line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
|
2
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
package Parse::Gnaw::Blocks::ParsingMethods; |
4
|
|
|
|
|
|
|
|
5
|
|
|
|
|
|
|
our $VERSION = '0.001'; |
6
|
|
|
|
|
|
|
|
7
|
|
|
|
|
|
|
#BEGIN {print "Parse::Gnaw::Blocks::ParsingMethods\n";} |
8
|
|
|
|
|
|
|
|
9
|
19
|
|
|
19
|
|
128
|
use warnings; |
|
19
|
|
|
|
|
32
|
|
|
19
|
|
|
|
|
575
|
|
10
|
19
|
|
|
19
|
|
96
|
use strict; |
|
19
|
|
|
|
|
35
|
|
|
19
|
|
|
|
|
533
|
|
11
|
19
|
|
|
19
|
|
96
|
use Carp; |
|
19
|
|
|
|
|
35
|
|
|
19
|
|
|
|
|
1316
|
|
12
|
19
|
|
|
19
|
|
103
|
use Data::Dumper; |
|
19
|
|
|
|
|
41
|
|
|
19
|
|
|
|
|
906
|
|
13
|
19
|
|
|
19
|
|
106
|
use Storable 'dclone'; |
|
19
|
|
|
|
|
32
|
|
|
19
|
|
|
|
|
918
|
|
14
|
|
|
|
|
|
|
|
15
|
19
|
|
|
19
|
|
107
|
use Parse::Gnaw::Blocks::LetterConstants; |
|
19
|
|
|
|
|
53
|
|
|
19
|
|
|
|
|
1828
|
|
16
|
19
|
|
|
19
|
|
104
|
use Parse::Gnaw::LinkedListConstants; |
|
19
|
|
|
|
|
38
|
|
|
19
|
|
|
|
|
42835
|
|
17
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
=head1 NAME |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
Parse::Gnaw::Blocks::ParsingMethods - A base package containing all the methods that derived "letter" type classes will inherit. |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
|
25
|
|
|
|
|
|
|
=head2 parse_grammarref |
26
|
|
|
|
|
|
|
|
27
|
|
|
|
|
|
|
We get a starting letter, and a reference to a rule. |
28
|
|
|
|
|
|
|
Because rules can be called like subroutines, we have to process the rule in such a way |
29
|
|
|
|
|
|
|
that it can be called from anywhere and we don't know who is calling us (or where we will return to). |
30
|
|
|
|
|
|
|
|
31
|
|
|
|
|
|
|
Note that we need to keep track of whether a rule consumes the current letter or not. |
32
|
|
|
|
|
|
|
some rules might not consume anything (rules that execute callbacks, for example, or configure flags) |
33
|
|
|
|
|
|
|
So, we need to call rules until we consume at least the current letter. |
34
|
|
|
|
|
|
|
|
35
|
|
|
|
|
|
|
if that is the last subrule, then return current letter. |
36
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
if there are more subrules, we need to get a list of possible connections to go to next. |
38
|
|
|
|
|
|
|
then loop through each possible connection and try the subrule. |
39
|
|
|
|
|
|
|
If it succeeds, great. |
40
|
|
|
|
|
|
|
If it fails, eval/trap the failure and loop to next connection. |
41
|
|
|
|
|
|
|
|
42
|
|
|
|
|
|
|
============================ |
43
|
|
|
|
|
|
|
This is a recursive call. |
44
|
|
|
|
|
|
|
============================ |
45
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
|
47
|
|
|
|
|
|
|
grammar is ('a', 'b', 'c') |
48
|
|
|
|
|
|
|
|
49
|
|
|
|
|
|
|
text is |
50
|
|
|
|
|
|
|
c b b |
51
|
|
|
|
|
|
|
j a k |
52
|
|
|
|
|
|
|
m n o |
53
|
|
|
|
|
|
|
|
54
|
|
|
|
|
|
|
start at "g", look for 'a'. fail, move to next starting position. repeat until hit 'a' in center position. |
55
|
|
|
|
|
|
|
match 'a' at center. look a next rule, it's defined and true, so we need to look for it. |
56
|
|
|
|
|
|
|
Howerver, we need to try every possible direction from "A". |
57
|
|
|
|
|
|
|
'b', 'b', 'k', 'o', 'n', 'm', 'j', 'c' |
58
|
|
|
|
|
|
|
|
59
|
|
|
|
|
|
|
This means every connection needs to trap for die "GRAMMARFAIL" |
60
|
|
|
|
|
|
|
if we try direction 'h', and it dies, |
61
|
|
|
|
|
|
|
we need to trap the error and try the next option. |
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
|
64
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
=cut |
66
|
|
|
|
|
|
|
sub parse_grammarref{ |
67
|
35
|
|
|
35
|
1
|
74
|
my ($llobj, $grammarref, $then_call )=@_; |
68
|
|
|
|
|
|
|
|
69
|
35
|
|
|
|
|
58
|
my $debug=0; |
70
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
# first element in grammarref array has information about this rule. get it. |
72
|
35
|
|
|
|
|
65
|
my $first_subrule = $grammarref->[0]; |
73
|
|
|
|
|
|
|
|
74
|
35
|
|
|
|
|
56
|
my $ruleinfo = $first_subrule->[2]; |
75
|
35
|
|
|
|
|
78
|
my $rulename = $ruleinfo->{payload}; |
76
|
35
|
50
|
|
|
|
144
|
my $quantifiertype= exists($ruleinfo->{quantifier}) ? $ruleinfo->{quantifier} : ''; |
77
|
35
|
100
|
|
|
|
93
|
my $isquantifier = ($quantifiertype eq '') ? '' : 1; |
78
|
|
|
|
|
|
|
|
79
|
35
|
|
|
|
|
313
|
my ($min,$max)=(1,1); # if not a quantifier, then want to match this rule exactly once |
80
|
|
|
|
|
|
|
|
81
|
35
|
100
|
|
|
|
111
|
if($isquantifier){ |
82
|
6
|
|
|
|
|
13
|
$min=$ruleinfo->{min}; |
83
|
6
|
|
|
|
|
14
|
$max=$ruleinfo->{max}; |
84
|
|
|
|
|
|
|
} |
85
|
|
|
|
|
|
|
|
86
|
35
|
|
|
|
|
70
|
my $start_letter = $llobj->[LIST__CURR_START]; |
87
|
35
|
|
|
|
|
67
|
my $entry_letter_payload = $start_letter->[LETTER__DATA_PAYLOAD]; |
88
|
35
|
50
|
|
|
|
94
|
my $minstr=defined($min) ? $min : 'undef'; |
89
|
35
|
50
|
|
|
|
83
|
my $maxstr=defined($max) ? $max : 'undef'; |
90
|
|
|
|
|
|
|
|
91
|
35
|
|
|
|
|
231
|
my $parse_grammarref_identification_string = "parse_grammarref, called with rule='$rulename', ". |
92
|
|
|
|
|
|
|
"letter='$entry_letter_payload', quantifiertype='$quantifiertype', isquantifier='$isquantifier', ". |
93
|
|
|
|
|
|
|
"min='$minstr', max='$maxstr', then_call='$then_call'"; |
94
|
35
|
50
|
|
|
|
124
|
if($debug){warn "$parse_grammarref_identification_string: begin";} |
|
0
|
|
|
|
|
0
|
|
95
|
|
|
|
|
|
|
|
96
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
# We're going to pretend that every rule is like a quantifier |
98
|
|
|
|
|
|
|
# if it's not REALLY a quantifier, then min/max is 1/1 and we must match exactly once. |
99
|
|
|
|
|
|
|
# if it IS really a quantifier, then we need to match min times |
100
|
|
|
|
|
|
|
|
101
|
|
|
|
|
|
|
# first, match "min" number of times. |
102
|
|
|
|
|
|
|
# if we fail that,then this rule fails, therefore, no need to trap error |
103
|
|
|
|
|
|
|
|
104
|
|
|
|
|
|
|
# if the minimum is 3, then we need to call the rule 3 times. |
105
|
35
|
|
|
|
|
110
|
foreach my $minimal_match (1 .. $min){ |
106
|
|
|
|
|
|
|
|
107
|
38
|
50
|
|
|
|
102
|
if($debug){warn "parse_grammarref($rulename): about to try one_iteration_of_grammarref";} |
|
0
|
|
|
|
|
0
|
|
108
|
|
|
|
|
|
|
|
109
|
38
|
|
|
|
|
807
|
my $discard_retval = $llobj->one_iteration_of_grammarref($grammarref,1); |
110
|
|
|
|
|
|
|
|
111
|
27
|
50
|
|
|
|
115
|
if($debug){warn "parse_grammarref($rulename): just tried one_iteration_of_grammarref and succeeded";} |
|
0
|
|
|
|
|
0
|
|
112
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
|
114
|
|
|
|
|
|
|
} |
115
|
|
|
|
|
|
|
|
116
|
|
|
|
|
|
|
# matched MIN number of times |
117
|
|
|
|
|
|
|
# now try 0 to ($max-$min) more times. |
118
|
|
|
|
|
|
|
# if min=3 and max=7, then we would try 0 to 4 more times. |
119
|
|
|
|
|
|
|
|
120
|
24
|
|
|
|
|
54
|
my $remainingattempts = -1; |
121
|
24
|
100
|
|
|
|
70
|
if($max>0){ |
122
|
18
|
|
|
|
|
39
|
$remainingattempts=($max-$min); |
123
|
|
|
|
|
|
|
} |
124
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
|
126
|
24
|
50
|
|
|
|
62
|
if($debug){warn "remainingattempts='$remainingattempts'";} |
|
0
|
|
|
|
|
0
|
|
127
|
|
|
|
|
|
|
|
128
|
24
|
|
100
|
|
|
150
|
for(my $quant_counter=0; (($remainingattempts<0) or ($quant_counter<=$remainingattempts)); $quant_counter++){ |
129
|
|
|
|
|
|
|
|
130
|
32
|
50
|
|
|
|
122
|
if($debug){warn "quant_counter='$quant_counter'";} |
|
0
|
|
|
|
|
0
|
|
131
|
|
|
|
|
|
|
|
132
|
|
|
|
|
|
|
################################################################################## |
133
|
|
|
|
|
|
|
# try one iteration of grammarref |
134
|
|
|
|
|
|
|
################################################################################## |
135
|
32
|
100
|
|
|
|
91
|
if($quant_counter>0){ |
136
|
8
|
50
|
|
|
|
25
|
if($debug){warn "parse_grammarref($rulename): try anotehr iteration of grammarref";} |
|
0
|
|
|
|
|
0
|
|
137
|
8
|
|
|
|
|
23
|
my $discard_retval = $llobj->one_iteration_of_grammarref($grammarref,1); |
138
|
|
|
|
|
|
|
} |
139
|
|
|
|
|
|
|
|
140
|
|
|
|
|
|
|
|
141
|
|
|
|
|
|
|
################################################################################## |
142
|
|
|
|
|
|
|
# then try the then_call rule. |
143
|
|
|
|
|
|
|
################################################################################## |
144
|
32
|
100
|
66
|
|
|
223
|
if( not defined($then_call) or ($then_call eq '') ){ |
145
|
|
|
|
|
|
|
# there is no then_call, we're done and we must have matched. huzzah! return |
146
|
13
|
50
|
|
|
|
45
|
if($debug){warn "parse_grammarref($rulename): no then_call, return success"; } |
|
0
|
|
|
|
|
0
|
|
147
|
13
|
|
|
|
|
69
|
return; |
148
|
|
|
|
|
|
|
} else { |
149
|
19
|
50
|
|
|
|
46
|
if($debug){warn "parse_grammarref($rulename): try the then_call rule";} |
|
0
|
|
|
|
|
0
|
|
150
|
|
|
|
|
|
|
# there is a then_call, try it, if it fails, catch and try another quant_counter. |
151
|
|
|
|
|
|
|
|
152
|
19
|
|
|
|
|
31
|
my $save_position=$llobj->[LIST__CURR_START]; # if all fail, go back to here |
153
|
|
|
|
|
|
|
|
154
|
|
|
|
|
|
|
#warn "then_call is '$then_call'"; |
155
|
|
|
|
|
|
|
|
156
|
19
|
|
|
|
|
332
|
eval{ |
157
|
19
|
|
|
|
|
67
|
my $thencall_grammarref=$llobj->convert_rule_name_to_rule_reference($then_call); |
158
|
19
|
|
|
|
|
80
|
my $discard_retval = $llobj->one_iteration_of_grammarref($thencall_grammarref,1); |
159
|
|
|
|
|
|
|
}; |
160
|
19
|
100
|
|
|
|
60
|
if($@){ |
161
|
9
|
50
|
|
|
|
48
|
if($@ =~ m{GRAMMARFAIL}){ |
162
|
9
|
|
|
|
|
49
|
$llobj->[LIST__CURR_START]=$save_position; |
163
|
|
|
|
|
|
|
} else { |
164
|
0
|
|
|
|
|
0
|
die $@; |
165
|
|
|
|
|
|
|
} |
166
|
|
|
|
|
|
|
} else { |
167
|
|
|
|
|
|
|
# didn't die. must have matched. huzzah! |
168
|
10
|
50
|
|
|
|
32
|
if($debug){warn "parse_grammarref($rulename): returning";} |
|
0
|
|
|
|
|
0
|
|
169
|
10
|
|
|
|
|
52
|
return; |
170
|
|
|
|
|
|
|
} |
171
|
|
|
|
|
|
|
} |
172
|
|
|
|
|
|
|
} |
173
|
|
|
|
|
|
|
|
174
|
|
|
|
|
|
|
# note that if the above for(my $quant_counter=0; loop |
175
|
|
|
|
|
|
|
# ever finds a match of the quantifier plus successfully runs then_call |
176
|
|
|
|
|
|
|
# then it will return in the "else" part of "if($@)". |
177
|
1
|
50
|
|
|
|
4
|
if($debug){warn "parse_grammarref($rulename): tried looping on quant_counter but failed to match";} |
|
0
|
|
|
|
|
0
|
|
178
|
1
|
|
|
|
|
13
|
die "GRAMMARFAIL"; |
179
|
|
|
|
|
|
|
} |
180
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
=head2 one_iteration_of_grammarref |
182
|
|
|
|
|
|
|
|
183
|
|
|
|
|
|
|
This runs an entire rule exactly one time. |
184
|
|
|
|
|
|
|
|
185
|
|
|
|
|
|
|
It does not call then_call. |
186
|
|
|
|
|
|
|
|
187
|
|
|
|
|
|
|
It does not address quantifier issues. |
188
|
|
|
|
|
|
|
|
189
|
|
|
|
|
|
|
|
190
|
|
|
|
|
|
|
|
191
|
|
|
|
|
|
|
=cut |
192
|
|
|
|
|
|
|
|
193
|
|
|
|
|
|
|
my $counter=0; |
194
|
|
|
|
|
|
|
|
195
|
|
|
|
|
|
|
# pass in subrule_iterator, return subrule_iterator |
196
|
|
|
|
|
|
|
sub one_iteration_of_grammarref{ |
197
|
107
|
|
|
107
|
1
|
174
|
my ($llobj, $grammarref,$subrule_iterator)=@_; # no "$thencall" here. |
198
|
|
|
|
|
|
|
|
199
|
107
|
|
|
|
|
138
|
my $debug=0; |
200
|
|
|
|
|
|
|
|
201
|
107
|
|
|
|
|
147
|
my $size_rule=scalar(@$grammarref); |
202
|
|
|
|
|
|
|
|
203
|
107
|
|
|
|
|
129
|
my $number_of_possible_connections=1; |
204
|
|
|
|
|
|
|
|
205
|
107
|
|
|
|
|
151
|
my $initial_subrule=$grammarref->[$subrule_iterator]; |
206
|
107
|
|
|
|
|
243
|
my $sub_method=$initial_subrule->[0]; |
207
|
107
|
|
|
|
|
234
|
my $sub_payload=$initial_subrule->[1]; |
208
|
107
|
|
|
|
|
236
|
my $initial_state_of_call = "called one_iteration_of_grammarref with sub_method='$sub_method' and sub_payload='$sub_payload'"; |
209
|
107
|
50
|
|
|
|
242
|
if($debug){warn $initial_state_of_call;} |
|
0
|
|
|
|
|
0
|
|
210
|
107
|
|
|
|
|
795
|
my @caller=caller(0); |
211
|
|
|
|
|
|
|
#print Dumper \@caller; |
212
|
|
|
|
|
|
|
#warn "called from (see above)"; |
213
|
|
|
|
|
|
|
|
214
|
107
|
|
33
|
|
|
574
|
while(($number_of_possible_connections==1) and ($subrule_iterator<$size_rule)){ |
215
|
|
|
|
|
|
|
|
216
|
161
|
50
|
|
|
|
330
|
if($debug){warn "WHILE: $initial_state_of_call subrule_iterator=$subrule_iterator";} |
|
0
|
|
|
|
|
0
|
|
217
|
|
|
|
|
|
|
############################################################################################# |
218
|
|
|
|
|
|
|
############################################################################################# |
219
|
|
|
|
|
|
|
############################################################################################# |
220
|
|
|
|
|
|
|
# if currentletter is not consumed yet |
221
|
|
|
|
|
|
|
# go through the grammar rules until its consumed, then stop this part of loop. |
222
|
|
|
|
|
|
|
############################################################################################# |
223
|
|
|
|
|
|
|
############################################################################################# |
224
|
|
|
|
|
|
|
############################################################################################# |
225
|
161
|
|
66
|
|
|
912
|
while( ($llobj->[LIST__CURR_START]->[LETTER__LETTER_HAS_BEEN_CONSUMED]==0) and ($subrule_iterator<$size_rule) ){ |
226
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
# get the subrule, i.e. ['lit','a',...] |
228
|
131
|
|
|
|
|
187
|
my $subrule = $grammarref->[$subrule_iterator]; |
229
|
131
|
|
|
|
|
148
|
$subrule_iterator++; |
230
|
|
|
|
|
|
|
|
231
|
|
|
|
|
|
|
# get the subrule name, i.e. 'lit' |
232
|
131
|
|
|
|
|
184
|
my $methodname=$subrule->[0]; |
233
|
|
|
|
|
|
|
|
234
|
|
|
|
|
|
|
#my $payloadify = $subrule->[1]; warn "trying methodname '$methodname' with payload '$payloadify'"; |
235
|
|
|
|
|
|
|
|
236
|
|
|
|
|
|
|
# if we can't call the methodname, then die a miserable death. |
237
|
131
|
50
|
|
|
|
647
|
if(not($llobj->can($methodname))){ |
238
|
0
|
|
|
|
|
0
|
print "no method found for '$methodname'\n"; warn; |
|
0
|
|
|
|
|
0
|
|
239
|
0
|
|
|
|
|
0
|
my $hashref=$subrule->[2]; |
240
|
0
|
|
|
|
|
0
|
print Dumper $hashref; |
241
|
0
|
|
|
|
|
0
|
my $filename=$hashref->{filename}; |
242
|
0
|
|
|
|
|
0
|
my $linenum =$hashref->{linenum}; |
243
|
0
|
|
|
|
|
0
|
confess "grammar method not defined '$methodname' in $filename line $linenum"; |
244
|
|
|
|
|
|
|
} |
245
|
|
|
|
|
|
|
|
246
|
131
|
50
|
|
|
|
299
|
if($debug){warn "methodname is '$methodname', subrule is"; print Dumper $subrule; } |
|
0
|
|
|
|
|
0
|
|
|
0
|
|
|
|
|
0
|
|
247
|
|
|
|
|
|
|
|
248
|
|
|
|
|
|
|
# call the method, i.e. $letter->lit(['lit','a',...]); |
249
|
131
|
|
|
|
|
339
|
$llobj->$methodname($subrule); |
250
|
|
|
|
|
|
|
|
251
|
81
|
50
|
|
|
|
467
|
unless(defined($llobj->[LIST__CURR_START])){ |
252
|
0
|
|
|
|
|
0
|
die "ERROR: called method and got undefined letter back (method == $methodname)"; |
253
|
|
|
|
|
|
|
} |
254
|
|
|
|
|
|
|
} |
255
|
|
|
|
|
|
|
|
256
|
|
|
|
|
|
|
# we consumed the current letter, so if there ARE MORE RULES in this grammar, |
257
|
|
|
|
|
|
|
# we need to figure out which connection to use to get to move to the next letter. |
258
|
|
|
|
|
|
|
# If there are NO MORE RULES left, just return letter (marked as consumed) |
259
|
|
|
|
|
|
|
# because we can't iterate the connections from that letter without knowing |
260
|
|
|
|
|
|
|
# what the next rule is. The next rule to get called will see letter is consumed |
261
|
|
|
|
|
|
|
# and skip to the possible connections section. |
262
|
111
|
100
|
|
|
|
261
|
if($subrule_iterator>=$size_rule){ |
263
|
45
|
|
|
|
|
210
|
return $subrule_iterator; |
264
|
|
|
|
|
|
|
} |
265
|
|
|
|
|
|
|
|
266
|
|
|
|
|
|
|
|
267
|
|
|
|
|
|
|
############################################################################################# |
268
|
|
|
|
|
|
|
############################################################################################# |
269
|
|
|
|
|
|
|
############################################################################################# |
270
|
|
|
|
|
|
|
# current letter is consumed and there are more subrules in this rule. |
271
|
|
|
|
|
|
|
# get a list of all possible letters connected to the current letter |
272
|
|
|
|
|
|
|
# for each possible connection, try setting current letter to that connected letter |
273
|
|
|
|
|
|
|
# and see if rest of rule matches. If doesn't match, trap failure, and try next possible letter. |
274
|
|
|
|
|
|
|
############################################################################################# |
275
|
|
|
|
|
|
|
############################################################################################# |
276
|
|
|
|
|
|
|
############################################################################################# |
277
|
|
|
|
|
|
|
|
278
|
|
|
|
|
|
|
# get a list of letter objects to try. |
279
|
66
|
|
|
|
|
252
|
my @list_of_possible_next_letters = $llobj->[LIST__CURR_START]->get_list_of_connecting_letters(); |
280
|
|
|
|
|
|
|
|
281
|
66
|
|
|
|
|
108
|
$number_of_possible_connections=scalar(@list_of_possible_next_letters); |
282
|
66
|
50
|
|
|
|
159
|
if($number_of_possible_connections==0) { |
283
|
0
|
|
|
|
|
0
|
die "somehow we mannaged to get into a letter that has no connections?"; |
284
|
|
|
|
|
|
|
} |
285
|
|
|
|
|
|
|
|
286
|
|
|
|
|
|
|
# if there is only 1 possible connection, then we can avoid recursion here. |
287
|
|
|
|
|
|
|
# just move current letter to the next possible letter, and loop around. |
288
|
|
|
|
|
|
|
# when we're parsing a simple string, this should save us time and memory. |
289
|
66
|
100
|
|
|
|
2249
|
if($number_of_possible_connections==1){ |
290
|
54
|
50
|
|
|
|
129
|
if($debug){warn "only 1 connection";} |
|
0
|
|
|
|
|
0
|
|
291
|
54
|
50
|
|
|
|
146
|
die "stuck in a loop" if($counter++>4000); |
292
|
|
|
|
|
|
|
|
293
|
54
|
|
|
|
|
114
|
$llobj->[LIST__CURR_START] = shift(@list_of_possible_next_letters); |
294
|
54
|
|
|
|
|
318
|
$llobj->[LIST__CURR_START] ->[LETTER__LETTER_HAS_BEEN_CONSUMED]=0; |
295
|
|
|
|
|
|
|
} else { |
296
|
12
|
50
|
|
|
|
22
|
if($debug){warn "multiple connections";} |
|
0
|
|
|
|
|
0
|
|
297
|
|
|
|
|
|
|
# else there are multiple possible connections. |
298
|
|
|
|
|
|
|
# will have to try each one in sequence until we get a match. |
299
|
|
|
|
|
|
|
# will have to be recursive because next letter will also have a bunch of connections |
300
|
|
|
|
|
|
|
# and we will have to loop through and try each possible connection for that letter. |
301
|
|
|
|
|
|
|
|
302
|
12
|
|
|
|
|
15
|
my $save_position=$llobj->[LIST__CURR_START]; # if all fail, go back to here |
303
|
12
|
|
|
|
|
14
|
my $save_iterator=$subrule_iterator; |
304
|
|
|
|
|
|
|
|
305
|
12
|
|
|
|
|
20
|
TRYCONNECTION : foreach my $possible_letter (@list_of_possible_next_letters){ |
306
|
|
|
|
|
|
|
|
307
|
42
|
|
|
|
|
50
|
$llobj->[LIST__CURR_START] = $possible_letter; |
308
|
|
|
|
|
|
|
|
309
|
|
|
|
|
|
|
# mark letter as not consumed. |
310
|
42
|
|
|
|
|
51
|
$llobj->[LIST__CURR_START]->[LETTER__LETTER_HAS_BEEN_CONSUMED]=0; |
311
|
|
|
|
|
|
|
|
312
|
42
|
|
|
|
|
50
|
eval{ |
313
|
42
|
|
|
|
|
101
|
$subrule_iterator = $llobj->one_iteration_of_grammarref($grammarref,$subrule_iterator); |
314
|
|
|
|
|
|
|
}; |
315
|
42
|
100
|
|
|
|
81
|
if($@){ |
316
|
36
|
50
|
|
|
|
81
|
if($@ =~ m{GRAMMARFAIL}){ |
317
|
|
|
|
|
|
|
# grammar failed, so this connection didn't work. Try another connection. |
318
|
36
|
|
|
|
|
40
|
$llobj->[LIST__CURR_START] =$save_position; |
319
|
36
|
|
|
|
|
45
|
$subrule_iterator =$save_iterator; |
320
|
36
|
|
|
|
|
69
|
next TRYCONNECTION; |
321
|
|
|
|
|
|
|
} else { |
322
|
|
|
|
|
|
|
# else we died, and it wasn't a grammar failure. rethrow the die |
323
|
0
|
|
|
|
|
0
|
die $@; |
324
|
|
|
|
|
|
|
} |
325
|
|
|
|
|
|
|
} else { |
326
|
|
|
|
|
|
|
###### |
327
|
|
|
|
|
|
|
# we eval'ed and didn't get $@. must have matched. Hazzah! |
328
|
|
|
|
|
|
|
###### |
329
|
|
|
|
|
|
|
|
330
|
|
|
|
|
|
|
# at the end of recursive calls, the iterator should equal the size of the rule. |
331
|
|
|
|
|
|
|
# (i.e. if rule has 3 elements in it, iterator will go 0,1,2 to try each subrule |
332
|
|
|
|
|
|
|
# and then it will increment to 3 and should return as matcihng the entire rule) |
333
|
|
|
|
|
|
|
# if iterator doesn't equal size of rule, then something went wrong in the process of parsing. |
334
|
6
|
50
|
|
|
|
12
|
unless($subrule_iterator==$size_rule){ |
335
|
0
|
|
|
|
|
0
|
die "ERROR: somehow managed to return a rule without matching all of the rule."; |
336
|
|
|
|
|
|
|
} |
337
|
|
|
|
|
|
|
|
338
|
|
|
|
|
|
|
# OK, we didn't get $@, AND it looks like the recursive calls matched the entire rule. |
339
|
|
|
|
|
|
|
# return as a successful match. |
340
|
6
|
|
|
|
|
26
|
return $subrule_iterator; |
341
|
|
|
|
|
|
|
} |
342
|
|
|
|
|
|
|
} |
343
|
6
|
|
|
|
|
40
|
die "GRAMMARFAIL"; #tried all possible connections. none worked. Die. |
344
|
|
|
|
|
|
|
} |
345
|
|
|
|
|
|
|
|
346
|
|
|
|
|
|
|
# if there is only 1 possible connection, the if($number_of_possible_connections==1) statement will kick out here |
347
|
|
|
|
|
|
|
# we will use the enclosing while() loop to loop around and try the next letter. |
348
|
|
|
|
|
|
|
} |
349
|
|
|
|
|
|
|
|
350
|
|
|
|
|
|
|
# end of while(($number_of_possible_connections==1) and ($subrule_iterator<$size_rule)){ |
351
|
|
|
|
|
|
|
# Should only reach this point when we're parsing a one-dimensional string (connectins==1) |
352
|
|
|
|
|
|
|
# and the entire rule matched (iterator==size of rule) |
353
|
|
|
|
|
|
|
# return as a successful match. |
354
|
0
|
|
|
|
|
0
|
return $subrule_iterator; |
355
|
|
|
|
|
|
|
} |
356
|
|
|
|
|
|
|
|
357
|
|
|
|
|
|
|
|
358
|
|
|
|
|
|
|
|
359
|
|
|
|
|
|
|
=head2 rule |
360
|
|
|
|
|
|
|
|
361
|
|
|
|
|
|
|
The "rule" method is just a placeholder for the first index into each rule array. |
362
|
|
|
|
|
|
|
This is where we store the name of the rule and any othe rule-specific info. |
363
|
|
|
|
|
|
|
For now, it doesn't do anything. |
364
|
|
|
|
|
|
|
|
365
|
|
|
|
|
|
|
=cut |
366
|
|
|
|
|
|
|
|
367
|
|
|
|
|
|
|
sub rule { |
368
|
0
|
|
|
0
|
1
|
0
|
my ($llobj, $subrule)=@_; |
369
|
0
|
|
|
|
|
0
|
return; |
370
|
|
|
|
|
|
|
} |
371
|
|
|
|
|
|
|
|
372
|
|
|
|
|
|
|
|
373
|
|
|
|
|
|
|
=head2 call |
374
|
|
|
|
|
|
|
|
375
|
|
|
|
|
|
|
When one grammar rule needs to call another rule (including itself), |
376
|
|
|
|
|
|
|
this method will get executed. |
377
|
|
|
|
|
|
|
|
378
|
|
|
|
|
|
|
Note that this supports recursive calling. A rule can call itself. |
379
|
|
|
|
|
|
|
A first rule can call a second rule which calls the first rule. |
380
|
|
|
|
|
|
|
|
381
|
|
|
|
|
|
|
The main reason this works is that when a rule "calls" another rule, |
382
|
|
|
|
|
|
|
it doesn't actually CONTAIN the rule. A rule is actually made up of |
383
|
|
|
|
|
|
|
a perl array. |
384
|
|
|
|
|
|
|
|
385
|
|
|
|
|
|
|
my $rule1=[ |
386
|
|
|
|
|
|
|
[ 'lit', 'a' ], |
387
|
|
|
|
|
|
|
[ 'lit', 'b' ], |
388
|
|
|
|
|
|
|
]; |
389
|
|
|
|
|
|
|
|
390
|
|
|
|
|
|
|
If a rule "calls" itself, it simply points to the name of the rule its calling: |
391
|
|
|
|
|
|
|
|
392
|
|
|
|
|
|
|
my $rule1=[ |
393
|
|
|
|
|
|
|
[ 'lit', 'a' ], |
394
|
|
|
|
|
|
|
[ 'call', 'rule1' ], |
395
|
|
|
|
|
|
|
[ 'lit', 'b' ], |
396
|
|
|
|
|
|
|
]; |
397
|
|
|
|
|
|
|
|
398
|
|
|
|
|
|
|
If a call to a rule resulted in the rule being called being expanded and |
399
|
|
|
|
|
|
|
embedded into the original rule, then recursive rules would explode. |
400
|
|
|
|
|
|
|
|
401
|
|
|
|
|
|
|
|
402
|
|
|
|
|
|
|
my $rule1; |
403
|
|
|
|
|
|
|
|
404
|
|
|
|
|
|
|
$rule1=[ |
405
|
|
|
|
|
|
|
[ 'lit', 'a' ], |
406
|
|
|
|
|
|
|
[ 'call', $rule1 ], |
407
|
|
|
|
|
|
|
[ 'lit', 'b' ], |
408
|
|
|
|
|
|
|
]; |
409
|
|
|
|
|
|
|
|
410
|
|
|
|
|
|
|
This would become problematic because it would want to expand itself forever |
411
|
|
|
|
|
|
|
(which would be grammatically correct, but explode your memory) or it would |
412
|
|
|
|
|
|
|
only expand one level (which would fit in memory, and be grammatically incorrect). |
413
|
|
|
|
|
|
|
|
414
|
|
|
|
|
|
|
If a call to another rule only contains the NAME of the rule being called, |
415
|
|
|
|
|
|
|
then it won't explode memory. |
416
|
|
|
|
|
|
|
|
417
|
|
|
|
|
|
|
A rule will call a rule only when it needs to, and not explode memory. |
418
|
|
|
|
|
|
|
|
419
|
|
|
|
|
|
|
So then the only other issue that can cause a recursive rule ot explode |
420
|
|
|
|
|
|
|
is if a rule calls itself before matching any text in the source string. |
421
|
|
|
|
|
|
|
|
422
|
|
|
|
|
|
|
This rule will explode when we try to match it against some text: |
423
|
|
|
|
|
|
|
|
424
|
|
|
|
|
|
|
$rule1=[ |
425
|
|
|
|
|
|
|
[ 'call', 'rule1' ], |
426
|
|
|
|
|
|
|
[ 'lit', 'b' ], |
427
|
|
|
|
|
|
|
]; |
428
|
|
|
|
|
|
|
|
429
|
|
|
|
|
|
|
The above example will break because rule1 will keep calling itself infinitely |
430
|
|
|
|
|
|
|
without ever matching anything. |
431
|
|
|
|
|
|
|
|
432
|
|
|
|
|
|
|
For the above example NOT to crash, we will eventually have to upgrade the |
433
|
|
|
|
|
|
|
"call" method to detect whether a recursive call is taking place, and |
434
|
|
|
|
|
|
|
if so, check to see that at least SOME text has been consumed. If not, |
435
|
|
|
|
|
|
|
skip the call and look for an alternation or something. |
436
|
|
|
|
|
|
|
|
437
|
|
|
|
|
|
|
For now, we can handle recurssion, but only if we match some text first: |
438
|
|
|
|
|
|
|
|
439
|
|
|
|
|
|
|
$rule1=[ |
440
|
|
|
|
|
|
|
['lit','a'], |
441
|
|
|
|
|
|
|
['call', 'rule1'], |
442
|
|
|
|
|
|
|
|
443
|
|
|
|
|
|
|
]; |
444
|
|
|
|
|
|
|
|
445
|
|
|
|
|
|
|
=cut |
446
|
|
|
|
|
|
|
|
447
|
|
|
|
|
|
|
sub call{ |
448
|
13
|
|
|
13
|
1
|
28
|
my ($llobj, $subrule)=@_; |
449
|
|
|
|
|
|
|
|
450
|
13
|
|
|
|
|
20
|
my $debug=0; |
451
|
|
|
|
|
|
|
|
452
|
13
|
50
|
|
|
|
43
|
if($debug){warn "call subrule is ";print Dumper $subrule;} |
|
0
|
|
|
|
|
0
|
|
|
0
|
|
|
|
|
0
|
|
453
|
|
|
|
|
|
|
|
454
|
|
|
|
|
|
|
#print "INSIDE CALL\n"; |
455
|
|
|
|
|
|
|
#print "letter IS\n"; print Dumper $letter; |
456
|
|
|
|
|
|
|
#print "SUBRULE IS\n"; print Dumper $subrule; |
457
|
13
|
|
|
|
|
20
|
my $hash_info=$subrule->[2]; |
458
|
|
|
|
|
|
|
|
459
|
13
|
|
|
|
|
33
|
my $rule = $hash_info->{payload}; |
460
|
13
|
|
|
|
|
31
|
my $package = $hash_info->{package}; |
461
|
|
|
|
|
|
|
|
462
|
|
|
|
|
|
|
# then_call is in caller hash. |
463
|
|
|
|
|
|
|
# when we get the grammarref, that will be the callee, so need to get next_call here and pass it in separtely |
464
|
|
|
|
|
|
|
# can't make it part of callee because multiple things could call this rule. |
465
|
13
|
|
|
|
|
31
|
my $then_call = $hash_info->{then_call}; |
466
|
|
|
|
|
|
|
|
467
|
13
|
50
|
|
|
|
42
|
if(length($package)){ |
468
|
13
|
|
|
|
|
35
|
$rule = $package.'::'.$rule; |
469
|
13
|
|
|
|
|
32
|
$then_call = $package.'::'.$then_call; |
470
|
|
|
|
|
|
|
} |
471
|
|
|
|
|
|
|
|
472
|
|
|
|
|
|
|
|
473
|
13
|
50
|
|
|
|
46
|
if($debug){warn "inside ParsingMethods::call. rule is '$rule', then_call is '$then_call'";} |
|
0
|
|
|
|
|
0
|
|
474
|
|
|
|
|
|
|
|
475
|
|
|
|
|
|
|
|
476
|
13
|
|
|
|
|
51
|
my $grammarref=$llobj->convert_rule_name_to_rule_reference($rule); |
477
|
|
|
|
|
|
|
|
478
|
13
|
50
|
|
|
|
50
|
if($debug){warn "grammarref for rule '$rule' is "; print Dumper $grammarref;} |
|
0
|
|
|
|
|
0
|
|
|
0
|
|
|
|
|
0
|
|
479
|
|
|
|
|
|
|
|
480
|
13
|
|
|
|
|
104
|
$llobj->parse_grammarref($grammarref, $then_call ); |
481
|
|
|
|
|
|
|
|
482
|
10
|
50
|
|
|
|
30
|
if($debug){warn "call returned";} |
|
0
|
|
|
|
|
0
|
|
483
|
10
|
|
|
|
|
34
|
return; # must have matched. |
484
|
|
|
|
|
|
|
|
485
|
|
|
|
|
|
|
} |
486
|
|
|
|
|
|
|
|
487
|
|
|
|
|
|
|
|
488
|
|
|
|
|
|
|
|
489
|
|
|
|
|
|
|
=head2 lit |
490
|
|
|
|
|
|
|
|
491
|
|
|
|
|
|
|
lit is short for literal. It is looking for the current letter object to match the letter value in $subrule. |
492
|
|
|
|
|
|
|
|
493
|
|
|
|
|
|
|
my $rule1=[ |
494
|
|
|
|
|
|
|
[ 'lit', 'a' ], |
495
|
|
|
|
|
|
|
[ 'lit', 'b' ], |
496
|
|
|
|
|
|
|
]; |
497
|
|
|
|
|
|
|
|
498
|
|
|
|
|
|
|
The above example is looking for 'a' followed by 'b'. |
499
|
|
|
|
|
|
|
|
500
|
|
|
|
|
|
|
=cut |
501
|
|
|
|
|
|
|
|
502
|
|
|
|
|
|
|
sub lit { |
503
|
112
|
|
|
112
|
1
|
161
|
my ($llobj, $subrule)=@_; |
504
|
|
|
|
|
|
|
|
505
|
112
|
|
|
|
|
223
|
my $grammar_letter=$subrule->[1]; |
506
|
112
|
|
|
|
|
185
|
my $letter_payload = $llobj->[LIST__CURR_START]->[LETTER__DATA_PAYLOAD]; |
507
|
|
|
|
|
|
|
#warn "lit rule '$grammar_letter' versus letter text '$letter_payload' "; |
508
|
112
|
100
|
|
|
|
256
|
if($grammar_letter ne $letter_payload){ |
509
|
44
|
|
|
|
|
707
|
die "GRAMMARFAIL"; |
510
|
|
|
|
|
|
|
} |
511
|
|
|
|
|
|
|
|
512
|
68
|
|
|
|
|
375
|
$llobj->[LIST__CURR_START]->[LETTER__LETTER_HAS_BEEN_CONSUMED]=1; |
513
|
|
|
|
|
|
|
} |
514
|
|
|
|
|
|
|
|
515
|
|
|
|
|
|
|
=head2 cc |
516
|
|
|
|
|
|
|
|
517
|
|
|
|
|
|
|
This is short for "character class". |
518
|
|
|
|
|
|
|
In perl regular expressions, this is represented with []. |
519
|
|
|
|
|
|
|
The letters in the square brackets are letters in teh character class you wnat to match. |
520
|
|
|
|
|
|
|
For example, [aeiou] would match a character class of any single vowel. |
521
|
|
|
|
|
|
|
|
522
|
|
|
|
|
|
|
=cut |
523
|
|
|
|
|
|
|
|
524
|
|
|
|
|
|
|
sub cc{ |
525
|
4
|
|
|
4
|
1
|
6
|
my ($llobj, $subrule)=@_; |
526
|
4
|
|
|
|
|
5
|
my $href_info=$subrule->[2]; |
527
|
4
|
|
|
|
|
7
|
my $hash_of_letters = $href_info->{hash_of_letters}; |
528
|
|
|
|
|
|
|
|
529
|
|
|
|
|
|
|
|
530
|
4
|
|
|
|
|
7
|
my $letter_payload = $llobj->[LIST__CURR_START]->[LETTER__DATA_PAYLOAD]; |
531
|
|
|
|
|
|
|
|
532
|
|
|
|
|
|
|
#print "called cc with letter_payload '$letter_payload' and class hash "; print Dumper $class_hashref; warn " "; |
533
|
|
|
|
|
|
|
|
534
|
4
|
100
|
|
|
|
20
|
unless(exists($hash_of_letters->{$letter_payload})){ |
535
|
|
|
|
|
|
|
#warn "dying "; |
536
|
2
|
|
|
|
|
25
|
die "GRAMMARFAIL"; |
537
|
|
|
|
|
|
|
} |
538
|
2
|
|
|
|
|
4
|
$llobj->[LIST__CURR_START]->[LETTER__LETTER_HAS_BEEN_CONSUMED]=1; |
539
|
|
|
|
|
|
|
} |
540
|
|
|
|
|
|
|
|
541
|
|
|
|
|
|
|
=head2 notcc |
542
|
|
|
|
|
|
|
|
543
|
|
|
|
|
|
|
This is short for "not character class". |
544
|
|
|
|
|
|
|
In perl regular expressions, this is represented with [^ ]. |
545
|
|
|
|
|
|
|
The letters in the square brackets are letters in teh character class you do NOT want to match. |
546
|
|
|
|
|
|
|
For example, [^aeiou] would NOT match a character class of any single vowel. |
547
|
|
|
|
|
|
|
Or it WOULD match any character that is NOT a vowel. |
548
|
|
|
|
|
|
|
|
549
|
|
|
|
|
|
|
=cut |
550
|
|
|
|
|
|
|
|
551
|
|
|
|
|
|
|
sub notcc{ |
552
|
2
|
|
|
2
|
1
|
3
|
my ($llobj, $subrule)=@_; |
553
|
2
|
|
|
|
|
3
|
my $href_info=$subrule->[2]; |
554
|
2
|
|
|
|
|
4
|
my $hash_of_letters = $href_info->{hash_of_letters}; |
555
|
|
|
|
|
|
|
|
556
|
|
|
|
|
|
|
|
557
|
2
|
|
|
|
|
4
|
my $letter_payload = $llobj->[LIST__CURR_START]->[LETTER__DATA_PAYLOAD]; |
558
|
|
|
|
|
|
|
|
559
|
|
|
|
|
|
|
#print "called cc with letter_payload '$letter_payload' and class hash "; print Dumper $class_hashref; warn " "; |
560
|
|
|
|
|
|
|
|
561
|
2
|
100
|
|
|
|
6
|
if(exists($hash_of_letters->{$letter_payload})){ |
562
|
|
|
|
|
|
|
#warn "dying "; |
563
|
1
|
|
|
|
|
11
|
die "GRAMMARFAIL"; |
564
|
|
|
|
|
|
|
} |
565
|
1
|
|
|
|
|
3
|
$llobj->[LIST__CURR_START]->[LETTER__LETTER_HAS_BEEN_CONSUMED]=1; |
566
|
|
|
|
|
|
|
} |
567
|
|
|
|
|
|
|
|
568
|
|
|
|
|
|
|
=head2 thrifty |
569
|
|
|
|
|
|
|
|
570
|
|
|
|
|
|
|
perform a thrifty quantifier match |
571
|
|
|
|
|
|
|
|
572
|
|
|
|
|
|
|
Note: Since we want to be able to read petabytes of streamed data, |
573
|
|
|
|
|
|
|
we will default to using thrifty matching. |
574
|
|
|
|
|
|
|
i.e. match as little as possible and move on. |
575
|
|
|
|
|
|
|
if we do greedy matching, then the first .* we run into will |
576
|
|
|
|
|
|
|
read in the entire stream (petabytes) into memory and crash the system. |
577
|
|
|
|
|
|
|
if it doesn't crash, it will back up until it finds amatch. |
578
|
|
|
|
|
|
|
We default to thrifty matching, meaning we only read in as little as possible |
579
|
|
|
|
|
|
|
to still find a match. This means we only read in just as much of the |
580
|
|
|
|
|
|
|
stream as we need to find a match. |
581
|
|
|
|
|
|
|
We can DO greedy matching, but it can be a problem if we're streaming massive quantities of data. |
582
|
|
|
|
|
|
|
|
583
|
|
|
|
|
|
|
basic thrifty algorithm: |
584
|
|
|
|
|
|
|
try the rule at least min times. |
585
|
|
|
|
|
|
|
if that matches, then return and let rest of grammar try. |
586
|
|
|
|
|
|
|
If rest of grammar dies, then revert to min location |
587
|
|
|
|
|
|
|
and try matching one more time. |
588
|
|
|
|
|
|
|
if that passes, then return and let rest of grammar try. |
589
|
|
|
|
|
|
|
if rest of grammar dies, then revert to min+1 location |
590
|
|
|
|
|
|
|
and try another rule. |
591
|
|
|
|
|
|
|
|
592
|
|
|
|
|
|
|
keep doing this until you reach "max" number of matches. |
593
|
|
|
|
|
|
|
if that doesn't make things happy, then quantifier dies |
594
|
|
|
|
|
|
|
and the expression fails. |
595
|
|
|
|
|
|
|
|
596
|
|
|
|
|
|
|
rule1 : 'a' rule2 'b' |
597
|
|
|
|
|
|
|
|
598
|
|
|
|
|
|
|
rule2 : 'c' d+ rule3 e+ |
599
|
|
|
|
|
|
|
|
600
|
|
|
|
|
|
|
rule3 : f g+ rule4 h |
601
|
|
|
|
|
|
|
|
602
|
|
|
|
|
|
|
rule4 : i* |
603
|
|
|
|
|
|
|
|
604
|
|
|
|
|
|
|
|
605
|
|
|
|
|
|
|
|
606
|
|
|
|
|
|
|
=cut |
607
|
|
|
|
|
|
|
|
608
|
|
|
|
|
|
|
sub thrifty { |
609
|
0
|
|
|
0
|
1
|
|
my ($llobj, $subrule)=@_; |
610
|
|
|
|
|
|
|
|
611
|
0
|
|
|
|
|
|
my $payload=$subrule->[1]; |
612
|
|
|
|
|
|
|
|
613
|
0
|
|
|
|
|
|
my $rule = $payload->{rule}; |
614
|
0
|
|
|
|
|
|
my $then_call = $payload->{then_call}; |
615
|
|
|
|
|
|
|
|
616
|
0
|
|
|
|
|
|
my $grammarref=$llobj->convert_grammar_name_to_array_ref($rule); |
617
|
|
|
|
|
|
|
|
618
|
0
|
|
|
|
|
|
$llobj->parse_grammarref($grammarref, $then_call ); |
619
|
|
|
|
|
|
|
|
620
|
|
|
|
|
|
|
|
621
|
|
|
|
|
|
|
|
622
|
|
|
|
|
|
|
|
623
|
0
|
|
|
|
|
|
return; # must have matched. |
624
|
|
|
|
|
|
|
} |
625
|
|
|
|
|
|
|
|
626
|
|
|
|
|
|
|
=head2 greedy |
627
|
|
|
|
|
|
|
|
628
|
|
|
|
|
|
|
basic greedy algorithm. |
629
|
|
|
|
|
|
|
try the rule max times. |
630
|
|
|
|
|
|
|
if not even zero match, die. |
631
|
|
|
|
|
|
|
at the end of every match, record the letter location of that specific match. |
632
|
|
|
|
|
|
|
|
633
|
|
|
|
|
|
|
return and let rest of grammar try. |
634
|
|
|
|
|
|
|
if rest of grammar dies, then revert to max-1 location, |
635
|
|
|
|
|
|
|
and try another rule. |
636
|
|
|
|
|
|
|
return and let rest of grammar try. |
637
|
|
|
|
|
|
|
if rest of grammar dies, then revert to max-2 location |
638
|
|
|
|
|
|
|
and try another rule. |
639
|
|
|
|
|
|
|
|
640
|
|
|
|
|
|
|
keep doing this until you reach "min" number of matches. |
641
|
|
|
|
|
|
|
we can't find a match even at "min", then quantifier dies |
642
|
|
|
|
|
|
|
and the expression fails. |
643
|
|
|
|
|
|
|
|
644
|
|
|
|
|
|
|
=cut |
645
|
|
|
|
|
|
|
|
646
|
|
|
|
|
|
|
sub greedy { |
647
|
0
|
|
|
0
|
1
|
|
my($llobj, $subrule, $overalldirectionforrule)=@_; |
648
|
|
|
|
|
|
|
|
649
|
0
|
|
|
|
|
|
my $payload=$subrule->[1]; |
650
|
|
|
|
|
|
|
|
651
|
0
|
|
|
|
|
|
my $min = $payload->{min}; |
652
|
0
|
|
|
|
|
|
my $max = $payload->{max}; |
653
|
0
|
|
|
|
|
|
my $rule = $payload->{rule}; |
654
|
|
|
|
|
|
|
|
655
|
|
|
|
|
|
|
|
656
|
|
|
|
|
|
|
} |
657
|
|
|
|
|
|
|
|
658
|
|
|
|
|
|
|
1; |
659
|
|
|
|
|
|
|
|