fully encoded symb is used. + + my ($mapid,$resid)=split(/\./,$parent_rid); + my $symb= + &Apache::lonnet::encode_symb($hash{'map_id_'.$mapid}, + $resid,$hash{'src_'.$parent_rid}); + + $seed = $symb; + } + + # TODO: Here for sure we need to pass along the username/domain + # so that we can impersonate users in lonprintout e.g. + + my $setcode; + if (defined($randomizationcode{$parent_rid})) { + if ($env{'form.CODE'} eq '') { + $env{'form.CODE'} = $randomizationcode{$parent_rid}; + $setcode = 1; + } + } + + my $rndseed=&Apache::lonnet::rndseed($seed); + &Apache::lonnet::setup_random_from_rndseed($rndseed); + + if ($setcode) { + undef($env{'form.CODE'}); + undef($setcode); + } + + # Take the set of map ids we have decoded and permute them to a + # random order based on the seed set above. All of this is + # processing the randomorder parameter if it is set, not + # randompick. + + @map_ids=&Math::Random::random_permutation(@map_ids); + } + + my $from = shift(@map_ids); + my $from_rid = $lpc.'.'.$from; + $hash{'map_start_'.$uri} = $from_rid; + $hash{'type_'.$from_rid}='start'; + + # Create links to reflect the random re-ordering done above. + # In the code to process the map XML, we did not process links or conditions + # if randomorder was set. This means that for an instructor to choose + + while (my $to = shift(@map_ids)) { + &make_link(++$linkpc,$lpc,$to,$from); + my $to_rid = $lpc.'.'.$to; + $hash{'type_'.$to_rid}='normal'; + $from = $to; + $from_rid = $to_rid; + } + + $hash{'map_finish_'.$uri}= $from_rid; + $hash{'type_'.$from_rid}='finish'; + } + + $parser = HTML::TokeParser->new(\$instr); + $parser->attr_encoded(1); + + # last parse out the mapalias params. These provide mnemonic + # tags to resources that can be used in conditions + + while (my $token = $parser->get_token) { + next if ($token->[0] ne 'S'); + if ($token->[1] eq 'param') { + &parse_mapalias_param($token,$lpc); + } + } +} + + +# -------------------------------------------------------------------- Resource +# +# Parses a resource tag to produce the value to push into the +# map_ids array. +# +# +# Information about the actual type of resource is provided by the file extension +# of the uri (e.g. .problem, .sequence etc. etc.). +# +# Parameters: +# $token - A token from HTML::TokeParser +# This is an array that describes the most recently parsed HTML item. +# $lpc - Map nesting level (?) +# $ispage - True if this resource is encapsulated in a .page (assembled resourcde). +# $uri - URI of the enclosing resource. +# $courseid - Course id of the course containing the resource being parsed. +# Returns: +# Value of the id attribute of the tag. +# +# Note: +# The token is an array that contains the following elements: +# [0] => 'S' indicating this is a start token +# [1] => 'resource' indicating this tag is a tag. +# [2] => Hash of attribute =>value pairs. +# [3] => @(keys [2]). +# [4] => unused. +# +# The attributes of the resourcde tag include: +# +# id - The resource id. +# src - The URI of the resource. +# type - The resource type (e.g. start and finish). +# title - The resource title. + + +sub parse_resource { + my ($token,$lpc,$ispage,$uri,$courseid) = @_; + + # I refuse to countenance code like this that has + # such a dirty side effect (and forcing this sub to be called within a loop). + # + # if ($token->[2]->{'type'} eq 'zombie') { next; } + # + # The original code both returns _and_ skips to the next pass of the >caller's< + # loop, that's just dirty. + # + + # Zombie resources don't produce anything useful. + + if ($token->[2]->{'type'} eq 'zombie') { + return undef; + } + + my $rid=$lpc.'.'.$token->[2]->{'id'}; # Resource id in hash is levelcounter.id-in-xml. + + # Save the hash element type and title: + + $hash{'kind_'.$rid}='res'; + $hash{'title_'.$rid}=$token->[2]->{'title'}; + + # Get the version free URI for the resource. + # If a 'version' attribute was supplied, and this resource's version + # information has not yet been stored, store it. + # + + my $turi=&versiontrack($token->[2]->{'src'}); + if ($token->[2]->{'version'}) { + unless ($hash{'version_'.$turi}) { + $hash{'version_'.$turi}=$1; + } + } + # Pull out the title and do entity substitution on &colon + # Q: Why no other entity substitutions? + + my $title=$token->[2]->{'title'}; + $title=~s/\&colon\;/\:/gs; + + + + # I think the point of all this code is to construct a final + # URI that apache and its rewrite rules can use to + # fetch the resource. Thi s sonly necessary if the resource + # is not a page. If the resource is a page then it must be + # assembled (at fetch time?). + + unless ($ispage) { + $turi=~/\.(\w+)$/; + my $embstyle=&Apache::loncommon::fileembstyle($1); + if ($token->[2]->{'external'} eq 'true') { # external + $turi=~s/^https?\:\/\//\/adm\/wrapper\/ext\//; + } elsif ($turi=~/^\/*uploaded\//) { # uploaded + if (($embstyle eq 'img') + || ($embstyle eq 'emb') + || ($embstyle eq 'wrp')) { + $turi='/adm/wrapper'.$turi; + } elsif ($embstyle eq 'ssi') { + #do nothing with these + } elsif ($turi!~/\.(sequence|page)$/) { + $turi='/adm/coursedocs/showdoc'.$turi; + } + } elsif ($turi=~ m{^/adm/$match_domain/$match_courseid/\d+/ext\.tool$}) { + $turi='/adm/wrapper'.$turi; + } elsif ($turi=~/\S/) { # normal non-empty internal resource + my $mapdir=$uri; + $mapdir=~s/[^\/]+$//; + $turi=&Apache::lonnet::hreflocation($mapdir,$turi); + if (($embstyle eq 'img') + || ($embstyle eq 'emb') + || ($embstyle eq 'wrp')) { + $turi='/adm/wrapper'.$turi; + } + } + } + # Store reverse lookup, remove query string resource 'ids'_uri => resource id. + # If the URI appears more than one time in the sequence, it's resourcde + # id's are constructed as a comma spearated list. + + my $idsuri=$turi; + $idsuri=~s/\?.+$//; + if (defined($hash{'ids_'.$idsuri})) { + $hash{'ids_'.$idsuri}.=','.$rid; + } else { + $hash{'ids_'.$idsuri}=''.$rid; + } + + + + if ($turi=~/\/(syllabus|aboutme|navmaps|smppg|bulletinboard|viewclasslist)$/) { + $turi.='?register=1'; + } + + + # resource id lookup: 'src'_resourc-di => URI decorated with a query + # parameter as above if necessary due to the resource type. + + $hash{'src_'.$rid}=$turi; + + # Mark the external-ness of the resource: + + if ($token->[2]->{'external'} eq 'true') { + $hash{'ext_'.$rid}='true:'; + } else { + $hash{'ext_'.$rid}='false:'; + } + + # If the resource is a start/finish resource set those + # entries in the has so that navigation knows where everything starts. + # TODO? If there is a malformed sequence that has no start or no finish + # resource, should this be detected and errors thrown? How would such a + # resource come into being other than being manually constructed by a person + # and then uploaded? Could that happen if an author decided a sequence was almost + # right edited it by hand and then reuploaded it to 'fix it' but accidently cut the + # start or finish resources? + # + # All resourcess also get a type_id => (start | finish | normal) hash entr. + # + if ($token->[2]->{'type'}) { + $hash{'type_'.$rid}=$token->[2]->{'type'}; + if ($token->[2]->{'type'} eq 'start') { + $hash{'map_start_'.$uri}="$rid"; + } + if ($token->[2]->{'type'} eq 'finish') { + $hash{'map_finish_'.$uri}="$rid"; + } + } else { + $hash{'type_'.$rid}='normal'; + } + + # Sequences end pages are constructed entities. They require that the + # map that defines _them_ be loaded as well into the hash...with this resourcde + # as the base of the nesting. + # Resources like that are also marked with is_map_id => 1 entries. + # + + if (($turi=~/\.sequence$/) || + ($turi=~/\.page$/)) { + $hash{'is_map_'.$rid}=1; + &loadmap($turi,$rid,$courseid); + } + return $token->[2]->{'id'}; +} + +#-------------------------------------------------------------------- link +# Links define how you are allowed to move from one resource to another. +# They are the transition edges in the directed graph that a map is. +# This sub takes informatino from a tag and constructs the +# navigation bits and pieces of a map. There is no requirement that the +# resources that are linke are already defined, however clearly the map is +# badly broken if they are not _eventually_ defined. +# +# Note that links can be unconditional or conditional. +# +# Parameters: +# linkpc - The link counter for this level of map nesting (this is +# reset to zero by loadmap prior to starting to process +# links for map). +# lpc - The map level ocounter (how deeply nested this map is in +# the hierarchy of maps that are recursively read in. +# to - resource id (within the XML) of the target of the edge. +# from - resource id (within the XML) of the source of the edge. +# condition- id of condition associated with the edge (also within the XML). +# + +sub make_link { + my ($linkpc,$lpc,$to,$from,$condition) = @_; + + # Compute fully qualified ids for the link, the + # and from/to by prepending lpc. + # + + my $linkid=$lpc.'.'.$linkpc; + my $goesto=$lpc.'.'.$to; + my $comesfrom=$lpc.'.'.$from; + my $undercond=0; + + + # If there is a condition, qualify it with the level counter. + + if ($condition) { + $undercond=$lpc.'.'.$condition; + } + + # Links are represnted by: + # goesto_.fuullyqualifedlinkid => fully qualified to + # comesfrom.fullyqualifiedlinkid => fully qualified from + # undercond_.fullyqualifiedlinkid => fully qualified condition id. + + $hash{'goesto_'.$linkid}=$goesto; + $hash{'comesfrom_'.$linkid}=$comesfrom; + $hash{'undercond_'.$linkid}=$undercond; + + # In addition: + # to_.fully qualified from => comma separated list of + # link ids with that from. + # Similarly: + # from_.fully qualified to => comma separated list of link ids` + # with that to. + # That allows us given a resource id to know all edges that go to it + # and leave from it. + # + + if (defined($hash{'to_'.$comesfrom})) { + $hash{'to_'.$comesfrom}.=','.$linkid; + } else { + $hash{'to_'.$comesfrom}=''.$linkid; + } + if (defined($hash{'from_'.$goesto})) { + $hash{'from_'.$goesto}.=','.$linkid; + } else { + $hash{'from_'.$goesto}=''.$linkid; + } +} + +# ------------------------------------------------------------------- Condition +# +# Processes tags, storing sufficient information about them +# in the hash so that they can be evaluated and used to conditionalize +# what is presented to the student. +# +# these can have the following attributes +# +# id = A unique identifier of the condition within the map. +# +# value = Is a perl script-let that, when evaluated in safe space +# determines whether or not the condition is true. +# Normally this takes the form of a test on an Apache::lonnet::EXT call +# to find the value of variable associated with a resource in the +# map identified by a mapalias. +# Here's a fragment of XML code that illustrates this: +# +# +# +# +# +# +# +# In this fragment: +# - The param tag establishes an alias to resource id 5 of 'mainproblem'. +# - The resource that is the start of the map is identified. +# - The resource tag identifies the resource associated with this tag +# and gives it the id 5. +# - The condition is true if the tries variable associated with mainproblem +# is less than 2 (that is the user has had more than 2 tries). +# The condition type is a stop condition which inhibits(?) the associated +# link if the condition is false. +# - The link to resource 5 from resource 1 is affected by this condition. +# +# type = Type of the condition. The type determines how the condition affects the +# link associated with it and is one of +# - 'force' +# - 'stop' +# anything else including not supplied..which treated as: +# - 'normal'. +# Presumably maps get created by the resource assembly tool and therefore +# illegal type values won't squirm their way into the XML. +# +# Side effects: +# - The kind_level-qualified-condition-id hash element is set to 'cond'. +# - The condition text is pushed into the cond array and its element number is +# set in the condid_level-qualified-condition-id element of the hash. +# - The condition type is colon appneded to the cond array element for this condition. +sub parse_condition { + my ($token,$lpc) = @_; + my $rid=$lpc.'.'.$token->[2]->{'id'}; + + $hash{'kind_'.$rid}='cond'; + + my $condition = $token->[2]->{'value'}; + $condition =~ s/[\n\r]+/ /gs; + push(@cond, $condition); + $hash{'condid_'.$rid}=$#cond; + if ($token->[2]->{'type'}) { + $cond[$#cond].=':'.$token->[2]->{'type'}; + } else { + $cond[$#cond].=':normal'; + } +} + +# ------------------------------------------------------------------- Parameter +# Parse a tag in the map. +# Parmameters: +# $token Token array for a start tag from HTML::TokeParser +# [0] = 'S' +# [1] = tagname ("param") +# [2] = Hash of {attribute} = values. +# [3] = Array of the keys in [2]. +# [4] = unused. +# $lpc Current map nesting level.a +# +# Typical attributes: +# to=n - Number of the resource the parameter applies to. +# type=xx - Type of parameter value (e.g. string_yesno or int_pos). +# name=xxx - Name of parameter (e.g. parameter_randompick or parameter_randomorder). +# value=xxx - value of the parameter. + +sub parse_param { + my ($token,$lpc) = @_; + my $referid=$lpc.'.'.$token->[2]->{'to'}; # Resource param applies to. + my $name=$token->[2]->{'name'}; # Name of parameter + my $part; + + if ($name=~/^parameter_(.*)_/) { + $part=$1; } else { - $errtext.='Map not loaded: The file does not exist. '; + $part=0; + } + + # Peel the parameter_ off the parameter name. + + $name=~s/^.*_([^_]*)$/$1/; + + # The value is: + # type.part.name.value + + my $newparam= + &escape($token->[2]->{'type'}).':'. + &escape($part.'.'.$name).'='. + &escape($token->[2]->{'value'}); + + # The hash key is param_resourceid. + # Multiple parameters for a single resource are & separated in the hash. + + + if (defined($hash{'param_'.$referid})) { + $hash{'param_'.$referid}.='&'.$newparam; + } else { + $hash{'param_'.$referid}=''.$newparam; + } + # + # These parameters have to do with randomly selecting + # resources, therefore a separate hash is also created to + # make it easy to locate them when actually computing the resource set later on + # See the code conditionalized by ($randomize) in loadmap(). + + if ($token->[2]->{'name'}=~/^parameter_(0_)*randompick$/) { # Random selection turned on + $randompick{$referid}=$token->[2]->{'value'}; + } + if ($token->[2]->{'name'}=~/^parameter_(0_)*randompickseed$/) { # Randomseed provided. + $randompickseed{$referid}=$token->[2]->{'value'}; + } + if ($token->[2]->{'name'}=~/^parameter_(0_)*randomorder$/) { # Random order turned on. + $randomorder{$referid}=$token->[2]->{'value'}; + } + + # These parameters have to do with how the URLs of resources are presented to + # course members(?). encrypturl presents encypted url's while + # hiddenresource hides the URL. + # + + if ($token->[2]->{'name'}=~/^parameter_(0_)*encrypturl$/) { + if ($token->[2]->{'value'}=~/^yes$/i) { + $encurl{$referid}=1; + } + } + if ($token->[2]->{'name'}=~/^parameter_(0_)*hiddenresource$/) { + if ($token->[2]->{'value'}=~/^yes$/i) { + $hiddenurl{$referid}=1; + } + } +} +# +# Parse mapalias parameters. +# these are tags of the form: +# +# A map alias is a textual name for a resource: +# - The to attribute identifies the resource (this gets level qualified below) +# - The value attributes provides the alias string. +# - name must be of the regexp form: /^parameter_(0_)*mapalias$/ +# - e.g. the string 'parameter_' followed by 0 or more "0_" strings +# terminating with the string 'mapalias'. +# Examples: +# 'parameter_mapalias', 'parameter_0_mapalias', parameter_0_0_mapalias' +# Invalid to ids are silently ignored. +# +# Parameters: +# token - The token array fromthe HMTML::TokeParser +# lpc - The current map level counter. +# +sub parse_mapalias_param { + my ($token,$lpc) = @_; + + # Fully qualify the to value and ignore the alias if there is no + # corresponding resource. + + my $referid=$lpc.'.'.$token->[2]->{'to'}; + return if (!exists($hash{'src_'.$referid})); + + # If this is a valid mapalias parameter, + # Append the target id to the count_mapalias element for that + # alias so that we can detect doubly defined aliases + # e.g.: + # + # + # + # The example above is trivial but the case that's important has to do with + # constructing a map that includes a nested map where the nested map may have + # aliases that conflict with aliases established in the enclosing map. + # + # ...and create/update the hash mapalias entry to actually store the alias. + # + + if ($token->[2]->{'name'}=~/^parameter_(0_)*mapalias$/) { + &count_mapalias($token->[2]->{'value'},$referid); + $hash{'mapalias_'.$token->[2]->{'value'}}=$referid; } } # --------------------------------------------------------- Simplify expression + +# +# Someone should really comment this to describe what it does to what and why. +# sub simplify { - my $expression=shift; + my $expression=shift; +# (0&1) = 1 + $expression=~s/\(0\&([_\.\d]+)\)/$1/g; # (8)=8 - $expression=~s/\((\d+)\)/$1/g; + $expression=~s/\(([_\.\d]+)\)/$1/g; # 8&8=8 - $expression=~s/(\D)(\d+)\&\2(\D)/$1$2$3/g; + $expression=~s/([^_\.\d])([_\.\d]+)\&\2([^_\.\d])/$1$2$3/g; # 8|8=8 - $expression=~s/(\D)(\d+)\|\2(\D)/$1$2$3/g; + $expression=~s/([^_\.\d])([_\.\d]+)(?:\|\2)+([^_\.\d])/$1$2$3/g; # (5&3)&4=5&3&4 - $expression=~s/\((\d+)((?:\&\d+)+)\)\&(\d+\D)/$1$2\&$3/g; + $expression=~s/\(([_\.\d]+)((?:\&[_\.\d]+)+)\)\&([_\.\d]+[^_\.\d])/$1$2\&$3/g; # (((5&3)|(4&6)))=((5&3)|(4&6)) - $expression=~ - s/\((\(\(\d+(?:\&\d+)*\)(?:\|\(\d+(?:\&\d+)*\))+\))\)/$1/g; + $expression=~ + s/\((\(\([_\.\d]+(?:\&[_\.\d]+)*\)(?:\|\([_\.\d]+(?:\&[_\.\d]+)*\))+\))\)/$1/g; # ((5&3)|(4&6))|(1&2)=(5&3)|(4&6)|(1&2) - $expression=~ - s/\((\(\d+(?:\&\d+)*\))((?:\|\(\d+(?:\&\d+)*\))+)\)\|(\(\d+(?:\&\d+)*\))/\($1$2\|$3\)/g; - return $expression; + $expression=~ + s/\((\([_\.\d]+(?:\&[_\.\d]+)*\))((?:\|\([_\.\d]+(?:\&[_\.\d]+)*\))+)\)\|(\([_\.\d]+(?:\&[_\.\d]+)*\))/\($1$2\|$3\)/g; + return $expression; } # -------------------------------------------------------- Build condition hash +# +# Traces a route recursively through the map after it has been loaded +# (I believe this really visits each resourcde that is reachable fromt he +# start top node. +# +# - Marks hidden resources as hidden. +# - Marks which resource URL's must be encrypted. +# - Figures out (if necessary) the first resource in the map. +# - Further builds the chunks of the big hash that define how +# conditions work +# +# Note that the tracing strategy won't visit resources that are not linked to +# anything or islands in the map (groups of resources that form a path but are not +# linked in to the path that can be traced from the start resource...but that's ok +# because by definition, those resources are not reachable by users of the course. +# +# Parameters: +# sofar - _URI of the prior entry or 0 if this is the top. +# rid - URI of the resource to visit. +# beenhere - list of resources (each resource enclosed by &'s) that have +# already been visited. +# encflag - If true the resource that resulted in a recursive call to us +# has an encoded URL (which means contained resources should too). +# hdnflag - If true,the resource that resulted in a recursive call to us +# was hidden (which means contained resources should be hidden too). +# Returns +# new value indicating how far the map has been traversed (the sofar). +# sub traceroute { - my ($sofar,$rid,$beenhere)=@_; - $sofar=simplify($sofar); - unless ($beenhere=~/\&$rid\&/) { - $beenhere.=$rid.'&'; - if (defined($hash{'conditions_'.$rid})) { - $hash{'conditions_'.$rid}=simplify( + my ($sofar,$rid,$beenhere,$encflag,$hdnflag)=@_; + my $newsofar=$sofar=simplify($sofar); + + unless ($beenhere=~/\&\Q$rid\E\&/) { + $beenhere.=$rid.'&'; + my ($mapid,$resid)=split(/\./,$rid); + my $symb=&Apache::lonnet::encode_symb($hash{'map_id_'.$mapid},$resid,$hash{'src_'.$rid}); + my $hidden=&Apache::lonnet::EXT('resource.0.hiddenresource',$symb); + + if ($hdnflag || lc($hidden) eq 'yes') { + $hiddenurl{$rid}=1; + } + if (!$hdnflag && lc($hidden) eq 'no') { + delete($hiddenurl{$rid}); + } + + my $encrypt=&Apache::lonnet::EXT('resource.0.encrypturl',$symb); + if ($encflag || lc($encrypt) eq 'yes') { $encurl{$rid}=1; } + + if (($retfrid eq '') && ($hash{'src_'.$rid}) + && ($hash{'src_'.$rid}!~/\.sequence$/)) { + $retfrid=$rid; + } + + if (defined($hash{'conditions_'.$rid})) { + $hash{'conditions_'.$rid}=simplify( '('.$hash{'conditions_'.$rid}.')|('.$sofar.')'); - } else { - $hash{'conditions_'.$rid}=$sofar; - } - if (defined($hash{'is_map_'.$rid})) { - if (defined($hash{'map_start_'.$hash{'src_'.$rid}})) { - &traceroute($sofar,$hash{'map_start_'.$hash{'src_'.$rid}},'&'); - if (defined($hash{'map_finish_'.$hash{'src_'.$rid}})) { - $sofar= - $hash{'conditions_'.$hash{'map_finish_'.$hash{'src_'.$rid}}}; - } - } - } - if (defined($hash{'to_'.$rid})) { - map { + } else { + $hash{'conditions_'.$rid}=$sofar; + } + + # if the expression is just the 0th condition keep it + # otherwise leave a pointer to this condition expression + + $newsofar = ($sofar eq '0') ? $sofar : '_'.$rid; + + # Recurse if the resource is a map: + + if (defined($hash{'is_map_'.$rid})) { + if (defined($hash{'map_start_'.$hash{'src_'.$rid}})) { + $sofar=$newsofar= + &traceroute($sofar, + $hash{'map_start_'.$hash{'src_'.$rid}}, + $beenhere, + $encflag || $encurl{$rid}, + $hdnflag || $hiddenurl{$rid}); + } + } + + # Processes links to this resource: + # - verify the existence of any conditionals on the link to here. + # - Recurse to any resources linked to us. + # + if (defined($hash{'to_'.$rid})) { + foreach my $id (split(/\,/,$hash{'to_'.$rid})) { my $further=$sofar; - if ($hash{'undercond_'.$_}) { - if (defined($hash{'condid_'.$hash{'undercond_'.$_}})) { - $further=simplify('('.$further.')&('. - $hash{'condid_'.$hash{'undercond_'.$_}}.')'); - } else { - $errtext.='Undefined condition ID: ' - .$hash{'undercond_'.$_}.'. '; - } - } - &traceroute($further,$hash{'goesto_'.$_},$beenhere); - } split(/\,/,$hash{'to_'.$rid}); - } + # + # If there's a condition associated with this link be sure + # it's been defined else that's an error: + # + if ($hash{'undercond_'.$id}) { + if (defined($hash{'condid_'.$hash{'undercond_'.$id}})) { + $further=simplify('('.'_'.$rid.')&('. + $hash{'condid_'.$hash{'undercond_'.$id}}.')'); + } else { + $errtext.= '
'. + &mt('Undefined condition ID: [_1]', + $hash{'undercond_'.$id}); + } + } + # Recurse to resoruces that have to's to us. + $newsofar=&traceroute($further,$hash{'goesto_'.$id},$beenhere, + $encflag,$hdnflag); + } + } } + return $newsofar; } -# ------------------------------------------ Cascading conditions, quick access +# ------------------------------ Cascading conditions, quick access, parameters + +# +# Seems a rather strangely named sub given what the comment above says it does. + sub accinit { my ($uri,$short,$fn)=@_; @@ -243,108 +985,712 @@ sub accinit { my %captured=(); my $condcounter=0; $acchash{'acc.cond.'.$short.'.0'}=0; - map { - if ($_=~/^conditions/) { - my $expr=$hash{$_}; - map { - my $sub=$_; - my $orig=$_; - $sub=~/\(\((\d+\&(:?\d+\&)*)(?:\d+\&*)+\)(?:\|\(\1(?:\d+\&*)+\))+\)/; - my $factor=$1; - $sub=~s/$factor//g; - $sub=~s/^\(/\($factor\(/; - $sub.=')'; - $sub=simplify($sub); - $orig=~s/(\W)/\\$1/g; - $expr=~s/$orig/$sub/; - } ($expr=~m/(\(\(\d+(?:\&\d+)+\)(?:\|\(\d+(?:\&\d+)+\))+\))/g); - $hash{$_}=$expr; - unless (defined($captured{$expr})) { - $condcounter++; - $captured{$expr}=$condcounter; - $acchash{'acc.cond.'.$short.'.'.$condcounter}=$expr; - } - } - } keys %hash; - map { - if ($_=~/^ids/) { - my $resid=$hash{$_}; - my $uri=$hash{'src_'.$resid}; - my @uriparts=split(/\//,$uri); - my $urifile=$uriparts[$#uriparts]; - $#uriparts--; - my $uripath=join('/',@uriparts); - $uripath=~s/^\/res\///; - if (defined($hash{'conditions_'.$resid})) { - $urifile.=':'.$captured{$hash{'conditions_'.$resid}}; - } else { - $urifile.=':0'; + + # This loop is only interested in conditions and + # parameters in the big hash: + + foreach my $key (keys(%hash)) { + + # conditions: + + if ($key=~/^conditions/) { + my $expr=$hash{$key}; + + # try to find and factor out common sub-expressions + # Any subexpression that is found is simplified, removed from + # the original condition expression and the simplified sub-expression + # substituted back in to the epxression..I'm not actually convinced this + # factors anything out...but instead maybe simplifies common factors(?) + + foreach my $sub ($expr=~m/(\(\([_\.\d]+(?:\&[_\.\d]+)+\)(?:\|\([_\.\d]+(?:\&[_\.\d]+)+\))+\))/g) { + my $orig=$sub; + + my ($factor) = ($sub=~/\(\(([_\.\d]+\&(:?[_\.\d]+\&)*)(?:[_\.\d]+\&*)+\)(?:\|\(\1(?:[_\.\d]+\&*)+\))+\)/); + next if (!defined($factor)); + + $sub=~s/\Q$factor\E//g; + $sub=~s/^\(/\($factor\(/; + $sub.=')'; + $sub=simplify($sub); + $expr=~s/\Q$orig\E/$sub/; + } + $hash{$key}=$expr; + + # If not yet seen, record in acchash and that we've seen it. + + unless (defined($captured{$expr})) { + $condcounter++; + $captured{$expr}=$condcounter; + $acchash{'acc.cond.'.$short.'.'.$condcounter}=$expr; + } + # Parameters: + + } elsif ($key=~/^param_(\d+)\.(\d+)/) { + my $prefix=&Apache::lonnet::encode_symb($hash{'map_id_'.$1},$2, + $hash{'src_'.$1.'.'.$2}); + foreach my $param (split(/\&/,$hash{$key})) { + my ($typename,$value)=split(/\=/,$param); + my ($type,$name)=split(/\:/,$typename); + $parmhash{$prefix.'.'.&unescape($name)}= + &unescape($value); + $parmhash{$prefix.'.'.&unescape($name).'.type'}= + &unescape($type); + } + } + } + # This loop only processes id entries in the big hash. + + foreach my $key (keys(%hash)) { + if ($key=~/^ids/) { + foreach my $resid (split(/\,/,$hash{$key})) { + my $uri=$hash{'src_'.$resid}; + my ($uripath,$urifile) = + &Apache::lonnet::split_uri_for_cond($uri); + if ($uripath) { + my $uricond='0'; + if (defined($hash{'conditions_'.$resid})) { + $uricond=$captured{$hash{'conditions_'.$resid}}; + } + if (defined($acchash{'acc.res.'.$short.'.'.$uripath})) { + if ($acchash{'acc.res.'.$short.'.'.$uripath}=~ + /(\&\Q$urifile\E\:[^\&]*)/) { + my $replace=$1; + my $regexp=$replace; + #$regexp=~s/\|/\\\|/g; + $acchash{'acc.res.'.$short.'.'.$uripath} =~ + s/\Q$regexp\E/$replace\|$uricond/; + } else { + $acchash{'acc.res.'.$short.'.'.$uripath}.= + $urifile.':'.$uricond.'&'; + } + } else { + $acchash{'acc.res.'.$short.'.'.$uripath}= + '&'.$urifile.':'.$uricond.'&'; + } + } + } + } + } + $acchash{'acc.res.'.$short.'.'}='&:0&'; + my $courseuri=$uri; + $courseuri=~s/^\/res\///; + my $regexp = 1; + &Apache::lonnet::delenv('(acc\.|httpref\.)',$regexp); + &Apache::lonnet::appenv(\%acchash); +} + +# ---------------- Selectively delete from randompick maps and hidden url parms + +sub hiddenurls { + my $randomoutentry=''; + foreach my $rid (keys(%randompick)) { + my $rndpick=$randompick{$rid}; + my $mpc=$hash{'map_pc_'.$hash{'src_'.$rid}}; +# ------------------------------------------- put existing resources into array + my @currentrids=(); + foreach my $key (sort(keys(%hash))) { + if ($key=~/^src_($mpc\.\d+)/) { + if ($hash{'src_'.$1}) { push @currentrids, $1; } } - if (defined($acchash{'acc.res.'.$short.'.'.$uripath})) { - $acchash{'acc.res.'.$short.'.'.$uripath}.=$urifile.'&'; - } else { - $acchash{'acc.res.'.$short.'.'.$uripath}='&'.$urifile.'&'; + } + # rids are number.number and we want to numercially sort on + # the second number + @currentrids=sort { + my (undef,$aid)=split(/\./,$a); + my (undef,$bid)=split(/\./,$b); + $aid <=> $bid; + } @currentrids; + next if ($#currentrids<$rndpick); +# -------------------------------- randomly eliminate the ones that should stay + my (undef,$id)=split(/\./,$rid); + if ($randompickseed{$rid}) { $id=$randompickseed{$rid}; } + my $setcode; + if (defined($randomizationcode{$rid})) { + if ($env{'form.CODE'} eq '') { + $env{'form.CODE'} = $randomizationcode{$rid}; + $setcode = 1; } } - } keys %hash; - my $courseuri=$uri; - $courseuri=~s/^\/res\///; - &Apache::lonnet::appenv(%acchash, - "request.course.id" => $short, - "request.course.fn" => $fn, - "request.course.uri" => $courseuri); + my $rndseed=&Apache::lonnet::rndseed($id); # use id instead of symb + if ($setcode) { + undef($env{'form.CODE'}); + undef($setcode); + } + &Apache::lonnet::setup_random_from_rndseed($rndseed); + my @whichids=&Math::Random::random_permuted_index($#currentrids+1); + for (my $i=1;$i<=$rndpick;$i++) { $currentrids[$whichids[$i]]=''; } + #&Apache::lonnet::logthis("$id,$rndseed,".join(':',@whichids)); +# -------------------------------------------------------- delete the leftovers + for (my $k=0; $k<=$#currentrids; $k++) { + if ($currentrids[$k]) { + $hash{'randomout_'.$currentrids[$k]}=1; + my ($mapid,$resid)=split(/\./,$currentrids[$k]); + if ($rescount{$mapid}) { + $rescount{$mapid} --; + } + if ($hash{'is_map_'.$currentrids[$k]}) { + if ($mapcount{$mapid}) { + $mapcount{$mapid} --; + } + } + $randomoutentry.='&'. + &Apache::lonnet::encode_symb($hash{'map_id_'.$mapid}, + $resid, + $hash{'src_'.$currentrids[$k]} + ).'&'; + } + } + } +# ------------------------------ take care of explicitly hidden urls or folders + foreach my $rid (keys(%hiddenurl)) { + $hash{'randomout_'.$rid}=1; + my ($mapid,$resid)=split(/\./,$rid); + if ($rescount{$mapid}) { + $rescount{$mapid} --; + } + if ($hash{'is_map_'.$rid}) { + if ($mapcount{$mapid}) { + $mapcount{$mapid} --; + } + } + $randomoutentry.='&'. + &Apache::lonnet::encode_symb($hash{'map_id_'.$mapid},$resid, + $hash{'src_'.$rid}).'&'; + } +# --------------------------------------- append randomout entry to environment + if ($randomoutentry) { + &Apache::lonnet::appenv({'acc.randomout' => $randomoutentry}); + } +} + +# -------------------------------------- populate big hash with map breadcrumbs + +# Create map_breadcrumbs_$pc from map_hierarchy_$pc by omitting intermediate +# maps not shown in Course Contents table. + +sub mapcrumbs { + foreach my $key (keys(%rescount)) { + if ($hash{'map_hierarchy_'.$key}) { + my $skipnext = 0; + foreach my $id (split(/,/,$hash{'map_hierarchy_'.$key}),$key) { + unless ($skipnext) { + $hash{'map_breadcrumbs_'.$key} .= "$id,"; + } + unless (($id == 0) || ($id == 1)) { + if ((!$rescount{$id}) || ($rescount{$id} == 1 && $mapcount{$id} == 1)) { + $skipnext = 1; + } else { + $skipnext = 0; + } + } + } + $hash{'map_breadcrumbs_'.$key} =~ s/,$//; + } + } } # ---------------------------------------------------- Read map and all submaps sub readmap { - my $short=shift; - $short=~s/^\///; - my %cenv=&Apache::lonnet::coursedescription($short); - my $fn=$cenv{'fn'}; - my $uri; - $short=~s/\//\_/g; - unless ($uri=$cenv{'url'}) { - &Apache::lonnet::logthis("WARNING: ". - "Could not load course $short."); - return 'No course data available.'; - } - @cond=('true:normal'); - if (tie(%hash,'GDBM_File',"$fn.db",&GDBM_WRCREAT,0640)) { - %hash=(); + my $short=shift; + $short=~s/^\///; + + # TODO: Hidden dependency on current user: + + my %cenv=&Apache::lonnet::coursedescription($short,{'freshen_cache'=>1}); + + my $fn=$cenv{'fn'}; + my $uri; + $short=~s/\//\_/g; + unless ($uri=$cenv{'url'}) { + &Apache::lonnet::logthis('WARNING: '. + "Could not load course $short."); + return ('',&mt('No course data available.'));; + } + @cond=('true:normal'); + + unless (open(LOCKFILE,">","$fn.db.lock")) { + # + # Most likely a permissions problem on the lockfile or its directory. + # + $retfurl = ''; + return ($retfurl,'
'.&mt('Map not loaded - Lock file could not be opened when reading map:').' '.$fn.'.'); + } + my $lock=0; + my $gotstate=0; + + # If we can get the lock without delay any files there are idle + # and from some prior request. We'll kill them off and regenerate them: + + if (flock(LOCKFILE,LOCK_EX|LOCK_NB)) { + $lock=1; # Remember that we hold the lock. + &unlink_tmpfiles($fn); + } + undef %randompick; + undef %randompickseed; + undef %randomorder; + undef %randomizationcode; + undef %hiddenurl; + undef %encurl; + undef %rescount; + undef %mapcount; + $retfrid=''; $errtext=''; + my ($untiedhash,$untiedparmhash,$tiedhash,$tiedparmhash); # More state flags. + + # if we got the lock, regenerate course regnerate empty files and tie them. + + if ($lock) { + if (tie(%hash,'GDBM_File',"$fn.db",&GDBM_WRCREAT(),0640)) { + $tiedhash = 1; + if (tie(%parmhash,'GDBM_File',$fn.'_parms.db',&GDBM_WRCREAT(),0640)) { + $tiedparmhash = 1; + $gotstate = &build_tmp_hashes($uri, + $fn, + $short, + \%cenv); # TODO: Need to provide requested user@dom + unless ($gotstate) { + &Apache::lonnet::logthis('Failed to write statemap at first attempt '.$fn.' for '.$uri.'.'); + } + $untiedparmhash = untie(%parmhash); + unless ($untiedparmhash) { + &Apache::lonnet::logthis('WARNING: '. + 'Could not untie coursemap parmhash '.$fn.' for '.$uri.'.'); + } + } + $untiedhash = untie(%hash); + unless ($untiedhash) { + &Apache::lonnet::logthis('WARNING: '. + 'Could not untie coursemap hash '.$fn.' for '.$uri.'.'); + } + } + flock(LOCKFILE,LOCK_UN); # RF: this is what I don't get unless there are other + # unlocked places the remainder happens..seems like if we + # just kept the lock here the rest of the code would have + # been much easier? + } + unless ($lock && $tiedhash && $tiedparmhash) { + # if we are here it is likely because we are already trying to + # initialize the course in another child, busy wait trying to + # tie the hashes for the next 90 seconds, if we succeed forward + # them on to navmaps, if we fail, throw up the Could not init + # course screen + # + # RF: I'm not seeing the case where the ties/unties can fail in a way + # that can be remedied by this. Since we owned the lock seems + # Tie/untie failures are a result of something like a permissions problem instead? + # + + # In any vent, undo what we did manage to do above first: + if ($lock) { + # Got the lock but not the DB files + flock(LOCKFILE,LOCK_UN); + $lock = 0; + } + if ($tiedhash) { + unless($untiedhash) { + untie(%hash); + } + } + if ($tiedparmhash) { + unless($untiedparmhash) { + untie(%parmhash); + } + } + # Log our failure: + + &Apache::lonnet::logthis('WARNING: '. + "Could not tie coursemap $fn for $uri."); + $tiedhash = ''; + $tiedparmhash = ''; + my $i=0; + + # Keep on retrying the lock for 90 sec until we succeed. + + while($i<90) { + $i++; + sleep(1); + if (flock(LOCKFILE,LOCK_EX|LOCK_NB)) { + + # Got the lock, tie the hashes...the assumption in this code is + # that some other worker thread has created the db files quite recently + # so no load is needed: + + $lock = 1; + if (tie(%hash,'GDBM_File',"$fn.db",&GDBM_READER(),0640)) { + $tiedhash = 1; + if (tie(%parmhash,'GDBM_File',$fn.'_parms.db',&GDBM_READER(),0640)) { + $tiedparmhash = 1; + if (-e "$fn.state") { + $retfurl='/adm/navmaps'; + + # BUG BUG: Side effect! + # Should conditionalize on something so that we can use this + # to load maps for courses that are not current? + # + &Apache::lonnet::appenv({"request.course.id" => $short, + "request.course.fn" => $fn, + "request.course.uri" => $uri, + "request.course.tied" => time}); + + $untiedhash = untie(%hash); + $untiedparmhash = untie(%parmhash); + $gotstate = 1; + last; + } + $untiedparmhash = untie(%parmhash); + } + $untiedhash = untie(%hash); + } + } + } + if ($lock) { + flock(LOCKFILE,LOCK_UN); + $lock = 0; + if ($tiedparmhash) { + unless ($untiedparmhash) { + &Apache::lonnet::logthis('WARNING: '. + 'Could not untie coursemap parmhash '.$fn.' for '.$uri.'.'); + } + } + if ($tiedparmhash) { + unless ($untiedhash) { + &Apache::lonnet::logthis('WARNING: '. + 'Could not untie coursemap hash '.$fn.' for '.$uri.'.'); + } + } + } + } + # I think this branch of code is all about what happens if we just did the stuff above, + # but found that the state file did not exist...again if we'd just held the lock + # would that have made this logic simpler..as generating all the files would be + # an atomic operation with respect to the lock. + # + unless ($gotstate) { + $lock = 0; + &Apache::lonnet::logthis('WARNING: '. + 'Could not read statemap '.$fn.' for '.$uri.'.'); + &unlink_tmpfiles($fn); + if (flock(LOCKFILE,LOCK_EX|LOCK_NB)) { + $lock=1; + } + undef %randompick; + undef %randompickseed; + undef %randomorder; + undef %randomizationcode; + undef %hiddenurl; + undef %encurl; + undef %rescount; + undef %mapcount; + $errtext=''; + $retfrid=''; + # + # Once more through the routine of tying and loading and so on. + # + if ($lock) { + if (tie(%hash,'GDBM_File',"$fn.db",&GDBM_WRCREAT(),0640)) { + if (tie(%parmhash,'GDBM_File',$fn.'_parms.db',&GDBM_WRCREAT(),0640)) { + $gotstate = &build_tmp_hashes($uri,$fn,$short,\%cenv); # TODO: User dependent? + unless ($gotstate) { + &Apache::lonnet::logthis('WARNING: '. + 'Failed to write statemap at second attempt '.$fn.' for '.$uri.'.'); + } + unless (untie(%parmhash)) { + &Apache::lonnet::logthis('WARNING: '. + 'Could not untie coursemap parmhash '.$fn.'.db for '.$uri.'.'); + } + } else { + &Apache::lonnet::logthis('WARNING: '. + 'Could not tie coursemap '.$fn.'__parms.db for '.$uri.'.'); + } + unless (untie(%hash)) { + &Apache::lonnet::logthis('WARNING: '. + 'Could not untie coursemap hash '.$fn.'.db for '.$uri.'.'); + } + } else { + &Apache::lonnet::logthis('WARNING: '. + 'Could not tie coursemap '.$fn.'.db for '.$uri.'.'); + } + flock(LOCKFILE,LOCK_UN); + $lock = 0; + } else { + # Failed to get the immediate lock. + + &Apache::lonnet::logthis('WARNING: '. + 'Could not obtain lock to tie coursemap hash '.$fn.'.db for '.$uri.'.'); + } + } + close(LOCKFILE); + unless (($errtext eq '') || ($env{'request.course.uri'} =~ m{^/uploaded/})) { + &Apache::lonmsg::author_res_msg($env{'request.course.uri'}, + $errtext); # TODO: User dependent? + } +# ------------------------------------------------- Check for critical messages + +# Depends on user must parameterize this as well..or separate as this is: +# more part of determining what someone sees on entering a course? + + my @what=&Apache::lonnet::dump('critical',$env{'user.domain'}, + $env{'user.name'}); + if ($what[0]) { + if (($what[0] ne 'con_lost') && ($what[0]!~/^error\:/)) { + $retfurl='/adm/email?critical=display'; + } + } + return ($retfurl,$errtext); +} + +# +# This sub is called when the course hash and the param hash have been tied and +# their lock file is held. +# Parameters: +# $uri - URI that identifies the course. +# $fn - The base path/filename of the files that make up the context +# being built. +# $short - Short course name. +# $cenvref - Reference to the course environment hash returned by +# Apache::lonnet::coursedescription +# +# Assumptions: +# The globals +# %hash, %paramhash are tied to their gdbm files and we hold the lock on them. +# +sub build_tmp_hashes { + my ($uri,$fn,$short,$cenvref) = @_; + + unless(ref($cenvref) eq 'HASH') { + return; + } + my %cenv = %{$cenvref}; + my $gotstate = 0; + %hash=(); # empty the global course and parameter hashes. + %parmhash=(); + $errtext=''; # No error messages yet. $pc=0; - loadmap($uri); + &clear_mapalias_count(); + &processversionfile(%cenv); + + # URI Of the map file. + + my $furi=&Apache::lonnet::clutter($uri); + # + # the map staring points. + # + $hash{'src_0.0'}=&versiontrack($furi); + $hash{'title_0.0'}=&Apache::lonnet::metadata($uri,'title'); + $hash{'ids_'.$furi}='0.0'; + $hash{'is_map_0.0'}=1; + + # Load the map.. note that loadmap may implicitly recurse if the map contains + # sub-maps. + + + &loadmap($uri,'0.0',$short); + + # The code below only executes if there is a starting point for the map> + # Q/BUG??? If there is no start resource for the map should that be an error? + # + if (defined($hash{'map_start_'.$uri})) { + &Apache::lonnet::appenv({"request.course.id" => $short, + "request.course.fn" => $fn, + "request.course.uri" => $uri, + "request.course.tied" => time}); + $env{'request.course.id'}=$short; &traceroute('0',$hash{'map_start_'.$uri},'&'); &accinit($uri,$short,$fn); + &hiddenurls(); + &mapcrumbs(); + } + $errtext .= &get_mapalias_errors(); +# ------------------------------------------------------- Put versions into src + foreach my $key (keys(%hash)) { + if ($key=~/^src_/) { + $hash{$key}=&putinversion($hash{$key}); + } elsif ($key =~ /^(map_(?:start|finish|pc)_)(.*)/) { + my ($type, $url) = ($1,$2); + my $value = $hash{$key}; + $hash{$type.&putinversion($url)}=$value; + } + } +# ---------------------------------------------------------------- Encrypt URLs + foreach my $id (keys(%encurl)) { +# $hash{'src_'.$id}=&Apache::lonenc::encrypted($hash{'src_'.$id}); + $hash{'encrypted_'.$id}=1; } - unless (untie(%hash)) { - &Apache::lonnet::logthis("WARNING: ". - "Could not untie coursemap $fn for $uri."); +# ----------------------------------------------- Close hashes to finally store +# --------------------------------- Routine must pass this point, no early outs + $hash{'first_rid'}=$retfrid; + my ($mapid,$resid)=split(/\./,$retfrid); + $hash{'first_mapurl'}=$hash{'map_id_'.$mapid}; + my $symb=&Apache::lonnet::encode_symb($hash{'map_id_'.$mapid},$resid,$hash{'src_'.$retfrid}); + $retfurl=&add_get_param($hash{'src_'.$retfrid},{ 'symb' => $symb }); + if ($hash{'encrypted_'.$retfrid}) { + $retfurl=&Apache::lonenc::encrypted($retfurl,(&Apache::lonnet::allowed('adv') ne 'F')); } + $hash{'first_url'}=$retfurl; +# ---------------------------------------------------- Store away initial state { - my $cfh; - if ($cfh=Apache::File->new(">$fn.state")) { - print $cfh join("\n",@cond); - } else { - &Apache::lonnet::logthis("WARNING: ". - "Could not write statemap $fn for $uri."); - } - } - } else { - &Apache::lonnet::logthis("WARNING: ". - "Could not tie coursemap $fn for $uri."); - } - return $errtext; + my $cfh; + if (open($cfh,">","$fn.state")) { + print $cfh join("\n",@cond); + $gotstate = 1; + } else { + &Apache::lonnet::logthis("WARNING: ". + "Could not write statemap $fn for $uri."); + } + } + return $gotstate; +} + +sub unlink_tmpfiles { + my ($fn) = @_; + my $file_dir = dirname($fn); + + if ("$file_dir/" eq LONCAPA::tempdir()) { + my @files = qw (.db _symb.db .state _parms.db); + foreach my $file (@files) { + if (-e $fn.$file) { + unless (unlink($fn.$file)) { + &Apache::lonnet::logthis("WARNING: ". + "Could not unlink ".$fn.$file.""); + } + } + } + } + return; +} + +# ------------------------------------------------------- Evaluate state string + +sub evalstate { + my $fn=$env{'request.course.fn'}.'.state'; + my $state=''; + if (-e $fn) { + my @conditions=(); + { + open(my $fh,"<",$fn); + @conditions=<$fh>; + close($fh); + } + my $safeeval = new Safe; + my $safehole = new Safe::Hole; + $safeeval->permit("entereval"); + $safeeval->permit(":base_math"); + $safeeval->deny(":base_io"); + $safehole->wrap(\&Apache::lonnet::EXT,$safeeval,'&EXT'); + foreach my $line (@conditions) { + chomp($line); + my ($condition,$weight)=split(/\:/,$line); + if ($safeeval->reval($condition)) { + if ($weight eq 'force') { + $state.='3'; + } else { + $state.='2'; + } + } else { + if ($weight eq 'stop') { + $state.='0'; + } else { + $state.='1'; + } + } + } + } + &Apache::lonnet::appenv({'user.state.'.$env{'request.course.id'} => $state}); + return $state; +} + +# This block seems to have code to manage/detect doubly defined +# aliases in maps. + +{ + my %mapalias_cache; + sub count_mapalias { + my ($value,$resid) = @_; + push(@{ $mapalias_cache{$value} }, $resid); + } + + sub get_mapalias_errors { + my $error_text; + foreach my $mapalias (sort(keys(%mapalias_cache))) { + next if (scalar(@{ $mapalias_cache{$mapalias} } ) == 1); + my $count; + my $which = + join('
  • ', + map { + my $id = $_; + if (exists($hash{'src_'.$id})) { + $count++; + } + my ($mapid) = split(/\./,$id); + &mt('Resource [_1][_2]in Map [_3]', + $hash{'title_'.$id},'
    ', + $hash{'title_'.$hash{'ids_'.$hash{'map_id_'.$mapid}}}); + } (@{ $mapalias_cache{$mapalias} })); + next if ($count < 2); + $error_text .= '
    '. + &mt('Error: Found the mapalias "[_1]" defined multiple times.', + $mapalias). + '
    • '.$which.'
    '; + } + &clear_mapalias_count(); + return $error_text; + } + sub clear_mapalias_count { + undef(%mapalias_cache); + } } - 1; __END__ +=head1 NAME + +Apache::lonuserstate - Construct and maintain state and binary representation +of course for user + +=head1 SYNOPSIS + +Invoked by lonroles.pm. + +&Apache::lonuserstate::readmap($cdom.'/'.$cnum); + +=head1 INTRODUCTION + +This module constructs and maintains state and binary representation +of course for user. + +This is part of the LearningOnline Network with CAPA project +described at http://www.lon-capa.org. + +=head1 SUBROUTINES + +=over + +=item loadmap() + +Loads map from disk + +=item simplify() + +Simplify expression + +=item traceroute() + +Build condition hash + +=item accinit() + +Cascading conditions, quick access, parameters +=item readmap() +Read map and all submaps +=item evalstate() +Evaluate state string +=back +=cut