This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
PATCH: uniprops.t take advantage of EBCDIC test.pl
[perl5.git] / lib / unicore / mktables
CommitLineData
d73e5302 1#!/usr/bin/perl -w
99870f4d
KW
2
3# !!!!!!!!!!!!!! IF YOU MODIFY THIS FILE !!!!!!!!!!!!!!!!!!!!!!!!!
4# Any files created or read by this program should be listed in 'mktables.lst'
5# Use -makelist to regenerate it.
6
23e33b60
KW
7# Needs 'no overloading' to run faster on miniperl. Code commented out at the
8# subroutine objaddr can be used instead to work as far back (untested) as
f998e60c
KW
9# 5.8: needs pack "U". But almost all occurrences of objaddr have been
10# removed in favor of using 'no overloading'. You also would have to go
11# through and replace occurrences like:
12# my $addr; { no overloading; $addr = 0+$self; }
13# with
14# my $addr = main::objaddr $self;
6c68572b
KW
15# (or reverse commit 9b01bafde4b022706c3d6f947a0963f821b2e50b
16# that instituted this change.)
17
23e33b60 18require 5.010_001;
d73e5302 19use strict;
99870f4d 20use warnings;
cf25bb62 21use Carp;
99870f4d
KW
22use File::Find;
23use File::Path;
d07a55ed 24use File::Spec;
99870f4d
KW
25use Text::Tabs;
26
27sub DEBUG () { 0 } # Set to 0 for production; 1 for development
28
29##########################################################################
30#
31# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
32# from the Unicode database files (lib/unicore/.../*.txt), It also generates
33# a pod file and a .t file
34#
35# The structure of this file is:
36# First these introductory comments; then
37# code needed for everywhere, such as debugging stuff; then
38# code to handle input parameters; then
39# data structures likely to be of external interest (some of which depend on
40# the input parameters, so follows them; then
41# more data structures and subroutine and package (class) definitions; then
42# the small actual loop to process the input files and finish up; then
43# a __DATA__ section, for the .t tests
44#
45# This program works on all releases of Unicode through at least 5.2. The
46# outputs have been scrutinized most intently for release 5.1. The others
47# have been checked for somewhat more than just sanity. It can handle all
48# existing Unicode character properties in those releases.
49#
99870f4d
KW
50# This program is mostly about Unicode character (or code point) properties.
51# A property describes some attribute or quality of a code point, like if it
52# is lowercase or not, its name, what version of Unicode it was first defined
53# in, or what its uppercase equivalent is. Unicode deals with these disparate
54# possibilities by making all properties into mappings from each code point
55# into some corresponding value. In the case of it being lowercase or not,
56# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
57# property maps each Unicode code point to a single value, called a "property
58# value". (Hence each Unicode property is a true mathematical function with
59# exactly one value per code point.)
60#
61# When using a property in a regular expression, what is desired isn't the
62# mapping of the code point to its property's value, but the reverse (or the
63# mathematical "inverse relation"): starting with the property value, "Does a
64# code point map to it?" These are written in a "compound" form:
65# \p{property=value}, e.g., \p{category=punctuation}. This program generates
66# files containing the lists of code points that map to each such regular
67# expression property value, one file per list
68#
69# There is also a single form shortcut that Perl adds for many of the commonly
70# used properties. This happens for all binary properties, plus script,
71# general_category, and block properties.
72#
73# Thus the outputs of this program are files. There are map files, mostly in
74# the 'To' directory; and there are list files for use in regular expression
75# matching, all in subdirectories of the 'lib' directory, with each
76# subdirectory being named for the property that the lists in it are for.
77# Bookkeeping, test, and documentation files are also generated.
78
79my $matches_directory = 'lib'; # Where match (\p{}) files go.
80my $map_directory = 'To'; # Where map files go.
81
82# DATA STRUCTURES
83#
84# The major data structures of this program are Property, of course, but also
85# Table. There are two kinds of tables, very similar to each other.
86# "Match_Table" is the data structure giving the list of code points that have
87# a particular property value, mentioned above. There is also a "Map_Table"
88# data structure which gives the property's mapping from code point to value.
89# There are two structures because the match tables need to be combined in
90# various ways, such as constructing unions, intersections, complements, etc.,
91# and the map ones don't. And there would be problems, perhaps subtle, if
92# a map table were inadvertently operated on in some of those ways.
93# The use of separate classes with operations defined on one but not the other
94# prevents accidentally confusing the two.
95#
96# At the heart of each table's data structure is a "Range_List", which is just
97# an ordered list of "Ranges", plus ancillary information, and methods to
98# operate on them. A Range is a compact way to store property information.
99# Each range has a starting code point, an ending code point, and a value that
100# is meant to apply to all the code points between the two end points,
101# inclusive. For a map table, this value is the property value for those
102# code points. Two such ranges could be written like this:
103# 0x41 .. 0x5A, 'Upper',
104# 0x61 .. 0x7A, 'Lower'
105#
106# Each range also has a type used as a convenience to classify the values.
107# Most ranges in this program will be Type 0, or normal, but there are some
108# ranges that have a non-zero type. These are used only in map tables, and
109# are for mappings that don't fit into the normal scheme of things. Mappings
110# that require a hash entry to communicate with utf8.c are one example;
111# another example is mappings for charnames.pm to use which indicate a name
112# that is algorithmically determinable from its code point (and vice-versa).
113# These are used to significantly compact these tables, instead of listing
114# each one of the tens of thousands individually.
115#
116# In a match table, the value of a range is irrelevant (and hence the type as
117# well, which will always be 0), and arbitrarily set to the null string.
118# Using the example above, there would be two match tables for those two
119# entries, one named Upper would contain the 0x41..0x5A range, and the other
120# named Lower would contain 0x61..0x7A.
121#
122# Actually, there are two types of range lists, "Range_Map" is the one
123# associated with map tables, and "Range_List" with match tables.
124# Again, this is so that methods can be defined on one and not the other so as
125# to prevent operating on them in incorrect ways.
126#
127# Eventually, most tables are written out to files to be read by utf8_heavy.pl
128# in the perl core. All tables could in theory be written, but some are
129# suppressed because there is no current practical use for them. It is easy
130# to change which get written by changing various lists that are near the top
131# of the actual code in this file. The table data structures contain enough
132# ancillary information to allow them to be treated as separate entities for
133# writing, such as the path to each one's file. There is a heading in each
134# map table that gives the format of its entries, and what the map is for all
135# the code points missing from it. (This allows tables to be more compact.)
678f13d5 136#
99870f4d
KW
137# The Property data structure contains one or more tables. All properties
138# contain a map table (except the $perl property which is a
139# pseudo-property containing only match tables), and any properties that
140# are usable in regular expression matches also contain various matching
141# tables, one for each value the property can have. A binary property can
142# have two values, True and False (or Y and N, which are preferred by Unicode
143# terminology). Thus each of these properties will have a map table that
144# takes every code point and maps it to Y or N (but having ranges cuts the
145# number of entries in that table way down), and two match tables, one
146# which has a list of all the code points that map to Y, and one for all the
147# code points that map to N. (For each of these, a third table is also
148# generated for the pseudo Perl property. It contains the identical code
149# points as the Y table, but can be written, not in the compound form, but in
150# a "single" form like \p{IsUppercase}.) Many properties are binary, but some
151# properties have several possible values, some have many, and properties like
152# Name have a different value for every named code point. Those will not,
153# unless the controlling lists are changed, have their match tables written
154# out. But all the ones which can be used in regular expression \p{} and \P{}
155# constructs will. Generally a property will have either its map table or its
156# match tables written but not both. Again, what gets written is controlled
157# by lists which can easily be changed.
678f13d5 158#
99870f4d
KW
159# For information about the Unicode properties, see Unicode's UAX44 document:
160
161my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
162
163# As stated earlier, this program will work on any release of Unicode so far.
164# Most obvious problems in earlier data have NOT been corrected except when
165# necessary to make Perl or this program work reasonably. For example, no
166# folding information was given in early releases, so this program uses the
167# substitute of lower case, just so that a regular expression with the /i
168# option will do something that actually gives the right results in many
169# cases. There are also a couple other corrections for version 1.1.5,
170# commented at the point they are made. As an example of corrections that
171# weren't made (but could be) is this statement from DerivedAge.txt: "The
172# supplementary private use code points and the non-character code points were
173# assigned in version 2.0, but not specifically listed in the UCD until
174# versions 3.0 and 3.1 respectively." (To be precise it was 3.0.1 not 3.0.0)
175# More information on Unicode version glitches is further down in these
176# introductory comments.
177#
178# This program works on all properties as of 5.2, though the files for some
678f13d5
KW
179# are suppressed from apparent lack of demand for them. You can change which
180# are output by changing lists in this program.
181#
99870f4d
KW
182# The old version of mktables emphasized the term "Fuzzy" to mean Unocde's
183# loose matchings rules (from Unicode TR18):
184#
185# The recommended names for UCD properties and property values are in
186# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
187# [PropValue]. There are both abbreviated names and longer, more
188# descriptive names. It is strongly recommended that both names be
189# recognized, and that loose matching of property names be used,
190# whereby the case distinctions, whitespace, hyphens, and underbar
191# are ignored.
192# The program still allows Fuzzy to override its determination of if loose
193# matching should be used, but it isn't currently used, as it is no longer
194# needed; the calculations it makes are good enough.
678f13d5 195#
99870f4d
KW
196# SUMMARY OF HOW IT WORKS:
197#
198# Process arguments
199#
200# A list is constructed containing each input file that is to be processed
201#
202# Each file on the list is processed in a loop, using the associated handler
203# code for each:
204# The PropertyAliases.txt and PropValueAliases.txt files are processed
205# first. These files name the properties and property values.
206# Objects are created of all the property and property value names
207# that the rest of the input should expect, including all synonyms.
208# The other input files give mappings from properties to property
209# values. That is, they list code points and say what the mapping
210# is under the given property. Some files give the mappings for
211# just one property; and some for many. This program goes through
212# each file and populates the properties from them. Some properties
213# are listed in more than one file, and Unicode has set up a
214# precedence as to which has priority if there is a conflict. Thus
215# the order of processing matters, and this program handles the
216# conflict possibility by processing the overriding input files
217# last, so that if necessary they replace earlier values.
218# After this is all done, the program creates the property mappings not
219# furnished by Unicode, but derivable from what it does give.
220# The tables of code points that match each property value in each
221# property that is accessible by regular expressions are created.
222# The Perl-defined properties are created and populated. Many of these
223# require data determined from the earlier steps
224# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 225# and Unicode are reconciled and warned about.
99870f4d
KW
226# All the properties are written to files
227# Any other files are written, and final warnings issued.
678f13d5 228#
99870f4d
KW
229# For clarity, a number of operators have been overloaded to work on tables:
230# ~ means invert (take all characters not in the set). The more
231# conventional '!' is not used because of the possibility of confusing
232# it with the actual boolean operation.
233# + means union
234# - means subtraction
235# & means intersection
236# The precedence of these is the order listed. Parentheses should be
237# copiously used. These are not a general scheme. The operations aren't
238# defined for a number of things, deliberately, to avoid getting into trouble.
239# Operations are done on references and affect the underlying structures, so
240# that the copy constructors for them have been overloaded to not return a new
241# clone, but the input object itself.
678f13d5 242#
99870f4d
KW
243# The bool operator is deliberately not overloaded to avoid confusion with
244# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
245#
246# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
247#
248# This program needs to be able to run under miniperl. Therefore, it uses a
249# minimum of other modules, and hence implements some things itself that could
250# be gotten from CPAN
251#
252# This program uses inputs published by the Unicode Consortium. These can
253# change incompatibly between releases without the Perl maintainers realizing
254# it. Therefore this program is now designed to try to flag these. It looks
255# at the directories where the inputs are, and flags any unrecognized files.
256# It keeps track of all the properties in the files it handles, and flags any
257# that it doesn't know how to handle. It also flags any input lines that
258# don't match the expected syntax, among other checks.
259#
260# It is also designed so if a new input file matches one of the known
261# templates, one hopefully just needs to add it to a list to have it
262# processed.
263#
264# As mentioned earlier, some properties are given in more than one file. In
265# particular, the files in the extracted directory are supposedly just
266# reformattings of the others. But they contain information not easily
267# derivable from the other files, including results for Unihan, which this
268# program doesn't ordinarily look at, and for unassigned code points. They
269# also have historically had errors or been incomplete. In an attempt to
270# create the best possible data, this program thus processes them first to
271# glean information missing from the other files; then processes those other
272# files to override any errors in the extracted ones. Much of the design was
273# driven by this need to store things and then possibly override them.
274#
275# It tries to keep fatal errors to a minimum, to generate something usable for
276# testing purposes. It always looks for files that could be inputs, and will
277# warn about any that it doesn't know how to handle (the -q option suppresses
278# the warning).
99870f4d
KW
279#
280# Why have files written out for binary 'N' matches?
281# For binary properties, if you know the mapping for either Y or N; the
678f13d5
KW
282# other is trivial to construct, so could be done at Perl run-time by just
283# complementing the result, instead of having a file for it. That is, if
284# someone types in \p{foo: N}, Perl could translate that to \P{foo: Y} and
285# not need a file. The problem is communicating to Perl that a given
286# property is binary. Perl can't figure it out from looking at the N (or
287# No), as some non-binary properties have these as property values. So
288# rather than inventing a way to communicate this info back to the core,
289# which would have required changes there as well, it was simpler just to
290# add the extra tables.
291#
292# Why is there more than one type of range?
293# This simplified things. There are some very specialized code points that
294# have to be handled specially for output, such as Hangul syllable names.
295# By creating a range type (done late in the development process), it
296# allowed this to be stored with the range, and overridden by other input.
297# Originally these were stored in another data structure, and it became a
298# mess trying to decide if a second file that was for the same property was
299# overriding the earlier one or not.
300#
301# Why are there two kinds of tables, match and map?
302# (And there is a base class shared by the two as well.) As stated above,
303# they actually are for different things. Development proceeded much more
304# smoothly when I (khw) realized the distinction. Map tables are used to
305# give the property value for every code point (actually every code point
306# that doesn't map to a default value). Match tables are used for regular
307# expression matches, and are essentially the inverse mapping. Separating
308# the two allows more specialized methods, and error checks so that one
309# can't just take the intersection of two map tables, for example, as that
310# is nonsensical.
99870f4d
KW
311#
312# There are no match tables generated for matches of the null string. These
678f13d5
KW
313# would like like qr/\p{JSN=}/ currently without modifying the regex code.
314# Perhaps something like them could be added if necessary. The JSN does have
315# a real code point U+110B that maps to the null string, but it is a
316# contributory property, and therefore not output by default. And it's easily
317# handled so far by making the null string the default where it is a
318# possibility.
99870f4d 319#
23e33b60
KW
320# DEBUGGING
321#
678f13d5
KW
322# This program is written so it will run under miniperl. Occasionally changes
323# will cause an error where the backtrace doesn't work well under miniperl.
324# To diagnose the problem, you can instead run it under regular perl, if you
325# have one compiled.
326#
327# There is a good trace facility. To enable it, first sub DEBUG must be set
328# to return true. Then a line like
329#
330# local $to_trace = 1 if main::DEBUG;
331#
332# can be added to enable tracing in its lexical scope or until you insert
333# another line:
334#
335# local $to_trace = 0 if main::DEBUG;
336#
337# then use a line like "trace $a, @b, %c, ...;
338#
339# Some of the more complex subroutines already have trace statements in them.
340# Permanent trace statements should be like:
341#
342# trace ... if main::DEBUG && $to_trace;
343#
344# If there is just one or a few files that you're debugging, you can easily
345# cause most everything else to be skipped. Change the line
346#
347# my $debug_skip = 0;
348#
349# to 1, and every file whose object is in @input_file_objects and doesn't have
350# a, 'non_skip => 1,' in its constructor will be skipped.
351#
99870f4d
KW
352# FUTURE ISSUES
353#
354# The program would break if Unicode were to change its names so that
355# interior white space, underscores, or dashes differences were significant
356# within property and property value names.
357#
358# It might be easier to use the xml versions of the UCD if this program ever
359# would need heavy revision, and the ability to handle old versions was not
360# required.
361#
362# There is the potential for name collisions, in that Perl has chosen names
363# that Unicode could decide it also likes. There have been such collisions in
364# the past, with mostly Perl deciding to adopt the Unicode definition of the
365# name. However in the 5.2 Unicode beta testing, there were a number of such
366# collisions, which were withdrawn before the final release, because of Perl's
367# and other's protests. These all involved new properties which began with
368# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
369# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
370# Unicode document, so they are unlikely to be used by Unicode for another
371# purpose. However, they might try something beginning with 'In', or use any
372# of the other Perl-defined properties. This program will warn you of name
373# collisions, and refuse to generate tables with them, but manual intervention
374# will be required in this event. One scheme that could be implemented, if
375# necessary, would be to have this program generate another file, or add a
376# field to mktables.lst that gives the date of first definition of a property.
377# Each new release of Unicode would use that file as a basis for the next
378# iteration. And the Perl synonym addition code could sort based on the age
379# of the property, so older properties get priority, and newer ones that clash
380# would be refused; hence existing code would not be impacted, and some other
381# synonym would have to be used for the new property. This is ugly, and
382# manual intervention would certainly be easier to do in the short run; lets
383# hope it never comes to this.
678f13d5 384#
99870f4d
KW
385# A NOTE ON UNIHAN
386#
387# This program can generate tables from the Unihan database. But it doesn't
388# by default, letting the CPAN module Unicode::Unihan handle them. Prior to
389# version 5.2, this database was in a single file, Unihan.txt. In 5.2 the
390# database was split into 8 different files, all beginning with the letters
391# 'Unihan'. This program will read those file(s) if present, but it needs to
392# know which of the many properties in the file(s) should have tables created
393# for them. It will create tables for any properties listed in
394# PropertyAliases.txt and PropValueAliases.txt, plus any listed in the
395# @cjk_properties array and the @cjk_property_values array. Thus, if a
396# property you want is not in those files of the release you are building
397# against, you must add it to those two arrays. Starting in 4.0, the
398# Unicode_Radical_Stroke was listed in those files, so if the Unihan database
399# is present in the directory, a table will be generated for that property.
400# In 5.2, several more properties were added. For your convenience, the two
401# arrays are initialized with all the 5.2 listed properties that are also in
402# earlier releases. But these are commented out. You can just uncomment the
403# ones you want, or use them as a template for adding entries for other
404# properties.
405#
406# You may need to adjust the entries to suit your purposes. setup_unihan(),
407# and filter_unihan_line() are the functions where this is done. This program
408# already does some adjusting to make the lines look more like the rest of the
409# Unicode DB; You can see what that is in filter_unihan_line()
410#
411# There is a bug in the 3.2 data file in which some values for the
412# kPrimaryNumeric property have commas and an unexpected comment. A filter
413# could be added for these; or for a particular installation, the Unihan.txt
414# file could be edited to fix them.
415# have to be
416#
678f13d5
KW
417# HOW TO ADD A FILE TO BE PROCESSED
418#
419# A new file from Unicode needs to have an object constructed for it in
420# @input_file_objects, probably at the end or at the end of the extracted
421# ones. The program should warn you if its name will clash with others on
422# restrictive file systems, like DOS. If so, figure out a better name, and
423# add lines to the README.perl file giving that. If the file is a character
424# property, it should be in the format that Unicode has by default
425# standardized for such files for the more recently introduced ones.
426# If so, the Input_file constructor for @input_file_objects can just be the
427# file name and release it first appeared in. If not, then it should be
428# possible to construct an each_line_handler() to massage the line into the
429# standardized form.
430#
431# For non-character properties, more code will be needed. You can look at
432# the existing entries for clues.
433#
434# UNICODE VERSIONS NOTES
435#
436# The Unicode UCD has had a number of errors in it over the versions. And
437# these remain, by policy, in the standard for that version. Therefore it is
438# risky to correct them, because code may be expecting the error. So this
439# program doesn't generally make changes, unless the error breaks the Perl
440# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
441# for U+1105, which causes real problems for the algorithms for Jamo
442# calculations, so it is changed here.
443#
444# But it isn't so clear cut as to what to do about concepts that are
445# introduced in a later release; should they extend back to earlier releases
446# where the concept just didn't exist? It was easier to do this than to not,
447# so that's what was done. For example, the default value for code points not
448# in the files for various properties was probably undefined until changed by
449# some version. No_Block for blocks is such an example. This program will
450# assign No_Block even in Unicode versions that didn't have it. This has the
451# benefit that code being written doesn't have to special case earlier
452# versions; and the detriment that it doesn't match the Standard precisely for
453# the affected versions.
454#
455# Here are some observations about some of the issues in early versions:
456#
457# The number of code points in \p{alpha} halve in 2.1.9. It turns out that
458# the reason is that the CJK block starting at 4E00 was removed from PropList,
459# and was not put back in until 3.1.0
460#
461# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
462# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
463# reason is that 3.2 introduced U+205F=medium math space, which was not
464# classed as white space, but Perl figured out that it should have been. 4.0
465# reclassified it correctly.
466#
467# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
468# this was erroneously a synonym for 202. In 4.0, ATB became 202, and ATBL
469# was left with no code points, as all the ones that mapped to 202 stayed
470# mapped to 202. Thus if your program used the numeric name for the class,
471# it would not have been affected, but if it used the mnemonic, it would have
472# been.
473#
474# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that code
475# points which eventually came to have this script property value, instead
476# mapped to "Unknown". But in the next release all these code points were
477# moved to \p{sc=common} instead.
99870f4d
KW
478#
479# The default for missing code points for BidiClass is complicated. Starting
480# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
481# tries to do the best it can for earlier releases. It is done in
482# process_PropertyAliases()
483#
484##############################################################################
485
486my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
487 # and errors
488my $MAX_LINE_WIDTH = 78;
489
490# Debugging aid to skip most files so as to not be distracted by them when
491# concentrating on the ones being debugged. Add
492# non_skip => 1,
493# to the constructor for those files you want processed when you set this.
494# Files with a first version number of 0 are special: they are always
495# processed regardless of the state of this flag.
496my $debug_skip = 0;
497
498# Set to 1 to enable tracing.
499our $to_trace = 0;
500
501{ # Closure for trace: debugging aid
502 my $print_caller = 1; # ? Include calling subroutine name
503 my $main_with_colon = 'main::';
504 my $main_colon_length = length($main_with_colon);
505
506 sub trace {
507 return unless $to_trace; # Do nothing if global flag not set
508
509 my @input = @_;
510
511 local $DB::trace = 0;
512 $DB::trace = 0; # Quiet 'used only once' message
513
514 my $line_number;
515
516 # Loop looking up the stack to get the first non-trace caller
517 my $caller_line;
518 my $caller_name;
519 my $i = 0;
520 do {
521 $line_number = $caller_line;
522 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
523 $caller = $main_with_colon unless defined $caller;
524
525 $caller_name = $caller;
526
527 # get rid of pkg
528 $caller_name =~ s/.*:://;
529 if (substr($caller_name, 0, $main_colon_length)
530 eq $main_with_colon)
531 {
532 $caller_name = substr($caller_name, $main_colon_length);
533 }
534
535 } until ($caller_name ne 'trace');
536
537 # If the stack was empty, we were called from the top level
538 $caller_name = 'main' if ($caller_name eq ""
539 || $caller_name eq 'trace');
540
541 my $output = "";
542 foreach my $string (@input) {
543 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
544 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
545 $output .= simple_dumper($string);
546 }
547 else {
548 $string = "$string" if ref $string;
549 $string = $UNDEF unless defined $string;
550 chomp $string;
551 $string = '""' if $string eq "";
552 $output .= " " if $output ne ""
553 && $string ne ""
554 && substr($output, -1, 1) ne " "
555 && substr($string, 0, 1) ne " ";
556 $output .= $string;
557 }
558 }
559
99f78760
KW
560 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
561 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
562 print STDERR $output, "\n";
563 return;
564 }
565}
566
567# This is for a rarely used development feature that allows you to compare two
568# versions of the Unicode standard without having to deal with changes caused
569# by the code points introduced in the later verson. Change the 0 to a SINGLE
570# dotted Unicode release number (e.g. 2.1). Only code points introduced in
571# that release and earlier will be used; later ones are thrown away. You use
572# the version number of the earliest one you want to compare; then run this
573# program on directory structures containing each release, and compare the
574# outputs. These outputs will therefore include only the code points common
575# to both releases, and you can see the changes caused just by the underlying
576# release semantic changes. For versions earlier than 3.2, you must copy a
577# version of DAge.txt into the directory.
578my $string_compare_versions = DEBUG && 0; # e.g., v2.1;
579my $compare_versions = DEBUG
580 && $string_compare_versions
581 && pack "C*", split /\./, $string_compare_versions;
582
583sub uniques {
584 # Returns non-duplicated input values. From "Perl Best Practices:
585 # Encapsulated Cleverness". p. 455 in first edition.
586
587 my %seen;
588 return grep { ! $seen{$_}++ } @_;
589}
590
591$0 = File::Spec->canonpath($0);
592
593my $make_test_script = 0; # ? Should we output a test script
594my $write_unchanged_files = 0; # ? Should we update the output files even if
595 # we don't think they have changed
596my $use_directory = ""; # ? Should we chdir somewhere.
597my $pod_directory; # input directory to store the pod file.
598my $pod_file = 'perluniprops';
599my $t_path; # Path to the .t test file
600my $file_list = 'mktables.lst'; # File to store input and output file names.
601 # This is used to speed up the build, by not
602 # executing the main body of the program if
603 # nothing on the list has changed since the
604 # previous build
605my $make_list = 1; # ? Should we write $file_list. Set to always
606 # make a list so that when the pumpking is
607 # preparing a release, s/he won't have to do
608 # special things
609my $glob_list = 0; # ? Should we try to include unknown .txt files
610 # in the input.
611my $output_range_counts = 1; # ? Should we include the number of code points
612 # in ranges in the output
9ef2b94f
KW
613my $output_names = 0; # ? Should character names be in the output
614my @viacode; # Contains the 1 million character names, if
615 # $output_names is true
616
99870f4d
KW
617# Verbosity levels; 0 is quiet
618my $NORMAL_VERBOSITY = 1;
619my $PROGRESS = 2;
620my $VERBOSE = 3;
621
622my $verbosity = $NORMAL_VERBOSITY;
623
624# Process arguments
625while (@ARGV) {
cf25bb62
JH
626 my $arg = shift @ARGV;
627 if ($arg eq '-v') {
99870f4d
KW
628 $verbosity = $VERBOSE;
629 }
630 elsif ($arg eq '-p') {
631 $verbosity = $PROGRESS;
632 $| = 1; # Flush buffers as we go.
633 }
634 elsif ($arg eq '-q') {
635 $verbosity = 0;
636 }
637 elsif ($arg eq '-w') {
638 $write_unchanged_files = 1; # update the files even if havent changed
639 }
640 elsif ($arg eq '-check') {
6ae7e459
YO
641 my $this = shift @ARGV;
642 my $ok = shift @ARGV;
643 if ($this ne $ok) {
644 print "Skipping as check params are not the same.\n";
645 exit(0);
646 }
00a8df5c 647 }
99870f4d
KW
648 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
649 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
650 }
3df51b85
KW
651 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
652 {
99870f4d 653 $make_test_script = 1;
99870f4d
KW
654 }
655 elsif ($arg eq '-makelist') {
656 $make_list = 1;
657 }
658 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
659 -d $use_directory or croak "Unknown directory '$use_directory'";
660 }
661 elsif ($arg eq '-L') {
662
663 # Existence not tested until have chdir'd
664 $file_list = shift;
665 }
666 elsif ($arg eq '-globlist') {
667 $glob_list = 1;
668 }
669 elsif ($arg eq '-c') {
670 $output_range_counts = ! $output_range_counts
671 }
9ef2b94f
KW
672 elsif ($arg eq '-output_names') {
673 $output_names = 1;
674 }
99870f4d
KW
675 else {
676 my $with_c = 'with';
677 $with_c .= 'out' if $output_range_counts; # Complements the state
678 croak <<END;
679usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
680 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
681 [-check A B ]
682 -c : Output comments $with_c number of code points in ranges
683 -q : Quiet Mode: Only output serious warnings.
684 -p : Set verbosity level to normal plus show progress.
685 -v : Set Verbosity level high: Show progress and non-serious
686 warnings
687 -w : Write files regardless
688 -C dir : Change to this directory before proceeding. All relative paths
689 except those specified by the -P and -T options will be done
690 with respect to this directory.
691 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 692 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
693 -L filelist : Use alternate 'filelist' instead of standard one
694 -globlist : Take as input all non-Test *.txt files in current and sub
695 directories
3df51b85
KW
696 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
697 overrides -T
99870f4d 698 -makelist : Rewrite the file list $file_list based on current setup
9ef2b94f
KW
699 -output_names : Output each character's name in the table files; useful for
700 doing what-ifs, looking at diffs; is slow, memory intensive,
701 resulting tables are usable but very large.
99870f4d
KW
702 -check A B : Executes $0 only if A and B are the same
703END
704 }
705}
706
707# Stores the most-recently changed file. If none have changed, can skip the
708# build
709my $youngest = -M $0; # Do this before the chdir!
710
711# Change directories now, because need to read 'version' early.
712if ($use_directory) {
3df51b85 713 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
714 $pod_directory = File::Spec->rel2abs($pod_directory);
715 }
3df51b85 716 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 717 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 718 }
99870f4d 719 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 720 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 721 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 722 }
3df51b85 723 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 724 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 725 }
00a8df5c
YO
726}
727
99870f4d
KW
728# Get Unicode version into regular and v-string. This is done now because
729# various tables below get populated based on it. These tables are populated
730# here to be near the top of the file, and so easily seeable by those needing
731# to modify things.
732open my $VERSION, "<", "version"
733 or croak "$0: can't open required file 'version': $!\n";
734my $string_version = <$VERSION>;
735close $VERSION;
736chomp $string_version;
737my $v_version = pack "C*", split /\./, $string_version; # v string
738
739# The following are the complete names of properties with property values that
740# are known to not match any code points in some versions of Unicode, but that
741# may change in the future so they should be matchable, hence an empty file is
742# generated for them.
743my @tables_that_may_be_empty = (
744 'Joining_Type=Left_Joining',
745 );
746push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
747push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
748push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
749 if $v_version ge v4.1.0;
750
751# The lists below are hashes, so the key is the item in the list, and the
752# value is the reason why it is in the list. This makes generation of
753# documentation easier.
754
755my %why_suppressed; # No file generated for these.
756
757# Files aren't generated for empty extraneous properties. This is arguable.
758# Extraneous properties generally come about because a property is no longer
759# used in a newer version of Unicode. If we generated a file without code
760# points, programs that used to work on that property will still execute
761# without errors. It just won't ever match (or will always match, with \P{}).
762# This means that the logic is now likely wrong. I (khw) think its better to
763# find this out by getting an error message. Just move them to the table
764# above to change this behavior
765my %why_suppress_if_empty_warn_if_not = (
766
767 # It is the only property that has ever officially been removed from the
768 # Standard. The database never contained any code points for it.
769 'Special_Case_Condition' => 'Obsolete',
770
771 # Apparently never official, but there were code points in some versions of
772 # old-style PropList.txt
773 'Non_Break' => 'Obsolete',
774);
775
776# These would normally go in the warn table just above, but they were changed
777# a long time before this program was written, so warnings about them are
778# moot.
779if ($v_version gt v3.2.0) {
780 push @tables_that_may_be_empty,
781 'Canonical_Combining_Class=Attached_Below_Left'
782}
783
784# These are listed in the Property aliases file in 5.2, but Unihan is ignored
785# unless explicitly added.
786if ($v_version ge v5.2.0) {
787 my $unihan = 'Unihan; remove from list if using Unihan';
23e33b60 788 foreach my $table qw (
99870f4d
KW
789 kAccountingNumeric
790 kOtherNumeric
791 kPrimaryNumeric
792 kCompatibilityVariant
793 kIICore
794 kIRG_GSource
795 kIRG_HSource
796 kIRG_JSource
797 kIRG_KPSource
798 kIRG_MSource
799 kIRG_KSource
800 kIRG_TSource
801 kIRG_USource
802 kIRG_VSource
803 kRSUnicode
804 )
805 {
806 $why_suppress_if_empty_warn_if_not{$table} = $unihan;
807 }
ca12659b
NC
808}
809
99870f4d
KW
810# Properties that this program ignores.
811my @unimplemented_properties = (
812'Unicode_Radical_Stroke' # Remove if changing to handle this one.
813);
d73e5302 814
99870f4d
KW
815# There are several types of obsolete properties defined by Unicode. These
816# must be hand-edited for every new Unicode release.
817my %why_deprecated; # Generates a deprecated warning message if used.
818my %why_stabilized; # Documentation only
819my %why_obsolete; # Documentation only
820
821{ # Closure
822 my $simple = 'Perl uses the more complete version of this property';
823 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
824
825 my $other_properties = 'other properties';
826 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
827 my $why_no_expand = "Easily computed, and yet doesn't cover the common encoding forms (UTF-16/8)",
828
829 %why_deprecated = (
830 'Grapheme_Link' => 'Deprecated by Unicode. Use ccc=vr (Canonical_Combining_Class=Virama) instead',
831 'Jamo_Short_Name' => $contributory,
832 'Line_Break=Surrogate' => 'Deprecated by Unicode because surrogates should never appear in well-formed text, and therefore shouldn\'t be the basis for line breaking',
833 'Other_Alphabetic' => $contributory,
834 'Other_Default_Ignorable_Code_Point' => $contributory,
835 'Other_Grapheme_Extend' => $contributory,
836 'Other_ID_Continue' => $contributory,
837 'Other_ID_Start' => $contributory,
838 'Other_Lowercase' => $contributory,
839 'Other_Math' => $contributory,
840 'Other_Uppercase' => $contributory,
841 );
842
843 %why_suppressed = (
844 # There is a lib/unicore/Decomposition.pl (used by normalize.pm) which
845 # contains the same information, but without the algorithmically
846 # determinable Hangul syllables'. This file is not published, so it's
847 # existence is not noted in the comment.
848 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize',
849
850 'ISO_Comment' => 'Apparently no demand for it, but can access it through Unicode::UCD::charinfo. Obsoleted, and code points for it removed in Unicode 5.2',
851 'Unicode_1_Name' => "$simple, and no apparent demand for it, but can access it through Unicode::UCD::charinfo. If there is no later name for a code point, then this one is used instead in charnames",
852
853 'Simple_Case_Folding' => "$simple. Can access this through Unicode::UCD::casefold",
854 'Simple_Lowercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
855 'Simple_Titlecase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
856 'Simple_Uppercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
857
858 'Name' => "Accessible via 'use charnames;'",
859 'Name_Alias' => "Accessible via 'use charnames;'",
860
861 # These are sort of jumping the gun; deprecation is proposed for
862 # Unicode version 6.0, but they have never been exposed by Perl, and
863 # likely are soon to be deprecated, so best not to expose them.
864 FC_NFKC_Closure => 'Use NFKC_Casefold instead',
865 Expands_On_NFC => $why_no_expand,
866 Expands_On_NFD => $why_no_expand,
867 Expands_On_NFKC => $why_no_expand,
868 Expands_On_NFKD => $why_no_expand,
869 );
870
871 # The following are suppressed because they were made contributory or
872 # deprecated by Unicode before Perl ever thought about supporting them.
873 foreach my $property ('Jamo_Short_Name', 'Grapheme_Link') {
874 $why_suppressed{$property} = $why_deprecated{$property};
875 }
cf25bb62 876
99870f4d
KW
877 # Customize the message for all the 'Other_' properties
878 foreach my $property (keys %why_deprecated) {
879 next if (my $main_property = $property) !~ s/^Other_//;
880 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
881 }
882}
883
884if ($v_version ge 4.0.0) {
885 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
886}
887if ($v_version ge 5.2.0) {
888 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
889}
890
891# Probably obsolete forever
892if ($v_version ge v4.1.0) {
893 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common"';
894}
895
896# This program can create files for enumerated-like properties, such as
897# 'Numeric_Type'. This file would be the same format as for a string
898# property, with a mapping from code point to its value, so you could look up,
899# for example, the script a code point is in. But no one so far wants this
900# mapping, or they have found another way to get it since this is a new
901# feature. So no file is generated except if it is in this list.
902my @output_mapped_properties = split "\n", <<END;
903END
904
905# If you are using the Unihan database, you need to add the properties that
906# you want to extract from it to this table. For your convenience, the
907# properties in the 5.2 PropertyAliases.txt file are listed, commented out
908my @cjk_properties = split "\n", <<'END';
909#cjkAccountingNumeric; kAccountingNumeric
910#cjkOtherNumeric; kOtherNumeric
911#cjkPrimaryNumeric; kPrimaryNumeric
912#cjkCompatibilityVariant; kCompatibilityVariant
913#cjkIICore ; kIICore
914#cjkIRG_GSource; kIRG_GSource
915#cjkIRG_HSource; kIRG_HSource
916#cjkIRG_JSource; kIRG_JSource
917#cjkIRG_KPSource; kIRG_KPSource
918#cjkIRG_KSource; kIRG_KSource
919#cjkIRG_TSource; kIRG_TSource
920#cjkIRG_USource; kIRG_USource
921#cjkIRG_VSource; kIRG_VSource
922#cjkRSUnicode; kRSUnicode ; Unicode_Radical_Stroke; URS
923END
924
925# Similarly for the property values. For your convenience, the lines in the
926# 5.2 PropertyAliases.txt file are listed. Just remove the first BUT NOT both
927# '#' marks
928my @cjk_property_values = split "\n", <<'END';
929## @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
930## @missing: 0000..10FFFF; cjkCompatibilityVariant; <code point>
931## @missing: 0000..10FFFF; cjkIICore; <none>
932## @missing: 0000..10FFFF; cjkIRG_GSource; <none>
933## @missing: 0000..10FFFF; cjkIRG_HSource; <none>
934## @missing: 0000..10FFFF; cjkIRG_JSource; <none>
935## @missing: 0000..10FFFF; cjkIRG_KPSource; <none>
936## @missing: 0000..10FFFF; cjkIRG_KSource; <none>
937## @missing: 0000..10FFFF; cjkIRG_TSource; <none>
938## @missing: 0000..10FFFF; cjkIRG_USource; <none>
939## @missing: 0000..10FFFF; cjkIRG_VSource; <none>
940## @missing: 0000..10FFFF; cjkOtherNumeric; NaN
941## @missing: 0000..10FFFF; cjkPrimaryNumeric; NaN
942## @missing: 0000..10FFFF; cjkRSUnicode; <none>
943END
944
945# The input files don't list every code point. Those not listed are to be
946# defaulted to some value. Below are hard-coded what those values are for
947# non-binary properties as of 5.1. Starting in 5.0, there are
948# machine-parsable comment lines in the files the give the defaults; so this
949# list shouldn't have to be extended. The claim is that all missing entries
950# for binary properties will default to 'N'. Unicode tried to change that in
951# 5.2, but the beta period produced enough protest that they backed off.
952#
953# The defaults for the fields that appear in UnicodeData.txt in this hash must
954# be in the form that it expects. The others may be synonyms.
955my $CODE_POINT = '<code point>';
956my %default_mapping = (
957 Age => "Unassigned",
958 # Bidi_Class => Complicated; set in code
959 Bidi_Mirroring_Glyph => "",
960 Block => 'No_Block',
961 Canonical_Combining_Class => 0,
962 Case_Folding => $CODE_POINT,
963 Decomposition_Mapping => $CODE_POINT,
964 Decomposition_Type => 'None',
965 East_Asian_Width => "Neutral",
966 FC_NFKC_Closure => $CODE_POINT,
967 General_Category => 'Cn',
968 Grapheme_Cluster_Break => 'Other',
969 Hangul_Syllable_Type => 'NA',
970 ISO_Comment => "",
971 Jamo_Short_Name => "",
972 Joining_Group => "No_Joining_Group",
973 # Joining_Type => Complicated; set in code
974 kIICore => 'N', # Is converted to binary
975 #Line_Break => Complicated; set in code
976 Lowercase_Mapping => $CODE_POINT,
977 Name => "",
978 Name_Alias => "",
979 NFC_QC => 'Yes',
980 NFD_QC => 'Yes',
981 NFKC_QC => 'Yes',
982 NFKD_QC => 'Yes',
983 Numeric_Type => 'None',
984 Numeric_Value => 'NaN',
985 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
986 Sentence_Break => 'Other',
987 Simple_Case_Folding => $CODE_POINT,
988 Simple_Lowercase_Mapping => $CODE_POINT,
989 Simple_Titlecase_Mapping => $CODE_POINT,
990 Simple_Uppercase_Mapping => $CODE_POINT,
991 Titlecase_Mapping => $CODE_POINT,
992 Unicode_1_Name => "",
993 Unicode_Radical_Stroke => "",
994 Uppercase_Mapping => $CODE_POINT,
995 Word_Break => 'Other',
996);
997
998# Below are files that Unicode furnishes, but this program ignores, and why
999my %ignored_files = (
1000 'CJKRadicals.txt' => 'Unihan data',
1001 'Index.txt' => 'An index, not actual data',
1002 'NamedSqProv.txt' => 'Not officially part of the Unicode standard; Append it to NamedSequences.txt if you want to process the contents.',
1003 'NamesList.txt' => 'Just adds commentary',
1004 'NormalizationCorrections.txt' => 'Data is already in other files.',
1005 'Props.txt' => 'Adds nothing to PropList.txt; only in very early releases',
1006 'ReadMe.txt' => 'Just comments',
1007 'README.TXT' => 'Just comments',
1008 'StandardizedVariants.txt' => 'Only for glyph changes, not a Unicode character property. Does not fit into current scheme where one code point is mapped',
1009);
1010
678f13d5 1011### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1012
1013my $HEADER=<<"EOF";
1014# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85
KW
1015# This file is machine-generated by $0 from the Unicode
1016# database, Version $string_version. Any changes made here will be lost!
cf25bb62
JH
1017EOF
1018
b6922eda 1019my $INTERNAL_ONLY=<<"EOF";
99870f4d
KW
1020
1021# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
b6922eda 1022# This file is for internal use by the Perl program only. The format and even
99870f4d
KW
1023# the name or existence of this file are subject to change without notice.
1024# Don't use it directly.
1025EOF
1026
1027my $DEVELOPMENT_ONLY=<<"EOF";
1028# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1029# This file contains information artificially constrained to code points
1030# present in Unicode release $string_compare_versions.
1031# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1032# not be used for production.
b6922eda
KW
1033
1034EOF
1035
99870f4d
KW
1036my $LAST_UNICODE_CODEPOINT_STRING = "10FFFF";
1037my $LAST_UNICODE_CODEPOINT = hex $LAST_UNICODE_CODEPOINT_STRING;
1038my $MAX_UNICODE_CODEPOINTS = $LAST_UNICODE_CODEPOINT + 1;
1039
1040# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1041# two must be 10; if there are 5, the first must not be a 0. Written this way
1042# to decrease backtracking
1043my $code_point_re =
1044 qr/ \b (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1045
1046# This matches the beginning of the line in the Unicode db files that give the
1047# defaults for code points not listed (i.e., missing) in the file. The code
1048# depends on this ending with a semi-colon, so it can assume it is a valid
1049# field when the line is split() by semi-colons
1050my $missing_defaults_prefix =
1051 qr/^#\s+\@missing:\s+0000\.\.$LAST_UNICODE_CODEPOINT_STRING\s*;/;
1052
1053# Property types. Unicode has more types, but these are sufficient for our
1054# purposes.
1055my $UNKNOWN = -1; # initialized to illegal value
1056my $NON_STRING = 1; # Either binary or enum
1057my $BINARY = 2;
1058my $ENUM = 3; # Include catalog
1059my $STRING = 4; # Anything else: string or misc
1060
1061# Some input files have lines that give default values for code points not
1062# contained in the file. Sometimes these should be ignored.
1063my $NO_DEFAULTS = 0; # Must evaluate to false
1064my $NOT_IGNORED = 1;
1065my $IGNORED = 2;
1066
1067# Range types. Each range has a type. Most ranges are type 0, for normal,
1068# and will appear in the main body of the tables in the output files, but
1069# there are other types of ranges as well, listed below, that are specially
1070# handled. There are pseudo-types as well that will never be stored as a
1071# type, but will affect the calculation of the type.
1072
1073# 0 is for normal, non-specials
1074my $MULTI_CP = 1; # Sequence of more than code point
1075my $HANGUL_SYLLABLE = 2;
1076my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1077my $NULL = 4; # The map is to the null string; utf8.c can't
1078 # handle these, nor is there an accepted syntax
1079 # for them in \p{} constructs
f86864ac 1080my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1081 # otherwise be $MULTI_CP type are instead type 0
1082
1083# process_generic_property_file() can accept certain overrides in its input.
1084# Each of these must begin AND end with $CMD_DELIM.
1085my $CMD_DELIM = "\a";
1086my $REPLACE_CMD = 'replace'; # Override the Replace
1087my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1088
1089my $NO = 0;
1090my $YES = 1;
1091
1092# Values for the Replace argument to add_range.
1093# $NO # Don't replace; add only the code points not
1094 # already present.
1095my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1096 # the comments at the subroutine definition.
1097my $UNCONDITIONALLY = 2; # Replace without conditions.
1098my $MULTIPLE = 4; # Don't replace, but add a duplicate record if
1099 # already there
1100
1101# Flags to give property statuses. The phrases are to remind maintainers that
1102# if the flag is changed, the indefinite article referring to it in the
1103# documentation may need to be as well.
1104my $NORMAL = "";
1105my $SUPPRESSED = 'z'; # The character should never actually be seen, since
1106 # it is suppressed
37e2e78e 1107my $PLACEHOLDER = 'P'; # Implies no pod entry generated
99870f4d
KW
1108my $DEPRECATED = 'D';
1109my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1110my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1111my $DISCOURAGED = 'X';
1112my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1113my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1114my $STRICTER = 'T';
1115my $a_bold_stricter = "a 'B<$STRICTER>'";
1116my $A_bold_stricter = "A 'B<$STRICTER>'";
1117my $STABILIZED = 'S';
1118my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1119my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1120my $OBSOLETE = 'O';
1121my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1122my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1123
1124my %status_past_participles = (
1125 $DISCOURAGED => 'discouraged',
1126 $SUPPRESSED => 'should never be generated',
1127 $STABILIZED => 'stabilized',
1128 $OBSOLETE => 'obsolete',
37e2e78e 1129 $DEPRECATED => 'deprecated',
99870f4d
KW
1130);
1131
1132# The format of the values of the map tables:
1133my $BINARY_FORMAT = 'b';
1134my $DECIMAL_FORMAT = 'd';
1135my $FLOAT_FORMAT = 'f';
1136my $INTEGER_FORMAT = 'i';
1137my $HEX_FORMAT = 'x';
1138my $RATIONAL_FORMAT = 'r';
1139my $STRING_FORMAT = 's';
1140
1141my %map_table_formats = (
1142 $BINARY_FORMAT => 'binary',
1143 $DECIMAL_FORMAT => 'single decimal digit',
1144 $FLOAT_FORMAT => 'floating point number',
1145 $INTEGER_FORMAT => 'integer',
1146 $HEX_FORMAT => 'positive hex whole number; a code point',
1147 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1148 $STRING_FORMAT => 'arbitrary string',
1149);
1150
1151# Unicode didn't put such derived files in a separate directory at first.
1152my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1153my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1154my $AUXILIARY = 'auxiliary';
1155
1156# Hashes that will eventually go into Heavy.pl for the use of utf8_heavy.pl
1157my %loose_to_file_of; # loosely maps table names to their respective
1158 # files
1159my %stricter_to_file_of; # same; but for stricter mapping.
1160my %nv_floating_to_rational; # maps numeric values floating point numbers to
1161 # their rational equivalent
1162my %loose_property_name_of; # Loosely maps property names to standard form
1163
1164# These constants names and values were taken from the Unicode standard,
1165# version 5.1, section 3.12. They are used in conjunction with Hangul
1166# syllables
1167my $SBase = 0xAC00;
1168my $LBase = 0x1100;
1169my $VBase = 0x1161;
1170my $TBase = 0x11A7;
1171my $SCount = 11172;
1172my $LCount = 19;
1173my $VCount = 21;
1174my $TCount = 28;
1175my $NCount = $VCount * $TCount;
1176
1177# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1178# with the above published constants.
1179my %Jamo;
1180my %Jamo_L; # Leading consonants
1181my %Jamo_V; # Vowels
1182my %Jamo_T; # Trailing consonants
1183
37e2e78e 1184my @backslash_X_tests; # List of tests read in for testing \X
99870f4d
KW
1185my @unhandled_properties; # Will contain a list of properties found in
1186 # the input that we didn't process.
f86864ac 1187my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1188 # listed in the pod
1189my @map_properties; # Properties that get map files written
1190my @named_sequences; # NamedSequences.txt contents.
1191my %potential_files; # Generated list of all .txt files in the directory
1192 # structure so we can warn if something is being
1193 # ignored.
1194my @files_actually_output; # List of files we generated.
1195my @more_Names; # Some code point names are compound; this is used
1196 # to store the extra components of them.
1197my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1198 # the minimum before we consider it equivalent to a
1199 # candidate rational
1200my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1201
1202# These store references to certain commonly used property objects
1203my $gc;
1204my $perl;
1205my $block;
1206
1207# Are there conflicting names because of beginning with 'In_', or 'Is_'
1208my $has_In_conflicts = 0;
1209my $has_Is_conflicts = 0;
1210
1211sub internal_file_to_platform ($) {
1212 # Convert our file paths which have '/' separators to those of the
1213 # platform.
1214
1215 my $file = shift;
1216 return undef unless defined $file;
1217
1218 return File::Spec->join(split '/', $file);
d07a55ed 1219}
5beb625e 1220
99870f4d
KW
1221sub file_exists ($) { # platform independent '-e'. This program internally
1222 # uses slash as a path separator.
1223 my $file = shift;
1224 return 0 if ! defined $file;
1225 return -e internal_file_to_platform($file);
1226}
5beb625e 1227
99870f4d 1228sub objaddr($) {
23e33b60
KW
1229 # Returns the address of the blessed input object.
1230 # It doesn't check for blessedness because that would do a string eval
1231 # every call, and the program is structured so that this is never called
1232 # for a non-blessed object.
99870f4d 1233
23e33b60 1234 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1235
1236 # Numifying a ref gives its address.
23e33b60 1237 return 0 + $_[0];
99870f4d
KW
1238}
1239
23e33b60
KW
1240# Commented code below should work on Perl 5.8.
1241## This 'require' doesn't necessarily work in miniperl, and even if it does,
1242## the native perl version of it (which is what would operate under miniperl)
1243## is extremely slow, as it does a string eval every call.
1244#my $has_fast_scalar_util = $\18 !~ /miniperl/
1245# && defined eval "require Scalar::Util";
1246#
1247#sub objaddr($) {
1248# # Returns the address of the blessed input object. Uses the XS version if
1249# # available. It doesn't check for blessedness because that would do a
1250# # string eval every call, and the program is structured so that this is
1251# # never called for a non-blessed object.
1252#
1253# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1254#
1255# # Check at least that is a ref.
1256# my $pkg = ref($_[0]) or return undef;
1257#
1258# # Change to a fake package to defeat any overloaded stringify
1259# bless $_[0], 'main::Fake';
1260#
1261# # Numifying a ref gives its address.
1262# my $addr = 0 + $_[0];
1263#
1264# # Return to original class
1265# bless $_[0], $pkg;
1266# return $addr;
1267#}
1268
99870f4d
KW
1269sub max ($$) {
1270 my $a = shift;
1271 my $b = shift;
1272 return $a if $a >= $b;
1273 return $b;
1274}
1275
1276sub min ($$) {
1277 my $a = shift;
1278 my $b = shift;
1279 return $a if $a <= $b;
1280 return $b;
1281}
1282
1283sub clarify_number ($) {
1284 # This returns the input number with underscores inserted every 3 digits
1285 # in large (5 digits or more) numbers. Input must be entirely digits, not
1286 # checked.
1287
1288 my $number = shift;
1289 my $pos = length($number) - 3;
1290 return $number if $pos <= 1;
1291 while ($pos > 0) {
1292 substr($number, $pos, 0) = '_';
1293 $pos -= 3;
5beb625e 1294 }
99870f4d 1295 return $number;
99598c8c
JH
1296}
1297
12ac2576 1298
99870f4d 1299package Carp;
7ebf06b3 1300
99870f4d
KW
1301# These routines give a uniform treatment of messages in this program. They
1302# are placed in the Carp package to cause the stack trace to not include them,
1303# although an alternative would be to use another package and set @CARP_NOT
1304# for it.
12ac2576 1305
99870f4d 1306our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1307
99f78760
KW
1308# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1309# and overload trying to load Scalar:Util under miniperl. See
1310# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1311undef $overload::VERSION;
1312
99870f4d
KW
1313sub my_carp {
1314 my $message = shift || "";
1315 my $nofold = shift || 0;
7ebf06b3 1316
99870f4d
KW
1317 if ($message) {
1318 $message = main::join_lines($message);
1319 $message =~ s/^$0: *//; # Remove initial program name
1320 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1321 $message = "\n$0: $message;";
12ac2576 1322
99870f4d
KW
1323 # Fold the message with program name, semi-colon end punctuation
1324 # (which looks good with the message that carp appends to it), and a
1325 # hanging indent for continuation lines.
1326 $message = main::simple_fold($message, "", 4) unless $nofold;
1327 $message =~ s/\n$//; # Remove the trailing nl so what carp
1328 # appends is to the same line
1329 }
12ac2576 1330
99870f4d 1331 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1332
99870f4d
KW
1333 carp $message;
1334 return;
1335}
7ebf06b3 1336
99870f4d
KW
1337sub my_carp_bug {
1338 # This is called when it is clear that the problem is caused by a bug in
1339 # this program.
7ebf06b3 1340
99870f4d
KW
1341 my $message = shift;
1342 $message =~ s/^$0: *//;
1343 $message = my_carp("Bug in $0. Please report it by running perlbug or if that is unavailable, by sending email to perbug\@perl.org:\n$message");
1344 carp $message;
1345 return;
1346}
7ebf06b3 1347
99870f4d
KW
1348sub carp_too_few_args {
1349 if (@_ != 2) {
1350 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1351 return;
12ac2576 1352 }
7ebf06b3 1353
99870f4d
KW
1354 my $args_ref = shift;
1355 my $count = shift;
7ebf06b3 1356
99870f4d
KW
1357 my_carp_bug("Need at least $count arguments to "
1358 . (caller 1)[3]
1359 . ". Instead got: '"
1360 . join ', ', @$args_ref
1361 . "'. No action taken.");
1362 return;
12ac2576
JP
1363}
1364
99870f4d
KW
1365sub carp_extra_args {
1366 my $args_ref = shift;
1367 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1368
99870f4d
KW
1369 unless (ref $args_ref) {
1370 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1371 return;
1372 }
1373 my ($package, $file, $line) = caller;
1374 my $subroutine = (caller 1)[3];
cf25bb62 1375
99870f4d
KW
1376 my $list;
1377 if (ref $args_ref eq 'HASH') {
1378 foreach my $key (keys %$args_ref) {
1379 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1380 }
99870f4d 1381 $list = join ', ', each %{$args_ref};
cf25bb62 1382 }
99870f4d
KW
1383 elsif (ref $args_ref eq 'ARRAY') {
1384 foreach my $arg (@$args_ref) {
1385 $arg = $UNDEF unless defined $arg;
1386 }
1387 $list = join ', ', @$args_ref;
1388 }
1389 else {
1390 my_carp_bug("Can't cope with ref "
1391 . ref($args_ref)
1392 . " . argument to 'carp_extra_args'. Not checking arguments.");
1393 return;
1394 }
1395
1396 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1397 return;
d73e5302
JH
1398}
1399
99870f4d
KW
1400package main;
1401
1402{ # Closure
1403
1404 # This program uses the inside-out method for objects, as recommended in
1405 # "Perl Best Practices". This closure aids in generating those. There
1406 # are two routines. setup_package() is called once per package to set
1407 # things up, and then set_access() is called for each hash representing a
1408 # field in the object. These routines arrange for the object to be
1409 # properly destroyed when no longer used, and for standard accessor
1410 # functions to be generated. If you need more complex accessors, just
1411 # write your own and leave those accesses out of the call to set_access().
1412 # More details below.
1413
1414 my %constructor_fields; # fields that are to be used in constructors; see
1415 # below
1416
1417 # The values of this hash will be the package names as keys to other
1418 # hashes containing the name of each field in the package as keys, and
1419 # references to their respective hashes as values.
1420 my %package_fields;
1421
1422 sub setup_package {
1423 # Sets up the package, creating standard DESTROY and dump methods
1424 # (unless already defined). The dump method is used in debugging by
1425 # simple_dumper().
1426 # The optional parameters are:
1427 # a) a reference to a hash, that gets populated by later
1428 # set_access() calls with one of the accesses being
1429 # 'constructor'. The caller can then refer to this, but it is
1430 # not otherwise used by these two routines.
1431 # b) a reference to a callback routine to call during destruction
1432 # of the object, before any fields are actually destroyed
1433
1434 my %args = @_;
1435 my $constructor_ref = delete $args{'Constructor_Fields'};
1436 my $destroy_callback = delete $args{'Destroy_Callback'};
1437 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1438
1439 my %fields;
1440 my $package = (caller)[0];
1441
1442 $package_fields{$package} = \%fields;
1443 $constructor_fields{$package} = $constructor_ref;
1444
1445 unless ($package->can('DESTROY')) {
1446 my $destroy_name = "${package}::DESTROY";
1447 no strict "refs";
1448
1449 # Use typeglob to give the anonymous subroutine the name we want
1450 *$destroy_name = sub {
1451 my $self = shift;
f998e60c 1452 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
1453
1454 $self->$destroy_callback if $destroy_callback;
1455 foreach my $field (keys %{$package_fields{$package}}) {
1456 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1457 delete $package_fields{$package}{$field}{$addr};
1458 }
1459 return;
1460 }
1461 }
1462
1463 unless ($package->can('dump')) {
1464 my $dump_name = "${package}::dump";
1465 no strict "refs";
1466 *$dump_name = sub {
1467 my $self = shift;
1468 return dump_inside_out($self, $package_fields{$package}, @_);
1469 }
1470 }
1471 return;
1472 }
1473
1474 sub set_access {
1475 # Arrange for the input field to be garbage collected when no longer
1476 # needed. Also, creates standard accessor functions for the field
1477 # based on the optional parameters-- none if none of these parameters:
1478 # 'addable' creates an 'add_NAME()' accessor function.
1479 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1480 # function.
1481 # 'settable' creates a 'set_NAME()' accessor function.
1482 # 'constructor' doesn't create an accessor function, but adds the
1483 # field to the hash that was previously passed to
1484 # setup_package();
1485 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1486 # 'add' etc. all mean 'addable'.
1487 # The read accessor function will work on both array and scalar
1488 # values. If another accessor in the parameter list is 'a', the read
1489 # access assumes an array. You can also force it to be array access
1490 # by specifying 'readable_array' instead of 'readable'
1491 #
1492 # A sort-of 'protected' access can be set-up by preceding the addable,
1493 # readable or settable with some initial portion of 'protected_' (but,
1494 # the underscore is required), like 'p_a', 'pro_set', etc. The
1495 # "protection" is only by convention. All that happens is that the
1496 # accessor functions' names begin with an underscore. So instead of
1497 # calling set_foo, the call is _set_foo. (Real protection could be
1498 # accomplished by having a new subroutine, end_package called at the
1499 # end of each package, and then storing the __LINE__ ranges and
1500 # checking them on every accessor. But that is way overkill.)
1501
1502 # We create anonymous subroutines as the accessors and then use
1503 # typeglobs to assign them to the proper package and name
1504
1505 my $name = shift; # Name of the field
1506 my $field = shift; # Reference to the inside-out hash containing the
1507 # field
1508
1509 my $package = (caller)[0];
1510
1511 if (! exists $package_fields{$package}) {
1512 croak "$0: Must call 'setup_package' before 'set_access'";
1513 }
d73e5302 1514
99870f4d
KW
1515 # Stash the field so DESTROY can get it.
1516 $package_fields{$package}{$name} = $field;
cf25bb62 1517
99870f4d
KW
1518 # Remaining arguments are the accessors. For each...
1519 foreach my $access (@_) {
1520 my $access = lc $access;
cf25bb62 1521
99870f4d 1522 my $protected = "";
cf25bb62 1523
99870f4d
KW
1524 # Match the input as far as it goes.
1525 if ($access =~ /^(p[^_]*)_/) {
1526 $protected = $1;
1527 if (substr('protected_', 0, length $protected)
1528 eq $protected)
1529 {
1530
1531 # Add 1 for the underscore not included in $protected
1532 $access = substr($access, length($protected) + 1);
1533 $protected = '_';
1534 }
1535 else {
1536 $protected = "";
1537 }
1538 }
1539
1540 if (substr('addable', 0, length $access) eq $access) {
1541 my $subname = "${package}::${protected}add_$name";
1542 no strict "refs";
1543
1544 # add_ accessor. Don't add if already there, which we
1545 # determine using 'eq' for scalars and '==' otherwise.
1546 *$subname = sub {
1547 use strict "refs";
1548 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1549 my $self = shift;
1550 my $value = shift;
f998e60c 1551 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
1552 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1553 if (ref $value) {
f998e60c 1554 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1555 }
1556 else {
f998e60c 1557 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1558 }
f998e60c 1559 push @{$field->{$addr}}, $value;
99870f4d
KW
1560 return;
1561 }
1562 }
1563 elsif (substr('constructor', 0, length $access) eq $access) {
1564 if ($protected) {
1565 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1566 }
1567 else {
1568 $constructor_fields{$package}{$name} = $field;
1569 }
1570 }
1571 elsif (substr('readable_array', 0, length $access) eq $access) {
1572
1573 # Here has read access. If one of the other parameters for
1574 # access is array, or this one specifies array (by being more
1575 # than just 'readable_'), then create a subroutine that
1576 # assumes the data is an array. Otherwise just a scalar
1577 my $subname = "${package}::${protected}$name";
1578 if (grep { /^a/i } @_
1579 or length($access) > length('readable_'))
1580 {
1581 no strict "refs";
1582 *$subname = sub {
1583 use strict "refs";
23e33b60 1584 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 1585 my $addr; { no overloading; $addr = 0+$_[0]; }
99870f4d
KW
1586 if (ref $field->{$addr} ne 'ARRAY') {
1587 my $type = ref $field->{$addr};
1588 $type = 'scalar' unless $type;
1589 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
1590 return;
1591 }
1592 return scalar @{$field->{$addr}} unless wantarray;
1593
1594 # Make a copy; had problems with caller modifying the
1595 # original otherwise
1596 my @return = @{$field->{$addr}};
1597 return @return;
1598 }
1599 }
1600 else {
1601
1602 # Here not an array value, a simpler function.
1603 no strict "refs";
1604 *$subname = sub {
1605 use strict "refs";
23e33b60 1606 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c
KW
1607 no overloading;
1608 return $field->{0+$_[0]};
99870f4d
KW
1609 }
1610 }
1611 }
1612 elsif (substr('settable', 0, length $access) eq $access) {
1613 my $subname = "${package}::${protected}set_$name";
1614 no strict "refs";
1615 *$subname = sub {
1616 use strict "refs";
23e33b60
KW
1617 if (main::DEBUG) {
1618 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
1619 Carp::carp_extra_args(\@_) if @_ > 2;
1620 }
1621 # $self is $_[0]; $value is $_[1]
f998e60c
KW
1622 no overloading;
1623 $field->{0+$_[0]} = $_[1];
99870f4d
KW
1624 return;
1625 }
1626 }
1627 else {
1628 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
1629 }
cf25bb62 1630 }
99870f4d 1631 return;
cf25bb62 1632 }
99870f4d
KW
1633}
1634
1635package Input_file;
1636
1637# All input files use this object, which stores various attributes about them,
1638# and provides for convenient, uniform handling. The run method wraps the
1639# processing. It handles all the bookkeeping of opening, reading, and closing
1640# the file, returning only significant input lines.
1641#
1642# Each object gets a handler which processes the body of the file, and is
1643# called by run(). Most should use the generic, default handler, which has
1644# code scrubbed to handle things you might not expect. A handler should
1645# basically be a while(next_line()) {...} loop.
1646#
1647# You can also set up handlers to
1648# 1) call before the first line is read for pre processing
1649# 2) call to adjust each line of the input before the main handler gets them
1650# 3) call upon EOF before the main handler exits its loop
1651# 4) call at the end for post processing
1652#
1653# $_ is used to store the input line, and is to be filtered by the
1654# each_line_handler()s. So, if the format of the line is not in the desired
1655# format for the main handler, these are used to do that adjusting. They can
1656# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
1657# so the $_ output of one is used as the input to the next. None of the other
1658# handlers are stackable, but could easily be changed to be so.
1659#
1660# Most of the handlers can call insert_lines() or insert_adjusted_lines()
1661# which insert the parameters as lines to be processed before the next input
1662# file line is read. This allows the EOF handler to flush buffers, for
1663# example. The difference between the two routines is that the lines inserted
1664# by insert_lines() are subjected to the each_line_handler()s. (So if you
1665# called it from such a handler, you would get infinite recursion.) Lines
1666# inserted by insert_adjusted_lines() go directly to the main handler without
1667# any adjustments. If the post-processing handler calls any of these, there
1668# will be no effect. Some error checking for these conditions could be added,
1669# but it hasn't been done.
1670#
1671# carp_bad_line() should be called to warn of bad input lines, which clears $_
1672# to prevent further processing of the line. This routine will output the
1673# message as a warning once, and then keep a count of the lines that have the
1674# same message, and output that count at the end of the file's processing.
1675# This keeps the number of messages down to a manageable amount.
1676#
1677# get_missings() should be called to retrieve any @missing input lines.
1678# Messages will be raised if this isn't done if the options aren't to ignore
1679# missings.
1680
1681sub trace { return main::trace(@_); }
1682
99870f4d
KW
1683{ # Closure
1684 # Keep track of fields that are to be put into the constructor.
1685 my %constructor_fields;
1686
1687 main::setup_package(Constructor_Fields => \%constructor_fields);
1688
1689 my %file; # Input file name, required
1690 main::set_access('file', \%file, qw{ c r });
1691
1692 my %first_released; # Unicode version file was first released in, required
1693 main::set_access('first_released', \%first_released, qw{ c r });
1694
1695 my %handler; # Subroutine to process the input file, defaults to
1696 # 'process_generic_property_file'
1697 main::set_access('handler', \%handler, qw{ c });
1698
1699 my %property;
1700 # name of property this file is for. defaults to none, meaning not
1701 # applicable, or is otherwise determinable, for example, from each line.
1702 main::set_access('property', \%property, qw{ c });
1703
1704 my %optional;
1705 # If this is true, the file is optional. If not present, no warning is
1706 # output. If it is present, the string given by this parameter is
1707 # evaluated, and if false the file is not processed.
1708 main::set_access('optional', \%optional, 'c', 'r');
1709
1710 my %non_skip;
1711 # This is used for debugging, to skip processing of all but a few input
1712 # files. Add 'non_skip => 1' to the constructor for those files you want
1713 # processed when you set the $debug_skip global.
1714 main::set_access('non_skip', \%non_skip, 'c');
1715
37e2e78e
KW
1716 my %skip;
1717 # This is used to skip processing of this input file semi-permanently.
1718 # It is used for files that we aren't planning to process anytime soon,
1719 # but want to allow to be in the directory and not raise a message that we
1720 # are not handling. Mostly for test files. This is in contrast to the
1721 # non_skip element, which is supposed to be used very temporarily for
1722 # debugging. Sets 'optional' to 1
1723 main::set_access('skip', \%skip, 'c');
1724
99870f4d
KW
1725 my %each_line_handler;
1726 # list of subroutines to look at and filter each non-comment line in the
1727 # file. defaults to none. The subroutines are called in order, each is
1728 # to adjust $_ for the next one, and the final one adjusts it for
1729 # 'handler'
1730 main::set_access('each_line_handler', \%each_line_handler, 'c');
1731
1732 my %has_missings_defaults;
1733 # ? Are there lines in the file giving default values for code points
1734 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
1735 # the norm, but IGNORED means it has such lines, but the handler doesn't
1736 # use them. Having these three states allows us to catch changes to the
1737 # UCD that this program should track
1738 main::set_access('has_missings_defaults',
1739 \%has_missings_defaults, qw{ c r });
1740
1741 my %pre_handler;
1742 # Subroutine to call before doing anything else in the file. If undef, no
1743 # such handler is called.
1744 main::set_access('pre_handler', \%pre_handler, qw{ c });
1745
1746 my %eof_handler;
1747 # Subroutine to call upon getting an EOF on the input file, but before
1748 # that is returned to the main handler. This is to allow buffers to be
1749 # flushed. The handler is expected to call insert_lines() or
1750 # insert_adjusted() with the buffered material
1751 main::set_access('eof_handler', \%eof_handler, qw{ c r });
1752
1753 my %post_handler;
1754 # Subroutine to call after all the lines of the file are read in and
1755 # processed. If undef, no such handler is called.
1756 main::set_access('post_handler', \%post_handler, qw{ c });
1757
1758 my %progress_message;
1759 # Message to print to display progress in lieu of the standard one
1760 main::set_access('progress_message', \%progress_message, qw{ c });
1761
1762 my %handle;
1763 # cache open file handle, internal. Is undef if file hasn't been
1764 # processed at all, empty if has;
1765 main::set_access('handle', \%handle);
1766
1767 my %added_lines;
1768 # cache of lines added virtually to the file, internal
1769 main::set_access('added_lines', \%added_lines);
1770
1771 my %errors;
1772 # cache of errors found, internal
1773 main::set_access('errors', \%errors);
1774
1775 my %missings;
1776 # storage of '@missing' defaults lines
1777 main::set_access('missings', \%missings);
1778
1779 sub new {
1780 my $class = shift;
1781
1782 my $self = bless \do{ my $anonymous_scalar }, $class;
f998e60c 1783 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
1784
1785 # Set defaults
1786 $handler{$addr} = \&main::process_generic_property_file;
1787 $non_skip{$addr} = 0;
37e2e78e 1788 $skip{$addr} = 0;
99870f4d
KW
1789 $has_missings_defaults{$addr} = $NO_DEFAULTS;
1790 $handle{$addr} = undef;
1791 $added_lines{$addr} = [ ];
1792 $each_line_handler{$addr} = [ ];
1793 $errors{$addr} = { };
1794 $missings{$addr} = [ ];
1795
1796 # Two positional parameters.
99f78760 1797 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
1798 $file{$addr} = main::internal_file_to_platform(shift);
1799 $first_released{$addr} = shift;
1800
1801 # The rest of the arguments are key => value pairs
1802 # %constructor_fields has been set up earlier to list all possible
1803 # ones. Either set or push, depending on how the default has been set
1804 # up just above.
1805 my %args = @_;
1806 foreach my $key (keys %args) {
1807 my $argument = $args{$key};
1808
1809 # Note that the fields are the lower case of the constructor keys
1810 my $hash = $constructor_fields{lc $key};
1811 if (! defined $hash) {
1812 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
1813 next;
1814 }
1815 if (ref $hash->{$addr} eq 'ARRAY') {
1816 if (ref $argument eq 'ARRAY') {
1817 foreach my $argument (@{$argument}) {
1818 next if ! defined $argument;
1819 push @{$hash->{$addr}}, $argument;
1820 }
1821 }
1822 else {
1823 push @{$hash->{$addr}}, $argument if defined $argument;
1824 }
1825 }
1826 else {
1827 $hash->{$addr} = $argument;
1828 }
1829 delete $args{$key};
1830 };
1831
1832 # If the file has a property for it, it means that the property is not
1833 # listed in the file's entries. So add a handler to the list of line
1834 # handlers to insert the property name into the lines, to provide a
1835 # uniform interface to the final processing subroutine.
1836 # the final code doesn't have to worry about that.
1837 if ($property{$addr}) {
1838 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
1839 }
1840
1841 if ($non_skip{$addr} && ! $debug_skip && $verbosity) {
1842 print "Warning: " . __PACKAGE__ . " constructor for $file{$addr} has useless 'non_skip' in it\n";
a3a8c5f0 1843 }
99870f4d 1844
37e2e78e
KW
1845 $optional{$addr} = 1 if $skip{$addr};
1846
99870f4d 1847 return $self;
d73e5302
JH
1848 }
1849
cf25bb62 1850
99870f4d
KW
1851 use overload
1852 fallback => 0,
1853 qw("") => "_operator_stringify",
1854 "." => \&main::_operator_dot,
1855 ;
cf25bb62 1856
99870f4d
KW
1857 sub _operator_stringify {
1858 my $self = shift;
cf25bb62 1859
99870f4d 1860 return __PACKAGE__ . " object for " . $self->file;
d73e5302 1861 }
d73e5302 1862
99870f4d
KW
1863 # flag to make sure extracted files are processed early
1864 my $seen_non_extracted_non_age = 0;
d73e5302 1865
99870f4d
KW
1866 sub run {
1867 # Process the input object $self. This opens and closes the file and
1868 # calls all the handlers for it. Currently, this can only be called
1869 # once per file, as it destroy's the EOF handler
d73e5302 1870
99870f4d
KW
1871 my $self = shift;
1872 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 1873
f998e60c 1874 my $addr; { no overloading; $addr = 0+$self; }
b6922eda 1875
99870f4d 1876 my $file = $file{$addr};
d73e5302 1877
99870f4d
KW
1878 # Don't process if not expecting this file (because released later
1879 # than this Unicode version), and isn't there. This means if someone
1880 # copies it into an earlier version's directory, we will go ahead and
1881 # process it.
1882 return if $first_released{$addr} gt $v_version && ! -e $file;
1883
1884 # If in debugging mode and this file doesn't have the non-skip
1885 # flag set, and isn't one of the critical files, skip it.
1886 if ($debug_skip
1887 && $first_released{$addr} ne v0
1888 && ! $non_skip{$addr})
1889 {
1890 print "Skipping $file in debugging\n" if $verbosity;
1891 return;
1892 }
1893
1894 # File could be optional
37e2e78e 1895 if ($optional{$addr}) {
99870f4d
KW
1896 return unless -e $file;
1897 my $result = eval $optional{$addr};
1898 if (! defined $result) {
1899 Carp::my_carp_bug("Got '$@' when tried to eval $optional{$addr}. $file Skipped.");
1900 return;
1901 }
1902 if (! $result) {
1903 if ($verbosity) {
1904 print STDERR "Skipping processing input file '$file' because '$optional{$addr}' is not true\n";
1905 }
1906 return;
1907 }
1908 }
1909
1910 if (! defined $file || ! -e $file) {
1911
1912 # If the file doesn't exist, see if have internal data for it
1913 # (based on first_released being 0).
1914 if ($first_released{$addr} eq v0) {
1915 $handle{$addr} = 'pretend_is_open';
1916 }
1917 else {
1918 if (! $optional{$addr} # File could be optional
1919 && $v_version ge $first_released{$addr})
1920 {
1921 print STDERR "Skipping processing input file '$file' because not found\n" if $v_version ge $first_released{$addr};
1922 }
1923 return;
1924 }
1925 }
1926 else {
1927
37e2e78e
KW
1928 # Here, the file exists. Some platforms may change the case of
1929 # its name
99870f4d 1930 if ($seen_non_extracted_non_age) {
517956bf 1931 if ($file =~ /$EXTRACTED/i) {
99870f4d 1932 Carp::my_carp_bug(join_lines(<<END
99f78760 1933$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
1934anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
1935have subtle problems
1936END
1937 ));
1938 }
1939 }
1940 elsif ($EXTRACTED_DIR
1941 && $first_released{$addr} ne v0
517956bf
CB
1942 && $file !~ /$EXTRACTED/i
1943 && lc($file) ne 'dage.txt')
99870f4d
KW
1944 {
1945 # We don't set this (by the 'if' above) if we have no
1946 # extracted directory, so if running on an early version,
1947 # this test won't work. Not worth worrying about.
1948 $seen_non_extracted_non_age = 1;
1949 }
1950
1951 # And mark the file as having being processed, and warn if it
1952 # isn't a file we are expecting. As we process the files,
1953 # they are deleted from the hash, so any that remain at the
1954 # end of the program are files that we didn't process.
517956bf
CB
1955 my $fkey = File::Spec->rel2abs($file);
1956 my $expecting = delete $potential_files{$fkey};
1957 $expecting = delete $potential_files{lc($fkey)} unless defined $expecting;
678f13d5
KW
1958 Carp::my_carp("Was not expecting '$file'.") if
1959 ! $expecting
99870f4d
KW
1960 && ! defined $handle{$addr};
1961
37e2e78e
KW
1962 # Having deleted from expected files, we can quit if not to do
1963 # anything. Don't print progress unless really want verbosity
1964 if ($skip{$addr}) {
1965 print "Skipping $file.\n" if $verbosity >= $VERBOSE;
1966 return;
1967 }
1968
99870f4d
KW
1969 # Open the file, converting the slashes used in this program
1970 # into the proper form for the OS
1971 my $file_handle;
1972 if (not open $file_handle, "<", $file) {
1973 Carp::my_carp("Can't open $file. Skipping: $!");
1974 return 0;
1975 }
1976 $handle{$addr} = $file_handle; # Cache the open file handle
1977 }
1978
1979 if ($verbosity >= $PROGRESS) {
1980 if ($progress_message{$addr}) {
1981 print "$progress_message{$addr}\n";
1982 }
1983 else {
1984 # If using a virtual file, say so.
1985 print "Processing ", (-e $file)
1986 ? $file
1987 : "substitute $file",
1988 "\n";
1989 }
1990 }
1991
1992
1993 # Call any special handler for before the file.
1994 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
1995
1996 # Then the main handler
1997 &{$handler{$addr}}($self);
1998
1999 # Then any special post-file handler.
2000 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2001
2002 # If any errors have been accumulated, output the counts (as the first
2003 # error message in each class was output when it was encountered).
2004 if ($errors{$addr}) {
2005 my $total = 0;
2006 my $types = 0;
2007 foreach my $error (keys %{$errors{$addr}}) {
2008 $total += $errors{$addr}->{$error};
2009 delete $errors{$addr}->{$error};
2010 $types++;
2011 }
2012 if ($total > 1) {
2013 my $message
2014 = "A total of $total lines had errors in $file. ";
2015
2016 $message .= ($types == 1)
2017 ? '(Only the first one was displayed.)'
2018 : '(Only the first of each type was displayed.)';
2019 Carp::my_carp($message);
2020 }
2021 }
2022
2023 if (@{$missings{$addr}}) {
2024 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2025 }
2026
2027 # If a real file handle, close it.
2028 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2029 ref $handle{$addr};
2030 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2031 # the file, as opposed to undef
2032 return;
2033 }
2034
2035 sub next_line {
2036 # Sets $_ to be the next logical input line, if any. Returns non-zero
2037 # if such a line exists. 'logical' means that any lines that have
2038 # been added via insert_lines() will be returned in $_ before the file
2039 # is read again.
2040
2041 my $self = shift;
2042 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2043
f998e60c 2044 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2045
2046 # Here the file is open (or if the handle is not a ref, is an open
2047 # 'virtual' file). Get the next line; any inserted lines get priority
2048 # over the file itself.
2049 my $adjusted;
2050
2051 LINE:
2052 while (1) { # Loop until find non-comment, non-empty line
2053 #local $to_trace = 1 if main::DEBUG;
2054 my $inserted_ref = shift @{$added_lines{$addr}};
2055 if (defined $inserted_ref) {
2056 ($adjusted, $_) = @{$inserted_ref};
2057 trace $adjusted, $_ if main::DEBUG && $to_trace;
2058 return 1 if $adjusted;
2059 }
2060 else {
2061 last if ! ref $handle{$addr}; # Don't read unless is real file
2062 last if ! defined ($_ = readline $handle{$addr});
2063 }
2064 chomp;
2065 trace $_ if main::DEBUG && $to_trace;
2066
2067 # See if this line is the comment line that defines what property
2068 # value that code points that are not listed in the file should
2069 # have. The format or existence of these lines is not guaranteed
2070 # by Unicode since they are comments, but the documentation says
2071 # that this was added for machine-readability, so probably won't
2072 # change. This works starting in Unicode Version 5.0. They look
2073 # like:
2074 #
2075 # @missing: 0000..10FFFF; Not_Reordered
2076 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2077 # @missing: 0000..10FFFF; ; NaN
2078 #
2079 # Save the line for a later get_missings() call.
2080 if (/$missing_defaults_prefix/) {
2081 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2082 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2083 }
2084 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2085 my @defaults = split /\s* ; \s*/x, $_;
2086
2087 # The first field is the @missing, which ends in a
2088 # semi-colon, so can safely shift.
2089 shift @defaults;
2090
2091 # Some of these lines may have empty field placeholders
2092 # which get in the way. An example is:
2093 # @missing: 0000..10FFFF; ; NaN
2094 # Remove them. Process starting from the top so the
2095 # splice doesn't affect things still to be looked at.
2096 for (my $i = @defaults - 1; $i >= 0; $i--) {
2097 next if $defaults[$i] ne "";
2098 splice @defaults, $i, 1;
2099 }
2100
2101 # What's left should be just the property (maybe) and the
2102 # default. Having only one element means it doesn't have
2103 # the property.
2104 my $default;
2105 my $property;
2106 if (@defaults >= 1) {
2107 if (@defaults == 1) {
2108 $default = $defaults[0];
2109 }
2110 else {
2111 $property = $defaults[0];
2112 $default = $defaults[1];
2113 }
2114 }
2115
2116 if (@defaults < 1
2117 || @defaults > 2
2118 || ($default =~ /^</
2119 && $default !~ /^<code *point>$/i
2120 && $default !~ /^<none>$/i))
2121 {
2122 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2123 }
2124 else {
2125
2126 # If the property is missing from the line, it should
2127 # be the one for the whole file
2128 $property = $property{$addr} if ! defined $property;
2129
2130 # Change <none> to the null string, which is what it
2131 # really means. If the default is the code point
2132 # itself, set it to <code point>, which is what
2133 # Unicode uses (but sometimes they've forgotten the
2134 # space)
2135 if ($default =~ /^<none>$/i) {
2136 $default = "";
2137 }
2138 elsif ($default =~ /^<code *point>$/i) {
2139 $default = $CODE_POINT;
2140 }
2141
2142 # Store them as a sub-arrays with both components.
2143 push @{$missings{$addr}}, [ $default, $property ];
2144 }
2145 }
2146
2147 # There is nothing for the caller to process on this comment
2148 # line.
2149 next;
2150 }
2151
2152 # Remove comments and trailing space, and skip this line if the
2153 # result is empty
2154 s/#.*//;
2155 s/\s+$//;
2156 next if /^$/;
2157
2158 # Call any handlers for this line, and skip further processing of
2159 # the line if the handler sets the line to null.
2160 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2161 &{$sub_ref}($self);
2162 next LINE if /^$/;
2163 }
2164
2165 # Here the line is ok. return success.
2166 return 1;
2167 } # End of looping through lines.
2168
2169 # If there is an EOF handler, call it (only once) and if it generates
2170 # more lines to process go back in the loop to handle them.
2171 if ($eof_handler{$addr}) {
2172 &{$eof_handler{$addr}}($self);
2173 $eof_handler{$addr} = ""; # Currently only get one shot at it.
2174 goto LINE if $added_lines{$addr};
2175 }
2176
2177 # Return failure -- no more lines.
2178 return 0;
2179
2180 }
2181
2182# Not currently used, not fully tested.
2183# sub peek {
2184# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
2185# # record. Not callable from an each_line_handler(), nor does it call
2186# # an each_line_handler() on the line.
2187#
2188# my $self = shift;
f998e60c 2189# my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2190#
2191# foreach my $inserted_ref (@{$added_lines{$addr}}) {
2192# my ($adjusted, $line) = @{$inserted_ref};
2193# next if $adjusted;
2194#
2195# # Remove comments and trailing space, and return a non-empty
2196# # resulting line
2197# $line =~ s/#.*//;
2198# $line =~ s/\s+$//;
2199# return $line if $line ne "";
2200# }
2201#
2202# return if ! ref $handle{$addr}; # Don't read unless is real file
2203# while (1) { # Loop until find non-comment, non-empty line
2204# local $to_trace = 1 if main::DEBUG;
2205# trace $_ if main::DEBUG && $to_trace;
2206# return if ! defined (my $line = readline $handle{$addr});
2207# chomp $line;
2208# push @{$added_lines{$addr}}, [ 0, $line ];
2209#
2210# $line =~ s/#.*//;
2211# $line =~ s/\s+$//;
2212# return $line if $line ne "";
2213# }
2214#
2215# return;
2216# }
2217
2218
2219 sub insert_lines {
2220 # Lines can be inserted so that it looks like they were in the input
2221 # file at the place it was when this routine is called. See also
2222 # insert_adjusted_lines(). Lines inserted via this routine go through
2223 # any each_line_handler()
2224
2225 my $self = shift;
2226
2227 # Each inserted line is an array, with the first element being 0 to
2228 # indicate that this line hasn't been adjusted, and needs to be
2229 # processed.
f998e60c
KW
2230 no overloading;
2231 push @{$added_lines{0+$self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
2232 return;
2233 }
2234
2235 sub insert_adjusted_lines {
2236 # Lines can be inserted so that it looks like they were in the input
2237 # file at the place it was when this routine is called. See also
2238 # insert_lines(). Lines inserted via this routine are already fully
2239 # adjusted, ready to be processed; each_line_handler()s handlers will
2240 # not be called. This means this is not a completely general
2241 # facility, as only the last each_line_handler on the stack should
2242 # call this. It could be made more general, by passing to each of the
2243 # line_handlers their position on the stack, which they would pass on
2244 # to this routine, and that would replace the boolean first element in
2245 # the anonymous array pushed here, so that the next_line routine could
2246 # use that to call only those handlers whose index is after it on the
2247 # stack. But this is overkill for what is needed now.
2248
2249 my $self = shift;
2250 trace $_[0] if main::DEBUG && $to_trace;
2251
2252 # Each inserted line is an array, with the first element being 1 to
2253 # indicate that this line has been adjusted
f998e60c
KW
2254 no overloading;
2255 push @{$added_lines{0+$self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
2256 return;
2257 }
2258
2259 sub get_missings {
2260 # Returns the stored up @missings lines' values, and clears the list.
2261 # The values are in an array, consisting of the default in the first
2262 # element, and the property in the 2nd. However, since these lines
2263 # can be stacked up, the return is an array of all these arrays.
2264
2265 my $self = shift;
2266 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2267
f998e60c 2268 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2269
2270 # If not accepting a list return, just return the first one.
2271 return shift @{$missings{$addr}} unless wantarray;
2272
2273 my @return = @{$missings{$addr}};
2274 undef @{$missings{$addr}};
2275 return @return;
2276 }
2277
2278 sub _insert_property_into_line {
2279 # Add a property field to $_, if this file requires it.
2280
f998e60c
KW
2281 my $self = shift;
2282 my $addr; { no overloading; $addr = 0+$self; }
2283 my $property = $property{$addr};
99870f4d
KW
2284 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2285
2286 $_ =~ s/(;|$)/; $property$1/;
2287 return;
2288 }
2289
2290 sub carp_bad_line {
2291 # Output consistent error messages, using either a generic one, or the
2292 # one given by the optional parameter. To avoid gazillions of the
2293 # same message in case the syntax of a file is way off, this routine
2294 # only outputs the first instance of each message, incrementing a
2295 # count so the totals can be output at the end of the file.
2296
2297 my $self = shift;
2298 my $message = shift;
2299 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2300
f998e60c 2301 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2302
2303 $message = 'Unexpected line' unless $message;
2304
2305 # No trailing punctuation so as to fit with our addenda.
2306 $message =~ s/[.:;,]$//;
2307
2308 # If haven't seen this exact message before, output it now. Otherwise
2309 # increment the count of how many times it has occurred
2310 unless ($errors{$addr}->{$message}) {
2311 Carp::my_carp("$message in '$_' in "
f998e60c 2312 . $file{$addr}
99870f4d
KW
2313 . " at line $.. Skipping this line;");
2314 $errors{$addr}->{$message} = 1;
2315 }
2316 else {
2317 $errors{$addr}->{$message}++;
2318 }
2319
2320 # Clear the line to prevent any further (meaningful) processing of it.
2321 $_ = "";
2322
2323 return;
2324 }
2325} # End closure
2326
2327package Multi_Default;
2328
2329# Certain properties in early versions of Unicode had more than one possible
2330# default for code points missing from the files. In these cases, one
2331# default applies to everything left over after all the others are applied,
2332# and for each of the others, there is a description of which class of code
2333# points applies to it. This object helps implement this by storing the
2334# defaults, and for all but that final default, an eval string that generates
2335# the class that it applies to.
2336
2337
2338{ # Closure
2339
2340 main::setup_package();
2341
2342 my %class_defaults;
2343 # The defaults structure for the classes
2344 main::set_access('class_defaults', \%class_defaults);
2345
2346 my %other_default;
2347 # The default that applies to everything left over.
2348 main::set_access('other_default', \%other_default, 'r');
2349
2350
2351 sub new {
2352 # The constructor is called with default => eval pairs, terminated by
2353 # the left-over default. e.g.
2354 # Multi_Default->new(
2355 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
2356 # - 0x200D',
2357 # 'R' => 'some other expression that evaluates to code points',
2358 # .
2359 # .
2360 # .
2361 # 'U'));
2362
2363 my $class = shift;
2364
2365 my $self = bless \do{my $anonymous_scalar}, $class;
f998e60c 2366 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2367
2368 while (@_ > 1) {
2369 my $default = shift;
2370 my $eval = shift;
2371 $class_defaults{$addr}->{$default} = $eval;
2372 }
2373
2374 $other_default{$addr} = shift;
2375
2376 return $self;
2377 }
2378
2379 sub get_next_defaults {
2380 # Iterates and returns the next class of defaults.
2381 my $self = shift;
2382 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2383
f998e60c 2384 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2385
2386 return each %{$class_defaults{$addr}};
2387 }
2388}
2389
2390package Alias;
2391
2392# An alias is one of the names that a table goes by. This class defines them
2393# including some attributes. Everything is currently setup in the
2394# constructor.
2395
2396
2397{ # Closure
2398
2399 main::setup_package();
2400
2401 my %name;
2402 main::set_access('name', \%name, 'r');
2403
2404 my %loose_match;
2405 # Determined by the constructor code if this name should match loosely or
2406 # not. The constructor parameters can override this, but it isn't fully
2407 # implemented, as should have ability to override Unicode one's via
2408 # something like a set_loose_match()
2409 main::set_access('loose_match', \%loose_match, 'r');
2410
2411 my %make_pod_entry;
2412 # Some aliases should not get their own entries because they are covered
2413 # by a wild-card, and some we want to discourage use of. Binary
2414 main::set_access('make_pod_entry', \%make_pod_entry, 'r');
2415
2416 my %status;
2417 # Aliases have a status, like deprecated, or even suppressed (which means
2418 # they don't appear in documentation). Enum
2419 main::set_access('status', \%status, 'r');
2420
2421 my %externally_ok;
2422 # Similarly, some aliases should not be considered as usable ones for
2423 # external use, such as file names, or we don't want documentation to
2424 # recommend them. Boolean
2425 main::set_access('externally_ok', \%externally_ok, 'r');
2426
2427 sub new {
2428 my $class = shift;
2429
2430 my $self = bless \do { my $anonymous_scalar }, $class;
f998e60c 2431 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2432
2433 $name{$addr} = shift;
2434 $loose_match{$addr} = shift;
2435 $make_pod_entry{$addr} = shift;
2436 $externally_ok{$addr} = shift;
2437 $status{$addr} = shift;
2438
2439 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2440
2441 # Null names are never ok externally
2442 $externally_ok{$addr} = 0 if $name{$addr} eq "";
2443
2444 return $self;
2445 }
2446}
2447
2448package Range;
2449
2450# A range is the basic unit for storing code points, and is described in the
2451# comments at the beginning of the program. Each range has a starting code
2452# point; an ending code point (not less than the starting one); a value
2453# that applies to every code point in between the two end-points, inclusive;
2454# and an enum type that applies to the value. The type is for the user's
2455# convenience, and has no meaning here, except that a non-zero type is
2456# considered to not obey the normal Unicode rules for having standard forms.
2457#
2458# The same structure is used for both map and match tables, even though in the
2459# latter, the value (and hence type) is irrelevant and could be used as a
2460# comment. In map tables, the value is what all the code points in the range
2461# map to. Type 0 values have the standardized version of the value stored as
2462# well, so as to not have to recalculate it a lot.
2463
2464sub trace { return main::trace(@_); }
2465
2466{ # Closure
2467
2468 main::setup_package();
2469
2470 my %start;
2471 main::set_access('start', \%start, 'r', 's');
2472
2473 my %end;
2474 main::set_access('end', \%end, 'r', 's');
2475
2476 my %value;
2477 main::set_access('value', \%value, 'r');
2478
2479 my %type;
2480 main::set_access('type', \%type, 'r');
2481
2482 my %standard_form;
2483 # The value in internal standard form. Defined only if the type is 0.
2484 main::set_access('standard_form', \%standard_form);
2485
2486 # Note that if these fields change, the dump() method should as well
2487
2488 sub new {
2489 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
2490 my $class = shift;
2491
2492 my $self = bless \do { my $anonymous_scalar }, $class;
f998e60c 2493 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2494
2495 $start{$addr} = shift;
2496 $end{$addr} = shift;
2497
2498 my %args = @_;
2499
2500 my $value = delete $args{'Value'}; # Can be 0
2501 $value = "" unless defined $value;
2502 $value{$addr} = $value;
2503
2504 $type{$addr} = delete $args{'Type'} || 0;
2505
2506 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2507
2508 if (! $type{$addr}) {
2509 $standard_form{$addr} = main::standardize($value);
2510 }
2511
2512 return $self;
2513 }
2514
2515 use overload
2516 fallback => 0,
2517 qw("") => "_operator_stringify",
2518 "." => \&main::_operator_dot,
2519 ;
2520
2521 sub _operator_stringify {
2522 my $self = shift;
f998e60c 2523 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2524
2525 # Output it like '0041..0065 (value)'
2526 my $return = sprintf("%04X", $start{$addr})
2527 . '..'
2528 . sprintf("%04X", $end{$addr});
2529 my $value = $value{$addr};
2530 my $type = $type{$addr};
2531 $return .= ' (';
2532 $return .= "$value";
2533 $return .= ", Type=$type" if $type != 0;
2534 $return .= ')';
2535
2536 return $return;
2537 }
2538
2539 sub standard_form {
2540 # The standard form is the value itself if the standard form is
2541 # undefined (that is if the value is special)
2542
2543 my $self = shift;
2544 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2545
f998e60c 2546 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2547
2548 return $standard_form{$addr} if defined $standard_form{$addr};
2549 return $value{$addr};
2550 }
2551
2552 sub dump {
2553 # Human, not machine readable. For machine readable, comment out this
2554 # entire routine and let the standard one take effect.
2555 my $self = shift;
2556 my $indent = shift;
2557 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2558
f998e60c 2559 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2560
2561 my $return = $indent
2562 . sprintf("%04X", $start{$addr})
2563 . '..'
2564 . sprintf("%04X", $end{$addr})
2565 . " '$value{$addr}';";
2566 if (! defined $standard_form{$addr}) {
2567 $return .= "(type=$type{$addr})";
2568 }
2569 elsif ($standard_form{$addr} ne $value{$addr}) {
2570 $return .= "(standard '$standard_form{$addr}')";
2571 }
2572 return $return;
2573 }
2574} # End closure
2575
2576package _Range_List_Base;
2577
2578# Base class for range lists. A range list is simply an ordered list of
2579# ranges, so that the ranges with the lowest starting numbers are first in it.
2580#
2581# When a new range is added that is adjacent to an existing range that has the
2582# same value and type, it merges with it to form a larger range.
2583#
2584# Ranges generally do not overlap, except that there can be multiple entries
2585# of single code point ranges. This is because of NameAliases.txt.
2586#
2587# In this program, there is a standard value such that if two different
2588# values, have the same standard value, they are considered equivalent. This
2589# value was chosen so that it gives correct results on Unicode data
2590
2591# There are a number of methods to manipulate range lists, and some operators
2592# are overloaded to handle them.
2593
99870f4d
KW
2594sub trace { return main::trace(@_); }
2595
2596{ # Closure
2597
2598 our $addr;
2599
2600 main::setup_package();
2601
2602 my %ranges;
2603 # The list of ranges
2604 main::set_access('ranges', \%ranges, 'readable_array');
2605
2606 my %max;
2607 # The highest code point in the list. This was originally a method, but
2608 # actual measurements said it was used a lot.
2609 main::set_access('max', \%max, 'r');
2610
2611 my %each_range_iterator;
2612 # Iterator position for each_range()
2613 main::set_access('each_range_iterator', \%each_range_iterator);
2614
2615 my %owner_name_of;
2616 # Name of parent this is attached to, if any. Solely for better error
2617 # messages.
2618 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
2619
2620 my %_search_ranges_cache;
2621 # A cache of the previous result from _search_ranges(), for better
2622 # performance
2623 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
2624
2625 sub new {
2626 my $class = shift;
2627 my %args = @_;
2628
2629 # Optional initialization data for the range list.
2630 my $initialize = delete $args{'Initialize'};
2631
2632 my $self;
2633
2634 # Use _union() to initialize. _union() returns an object of this
2635 # class, which means that it will call this constructor recursively.
2636 # But it won't have this $initialize parameter so that it won't
2637 # infinitely loop on this.
2638 return _union($class, $initialize, %args) if defined $initialize;
2639
2640 $self = bless \do { my $anonymous_scalar }, $class;
f998e60c 2641 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2642
2643 # Optional parent object, only for debug info.
2644 $owner_name_of{$addr} = delete $args{'Owner'};
2645 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
2646
2647 # Stringify, in case it is an object.
2648 $owner_name_of{$addr} = "$owner_name_of{$addr}";
2649
2650 # This is used only for error messages, and so a colon is added
2651 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
2652
2653 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2654
2655 # Max is initialized to a negative value that isn't adjacent to 0,
2656 # for simpler tests
2657 $max{$addr} = -2;
2658
2659 $_search_ranges_cache{$addr} = 0;
2660 $ranges{$addr} = [];
2661
2662 return $self;
2663 }
2664
2665 use overload
2666 fallback => 0,
2667 qw("") => "_operator_stringify",
2668 "." => \&main::_operator_dot,
2669 ;
2670
2671 sub _operator_stringify {
2672 my $self = shift;
f998e60c 2673 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2674
2675 return "Range_List attached to '$owner_name_of{$addr}'"
2676 if $owner_name_of{$addr};
2677 return "anonymous Range_List " . \$self;
2678 }
2679
2680 sub _union {
2681 # Returns the union of the input code points. It can be called as
2682 # either a constructor or a method. If called as a method, the result
2683 # will be a new() instance of the calling object, containing the union
2684 # of that object with the other parameter's code points; if called as
2685 # a constructor, the first parameter gives the class the new object
2686 # should be, and the second parameter gives the code points to go into
2687 # it.
2688 # In either case, there are two parameters looked at by this routine;
2689 # any additional parameters are passed to the new() constructor.
2690 #
2691 # The code points can come in the form of some object that contains
2692 # ranges, and has a conventionally named method to access them; or
2693 # they can be an array of individual code points (as integers); or
2694 # just a single code point.
2695 #
2696 # If they are ranges, this routine doesn't make any effort to preserve
2697 # the range values of one input over the other. Therefore this base
2698 # class should not allow _union to be called from other than
2699 # initialization code, so as to prevent two tables from being added
2700 # together where the range values matter. The general form of this
2701 # routine therefore belongs in a derived class, but it was moved here
2702 # to avoid duplication of code. The failure to overload this in this
2703 # class keeps it safe.
2704 #
2705
2706 my $self;
2707 my @args; # Arguments to pass to the constructor
2708
2709 my $class = shift;
2710
2711 # If a method call, will start the union with the object itself, and
2712 # the class of the new object will be the same as self.
2713 if (ref $class) {
2714 $self = $class;
2715 $class = ref $self;
2716 push @args, $self;
2717 }
2718
2719 # Add the other required parameter.
2720 push @args, shift;
2721 # Rest of parameters are passed on to the constructor
2722
2723 # Accumulate all records from both lists.
2724 my @records;
2725 for my $arg (@args) {
2726 #local $to_trace = 0 if main::DEBUG;
2727 trace "argument = $arg" if main::DEBUG && $to_trace;
2728 if (! defined $arg) {
2729 my $message = "";
2730 if (defined $self) {
f998e60c
KW
2731 no overloading;
2732 $message .= $owner_name_of{0+$self};
99870f4d
KW
2733 }
2734 Carp::my_carp_bug($message .= "Undefined argument to _union. No union done.");
2735 return;
2736 }
2737 $arg = [ $arg ] if ! ref $arg;
2738 my $type = ref $arg;
2739 if ($type eq 'ARRAY') {
2740 foreach my $element (@$arg) {
2741 push @records, Range->new($element, $element);
2742 }
2743 }
2744 elsif ($arg->isa('Range')) {
2745 push @records, $arg;
2746 }
2747 elsif ($arg->can('ranges')) {
2748 push @records, $arg->ranges;
2749 }
2750 else {
2751 my $message = "";
2752 if (defined $self) {
f998e60c
KW
2753 no overloading;
2754 $message .= $owner_name_of{0+$self};
99870f4d
KW
2755 }
2756 Carp::my_carp_bug($message . "Cannot take the union of a $type. No union done.");
2757 return;
2758 }
2759 }
2760
2761 # Sort with the range containing the lowest ordinal first, but if
2762 # two ranges start at the same code point, sort with the bigger range
2763 # of the two first, because it takes fewer cycles.
2764 @records = sort { ($a->start <=> $b->start)
2765 or
2766 # if b is shorter than a, b->end will be
2767 # less than a->end, and we want to select
2768 # a, so want to return -1
2769 ($b->end <=> $a->end)
2770 } @records;
2771
2772 my $new = $class->new(@_);
2773
2774 # Fold in records so long as they add new information.
2775 for my $set (@records) {
2776 my $start = $set->start;
2777 my $end = $set->end;
2778 my $value = $set->value;
2779 if ($start > $new->max) {
2780 $new->_add_delete('+', $start, $end, $value);
2781 }
2782 elsif ($end > $new->max) {
2783 $new->_add_delete('+', $new->max +1, $end, $value);
2784 }
2785 }
2786
2787 return $new;
2788 }
2789
2790 sub range_count { # Return the number of ranges in the range list
2791 my $self = shift;
2792 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2793
f998e60c
KW
2794 no overloading;
2795 return scalar @{$ranges{0+$self}};
99870f4d
KW
2796 }
2797
2798 sub min {
2799 # Returns the minimum code point currently in the range list, or if
2800 # the range list is empty, 2 beyond the max possible. This is a
2801 # method because used so rarely, that not worth saving between calls,
2802 # and having to worry about changing it as ranges are added and
2803 # deleted.
2804
2805 my $self = shift;
2806 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2807
f998e60c 2808 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2809
2810 # If the range list is empty, return a large value that isn't adjacent
2811 # to any that could be in the range list, for simpler tests
2812 return $LAST_UNICODE_CODEPOINT + 2 unless scalar @{$ranges{$addr}};
2813 return $ranges{$addr}->[0]->start;
2814 }
2815
2816 sub contains {
2817 # Boolean: Is argument in the range list? If so returns $i such that:
2818 # range[$i]->end < $codepoint <= range[$i+1]->end
2819 # which is one beyond what you want; this is so that the 0th range
2820 # doesn't return false
2821 my $self = shift;
2822 my $codepoint = shift;
2823 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2824
99870f4d
KW
2825 my $i = $self->_search_ranges($codepoint);
2826 return 0 unless defined $i;
2827
2828 # The search returns $i, such that
2829 # range[$i-1]->end < $codepoint <= range[$i]->end
2830 # So is in the table if and only iff it is at least the start position
2831 # of range $i.
f998e60c
KW
2832 no overloading;
2833 return 0 if $ranges{0+$self}->[$i]->start > $codepoint;
99870f4d
KW
2834 return $i + 1;
2835 }
2836
2837 sub value_of {
2838 # Returns the value associated with the code point, undef if none
2839
2840 my $self = shift;
2841 my $codepoint = shift;
2842 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2843
99870f4d
KW
2844 my $i = $self->contains($codepoint);
2845 return unless $i;
2846
2847 # contains() returns 1 beyond where we should look
f998e60c
KW
2848 no overloading;
2849 return $ranges{0+$self}->[$i-1]->value;
99870f4d
KW
2850 }
2851
2852 sub _search_ranges {
2853 # Find the range in the list which contains a code point, or where it
2854 # should go if were to add it. That is, it returns $i, such that:
2855 # range[$i-1]->end < $codepoint <= range[$i]->end
2856 # Returns undef if no such $i is possible (e.g. at end of table), or
2857 # if there is an error.
2858
2859 my $self = shift;
2860 my $code_point = shift;
2861 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2862
f998e60c 2863 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
2864
2865 return if $code_point > $max{$addr};
2866 my $r = $ranges{$addr}; # The current list of ranges
2867 my $range_list_size = scalar @$r;
2868 my $i;
2869
2870 use integer; # want integer division
2871
2872 # Use the cached result as the starting guess for this one, because,
2873 # an experiment on 5.1 showed that 90% of the time the cache was the
2874 # same as the result on the next call (and 7% it was one less).
2875 $i = $_search_ranges_cache{$addr};
2876 $i = 0 if $i >= $range_list_size; # Reset if no longer valid (prob.
2877 # from an intervening deletion
2878 #local $to_trace = 1 if main::DEBUG;
2879 trace "previous \$i is still valid: $i" if main::DEBUG && $to_trace && $code_point <= $r->[$i]->end && ($i == 0 || $r->[$i-1]->end < $code_point);
2880 return $i if $code_point <= $r->[$i]->end
2881 && ($i == 0 || $r->[$i-1]->end < $code_point);
2882
2883 # Here the cache doesn't yield the correct $i. Try adding 1.
2884 if ($i < $range_list_size - 1
2885 && $r->[$i]->end < $code_point &&
2886 $code_point <= $r->[$i+1]->end)
2887 {
2888 $i++;
2889 trace "next \$i is correct: $i" if main::DEBUG && $to_trace;
2890 $_search_ranges_cache{$addr} = $i;
2891 return $i;
2892 }
2893
2894 # Here, adding 1 also didn't work. We do a binary search to
2895 # find the correct position, starting with current $i
2896 my $lower = 0;
2897 my $upper = $range_list_size - 1;
2898 while (1) {
2899 trace "top of loop i=$i:", sprintf("%04X", $r->[$lower]->start), "[$lower] .. ", sprintf("%04X", $r->[$i]->start), "[$i] .. ", sprintf("%04X", $r->[$upper]->start), "[$upper]" if main::DEBUG && $to_trace;
2900
2901 if ($code_point <= $r->[$i]->end) {
2902
2903 # Here we have met the upper constraint. We can quit if we
2904 # also meet the lower one.
2905 last if $i == 0 || $r->[$i-1]->end < $code_point;
2906
2907 $upper = $i; # Still too high.
2908
2909 }
2910 else {
2911
2912 # Here, $r[$i]->end < $code_point, so look higher up.
2913 $lower = $i;
2914 }
2915
2916 # Split search domain in half to try again.
2917 my $temp = ($upper + $lower) / 2;
2918
2919 # No point in continuing unless $i changes for next time
2920 # in the loop.
2921 if ($temp == $i) {
2922
2923 # We can't reach the highest element because of the averaging.
2924 # So if one below the upper edge, force it there and try one
2925 # more time.
2926 if ($i == $range_list_size - 2) {
2927
2928 trace "Forcing to upper edge" if main::DEBUG && $to_trace;
2929 $i = $range_list_size - 1;
2930
2931 # Change $lower as well so if fails next time through,
2932 # taking the average will yield the same $i, and we will
2933 # quit with the error message just below.
2934 $lower = $i;
2935 next;
2936 }
2937 Carp::my_carp_bug("$owner_name_of{$addr}Can't find where the range ought to go. No action taken.");
2938 return;
2939 }
2940 $i = $temp;
2941 } # End of while loop
2942
2943 if (main::DEBUG && $to_trace) {
2944 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i;
2945 trace "i= [ $i ]", $r->[$i];
2946 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < $range_list_size - 1;
2947 }
2948
2949 # Here we have found the offset. Cache it as a starting point for the
2950 # next call.
2951 $_search_ranges_cache{$addr} = $i;
2952 return $i;
2953 }
2954
2955 sub _add_delete {
2956 # Add, replace or delete ranges to or from a list. The $type
2957 # parameter gives which:
2958 # '+' => insert or replace a range, returning a list of any changed
2959 # ranges.
2960 # '-' => delete a range, returning a list of any deleted ranges.
2961 #
2962 # The next three parameters give respectively the start, end, and
2963 # value associated with the range. 'value' should be null unless the
2964 # operation is '+';
2965 #
2966 # The range list is kept sorted so that the range with the lowest
2967 # starting position is first in the list, and generally, adjacent
2968 # ranges with the same values are merged into single larger one (see
2969 # exceptions below).
2970 #
2971 # There are more parameters, all are key => value pairs:
2972 # Type gives the type of the value. It is only valid for '+'.
2973 # All ranges have types; if this parameter is omitted, 0 is
2974 # assumed. Ranges with type 0 are assumed to obey the
2975 # Unicode rules for casing, etc; ranges with other types are
2976 # not. Otherwise, the type is arbitrary, for the caller's
2977 # convenience, and looked at only by this routine to keep
2978 # adjacent ranges of different types from being merged into
2979 # a single larger range, and when Replace =>
2980 # $IF_NOT_EQUIVALENT is specified (see just below).
2981 # Replace determines what to do if the range list already contains
2982 # ranges which coincide with all or portions of the input
2983 # range. It is only valid for '+':
2984 # => $NO means that the new value is not to replace
2985 # any existing ones, but any empty gaps of the
2986 # range list coinciding with the input range
2987 # will be filled in with the new value.
2988 # => $UNCONDITIONALLY means to replace the existing values with
2989 # this one unconditionally. However, if the
2990 # new and old values are identical, the
2991 # replacement is skipped to save cycles
2992 # => $IF_NOT_EQUIVALENT means to replace the existing values
2993 # with this one if they are not equivalent.
2994 # Ranges are equivalent if their types are the
2995 # same, and they are the same string, or if
2996 # both are type 0 ranges, if their Unicode
2997 # standard forms are identical. In this last
2998 # case, the routine chooses the more "modern"
2999 # one to use. This is because some of the
3000 # older files are formatted with values that
3001 # are, for example, ALL CAPs, whereas the
3002 # derived files have a more modern style,
3003 # which looks better. By looking for this
3004 # style when the pre-existing and replacement
3005 # standard forms are the same, we can move to
3006 # the modern style
3007 # => $MULTIPLE means that if this range duplicates an
3008 # existing one, but has a different value,
3009 # don't replace the existing one, but insert
3010 # this, one so that the same range can occur
3011 # multiple times.
3012 # => anything else is the same as => $IF_NOT_EQUIVALENT
3013 #
3014 # "same value" means identical for type-0 ranges, and it means having
3015 # the same standard forms for non-type-0 ranges.
3016
3017 return Carp::carp_too_few_args(\@_, 5) if main::DEBUG && @_ < 5;
3018
3019 my $self = shift;
3020 my $operation = shift; # '+' for add/replace; '-' for delete;
3021 my $start = shift;
3022 my $end = shift;
3023 my $value = shift;
3024
3025 my %args = @_;
3026
3027 $value = "" if not defined $value; # warning: $value can be "0"
3028
3029 my $replace = delete $args{'Replace'};
3030 $replace = $IF_NOT_EQUIVALENT unless defined $replace;
3031
3032 my $type = delete $args{'Type'};
3033 $type = 0 unless defined $type;
3034
3035 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3036
f998e60c 3037 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
3038
3039 if ($operation ne '+' && $operation ne '-') {
3040 Carp::my_carp_bug("$owner_name_of{$addr}First parameter to _add_delete must be '+' or '-'. No action taken.");
3041 return;
3042 }
3043 unless (defined $start && defined $end) {
3044 Carp::my_carp_bug("$owner_name_of{$addr}Undefined start and/or end to _add_delete. No action taken.");
3045 return;
3046 }
3047 unless ($end >= $start) {
3048 Carp::my_carp_bug("$owner_name_of{$addr}End of range (" . sprintf("%04X", $end) . ") must not be before start (" . sprintf("%04X", $start) . "). No action taken.");
3049 return;
3050 }
3051 #local $to_trace = 1 if main::DEBUG;
3052
3053 if ($operation eq '-') {
3054 if ($replace != $IF_NOT_EQUIVALENT) {
3055 Carp::my_carp_bug("$owner_name_of{$addr}Replace => \$IF_NOT_EQUIVALENT is required when deleting a range from a range list. Assuming Replace => \$IF_NOT_EQUIVALENT.");
3056 $replace = $IF_NOT_EQUIVALENT;
3057 }
3058 if ($type) {
3059 Carp::my_carp_bug("$owner_name_of{$addr}Type => 0 is required when deleting a range from a range list. Assuming Type => 0.");
3060 $type = 0;
3061 }
3062 if ($value ne "") {
3063 Carp::my_carp_bug("$owner_name_of{$addr}Value => \"\" is required when deleting a range from a range list. Assuming Value => \"\".");
3064 $value = "";
3065 }
3066 }
3067
3068 my $r = $ranges{$addr}; # The current list of ranges
3069 my $range_list_size = scalar @$r; # And its size
3070 my $max = $max{$addr}; # The current high code point in
3071 # the list of ranges
3072
3073 # Do a special case requiring fewer machine cycles when the new range
3074 # starts after the current highest point. The Unicode input data is
3075 # structured so this is common.
3076 if ($start > $max) {
3077
3078 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) type=$type" if main::DEBUG && $to_trace;
3079 return if $operation eq '-'; # Deleting a non-existing range is a
3080 # no-op
3081
3082 # If the new range doesn't logically extend the current final one
3083 # in the range list, create a new range at the end of the range
3084 # list. (max cleverly is initialized to a negative number not
3085 # adjacent to 0 if the range list is empty, so even adding a range
3086 # to an empty range list starting at 0 will have this 'if'
3087 # succeed.)
3088 if ($start > $max + 1 # non-adjacent means can't extend.
3089 || @{$r}[-1]->value ne $value # values differ, can't extend.
3090 || @{$r}[-1]->type != $type # types differ, can't extend.
3091 ) {
3092 push @$r, Range->new($start, $end,
3093 Value => $value,
3094 Type => $type);
3095 }
3096 else {
3097
3098 # Here, the new range starts just after the current highest in
3099 # the range list, and they have the same type and value.
3100 # Extend the current range to incorporate the new one.
3101 @{$r}[-1]->set_end($end);
3102 }
3103
3104 # This becomes the new maximum.
3105 $max{$addr} = $end;
3106
3107 return;
3108 }
3109 #local $to_trace = 0 if main::DEBUG;
3110
3111 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) replace=$replace" if main::DEBUG && $to_trace;
3112
3113 # Here, the input range isn't after the whole rest of the range list.
3114 # Most likely 'splice' will be needed. The rest of the routine finds
3115 # the needed splice parameters, and if necessary, does the splice.
3116 # First, find the offset parameter needed by the splice function for
3117 # the input range. Note that the input range may span multiple
3118 # existing ones, but we'll worry about that later. For now, just find
3119 # the beginning. If the input range is to be inserted starting in a
3120 # position not currently in the range list, it must (obviously) come
3121 # just after the range below it, and just before the range above it.
3122 # Slightly less obviously, it will occupy the position currently
3123 # occupied by the range that is to come after it. More formally, we
3124 # are looking for the position, $i, in the array of ranges, such that:
3125 #
3126 # r[$i-1]->start <= r[$i-1]->end < $start < r[$i]->start <= r[$i]->end
3127 #
3128 # (The ordered relationships within existing ranges are also shown in
3129 # the equation above). However, if the start of the input range is
3130 # within an existing range, the splice offset should point to that
3131 # existing range's position in the list; that is $i satisfies a
3132 # somewhat different equation, namely:
3133 #
3134 #r[$i-1]->start <= r[$i-1]->end < r[$i]->start <= $start <= r[$i]->end
3135 #
3136 # More briefly, $start can come before or after r[$i]->start, and at
3137 # this point, we don't know which it will be. However, these
3138 # two equations share these constraints:
3139 #
3140 # r[$i-1]->end < $start <= r[$i]->end
3141 #
3142 # And that is good enough to find $i.
3143
3144 my $i = $self->_search_ranges($start);
3145 if (! defined $i) {
3146 Carp::my_carp_bug("Searching $self for range beginning with $start unexpectedly returned undefined. Operation '$operation' not performed");
3147 return;
3148 }
3149
3150 # The search function returns $i such that:
3151 #
3152 # r[$i-1]->end < $start <= r[$i]->end
3153 #
3154 # That means that $i points to the first range in the range list
3155 # that could possibly be affected by this operation. We still don't
3156 # know if the start of the input range is within r[$i], or if it
3157 # points to empty space between r[$i-1] and r[$i].
3158 trace "[$i] is the beginning splice point. Existing range there is ", $r->[$i] if main::DEBUG && $to_trace;
3159
3160 # Special case the insertion of data that is not to replace any
3161 # existing data.
3162 if ($replace == $NO) { # If $NO, has to be operation '+'
3163 #local $to_trace = 1 if main::DEBUG;
3164 trace "Doesn't replace" if main::DEBUG && $to_trace;
3165
3166 # Here, the new range is to take effect only on those code points
3167 # that aren't already in an existing range. This can be done by
3168 # looking through the existing range list and finding the gaps in
3169 # the ranges that this new range affects, and then calling this
3170 # function recursively on each of those gaps, leaving untouched
3171 # anything already in the list. Gather up a list of the changed
3172 # gaps first so that changes to the internal state as new ranges
3173 # are added won't be a problem.
3174 my @gap_list;
3175
3176 # First, if the starting point of the input range is outside an
3177 # existing one, there is a gap from there to the beginning of the
3178 # existing range -- add a span to fill the part that this new
3179 # range occupies
3180 if ($start < $r->[$i]->start) {
3181 push @gap_list, Range->new($start,
3182 main::min($end,
3183 $r->[$i]->start - 1),
3184 Type => $type);
3185 trace "gap before $r->[$i] [$i], will add", $gap_list[-1] if main::DEBUG && $to_trace;
3186 }
3187
3188 # Then look through the range list for other gaps until we reach
3189 # the highest range affected by the input one.
3190 my $j;
3191 for ($j = $i+1; $j < $range_list_size; $j++) {
3192 trace "j=[$j]", $r->[$j] if main::DEBUG && $to_trace;
3193 last if $end < $r->[$j]->start;
3194
3195 # If there is a gap between when this range starts and the
3196 # previous one ends, add a span to fill it. Note that just
3197 # because there are two ranges doesn't mean there is a
3198 # non-zero gap between them. It could be that they have
3199 # different values or types
3200 if ($r->[$j-1]->end + 1 != $r->[$j]->start) {
3201 push @gap_list,
3202 Range->new($r->[$j-1]->end + 1,
3203 $r->[$j]->start - 1,
3204 Type => $type);
3205 trace "gap between $r->[$j-1] and $r->[$j] [$j], will add: $gap_list[-1]" if main::DEBUG && $to_trace;
3206 }
3207 }
3208
3209 # Here, we have either found an existing range in the range list,
3210 # beyond the area affected by the input one, or we fell off the
3211 # end of the loop because the input range affects the whole rest
3212 # of the range list. In either case, $j is 1 higher than the
3213 # highest affected range. If $j == $i, it means that there are no
3214 # affected ranges, that the entire insertion is in the gap between
3215 # r[$i-1], and r[$i], which we already have taken care of before
3216 # the loop.
3217 # On the other hand, if there are affected ranges, it might be
3218 # that there is a gap that needs filling after the final such
3219 # range to the end of the input range
3220 if ($r->[$j-1]->end < $end) {
3221 push @gap_list, Range->new(main::max($start,
3222 $r->[$j-1]->end + 1),
3223 $end,
3224 Type => $type);
3225 trace "gap after $r->[$j-1], will add $gap_list[-1]" if main::DEBUG && $to_trace;
3226 }
3227
3228 # Call recursively to fill in all the gaps.
3229 foreach my $gap (@gap_list) {
3230 $self->_add_delete($operation,
3231 $gap->start,
3232 $gap->end,
3233 $value,
3234 Type => $type);
3235 }
3236
3237 return;
3238 }
3239
3240 # Here, we have taken care of the case where $replace is $NO, which
3241 # means that whatever action we now take is done unconditionally. It
3242 # still could be that this call will result in a no-op, if duplicates
3243 # aren't allowed, and we are inserting a range that merely duplicates
3244 # data already in the range list; or also if deleting a non-existent
3245 # range.
3246 # $i still points to the first potential affected range. Now find the
3247 # highest range affected, which will determine the length parameter to
3248 # splice. (The input range can span multiple existing ones.) While
3249 # we are looking through the range list, see also if this is an
3250 # insertion that will change the values of at least one of the
3251 # affected ranges. We don't need to do this check unless this is an
3252 # insertion of non-multiples, and also since this is a boolean, we
3253 # don't need to do it if have already determined that it will make a
3254 # change; just unconditionally change them. $cdm is created to be 1
3255 # if either of these is true. (The 'c' in the name comes from below)
3256 my $cdm = ($operation eq '-' || $replace == $MULTIPLE);
3257 my $j; # This will point to the highest affected range
3258
3259 # For non-zero types, the standard form is the value itself;
3260 my $standard_form = ($type) ? $value : main::standardize($value);
3261
3262 for ($j = $i; $j < $range_list_size; $j++) {
3263 trace "Looking for highest affected range; the one at $j is ", $r->[$j] if main::DEBUG && $to_trace;
3264
3265 # If find a range that it doesn't overlap into, we can stop
3266 # searching
3267 last if $end < $r->[$j]->start;
3268
3269 # Here, overlaps the range at $j. If the value's don't match,
3270 # and this is supposedly an insertion, it becomes a change
3271 # instead. This is what the 'c' stands for in $cdm.
3272 if (! $cdm) {
3273 if ($r->[$j]->standard_form ne $standard_form) {
3274 $cdm = 1;
3275 }
3276 else {
3277
3278 # Here, the two values are essentially the same. If the
3279 # two are actually identical, replacing wouldn't change
3280 # anything so skip it.
3281 my $pre_existing = $r->[$j]->value;
3282 if ($pre_existing ne $value) {
3283
3284 # Here the new and old standardized values are the
3285 # same, but the non-standardized values aren't. If
3286 # replacing unconditionally, then replace
3287 if( $replace == $UNCONDITIONALLY) {
3288 $cdm = 1;
3289 }
3290 else {
3291
3292 # Here, are replacing conditionally. Decide to
3293 # replace or not based on which appears to look
3294 # the "nicest". If one is mixed case and the
3295 # other isn't, choose the mixed case one.
3296 my $new_mixed = $value =~ /[A-Z]/
3297 && $value =~ /[a-z]/;
3298 my $old_mixed = $pre_existing =~ /[A-Z]/
3299 && $pre_existing =~ /[a-z]/;
3300
3301 if ($old_mixed != $new_mixed) {
3302 $cdm = 1 if $new_mixed;
3303 if (main::DEBUG && $to_trace) {
3304 if ($cdm) {
3305 trace "Replacing $pre_existing with $value";
3306 }
3307 else {
3308 trace "Retaining $pre_existing over $value";
3309 }
3310 }
3311 }
3312 else {
3313
3314 # Here casing wasn't different between the two.
3315 # If one has hyphens or underscores and the
3316 # other doesn't, choose the one with the
3317 # punctuation.
3318 my $new_punct = $value =~ /[-_]/;
3319 my $old_punct = $pre_existing =~ /[-_]/;
3320
3321 if ($old_punct != $new_punct) {
3322 $cdm = 1 if $new_punct;
3323 if (main::DEBUG && $to_trace) {
3324 if ($cdm) {
3325 trace "Replacing $pre_existing with $value";
3326 }
3327 else {
3328 trace "Retaining $pre_existing over $value";
3329 }
3330 }
3331 } # else existing one is just as "good";
3332 # retain it to save cycles.
3333 }
3334 }
3335 }
3336 }
3337 }
3338 } # End of loop looking for highest affected range.
3339
3340 # Here, $j points to one beyond the highest range that this insertion
3341 # affects (hence to beyond the range list if that range is the final
3342 # one in the range list).
3343
3344 # The splice length is all the affected ranges. Get it before
3345 # subtracting, for efficiency, so we don't have to later add 1.
3346 my $length = $j - $i;
3347
3348 $j--; # $j now points to the highest affected range.
3349 trace "Final affected range is $j: $r->[$j]" if main::DEBUG && $to_trace;
3350
3351 # If inserting a multiple record, this is where it goes, after all the
3352 # existing ones for this range. This implies an insertion, and no
3353 # change to any existing ranges. Note that $j can be -1 if this new
3354 # range doesn't actually duplicate any existing, and comes at the
3355 # beginning of the list, in which case we can handle it like any other
3356 # insertion, and is easier to do so.
3357 if ($replace == $MULTIPLE && $j >= 0) {
3358
3359 # This restriction could be remedied with a little extra work, but
3360 # it won't hopefully ever be necessary
3361 if ($r->[$j]->start != $r->[$j]->end) {
3362 Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple when the other range ($r->[$j]) contains more than one code point. No action taken.");
3363 return;
3364 }
3365
3366 # Don't add an exact duplicate, as it isn't really a multiple
3367 return if $value eq $r->[$j]->value && $type eq $r->[$j]->type;
3368
3369 trace "Adding multiple record at $j+1 with $start..$end, $value" if main::DEBUG && $to_trace;
3370 my @return = splice @$r,
3371 $j+1,
3372 0,
3373 Range->new($start,
3374 $end,
3375 Value => $value,
3376 Type => $type);
3377 if (main::DEBUG && $to_trace) {
3378 trace "After splice:";
3379 trace 'j-2=[', $j-2, ']', $r->[$j-2] if $j >= 2;
3380 trace 'j-1=[', $j-1, ']', $r->[$j-1] if $j >= 1;
3381 trace "j =[", $j, "]", $r->[$j] if $j >= 0;
3382 trace 'j+1=[', $j+1, ']', $r->[$j+1] if $j < @$r - 1;
3383 trace 'j+2=[', $j+2, ']', $r->[$j+2] if $j < @$r - 2;
3384 trace 'j+3=[', $j+3, ']', $r->[$j+3] if $j < @$r - 3;
3385 }
3386 return @return;
3387 }
3388
3389 # Here, have taken care of $NO and $MULTIPLE replaces.
3390 # $j points to the highest affected range. But it can be < $i or even
3391 # -1. These happen only if the insertion is entirely in the gap
3392 # between r[$i-1] and r[$i]. Here's why: j < i means that the j loop
3393 # above exited first time through with $end < $r->[$i]->start. (And
3394 # then we subtracted one from j) This implies also that $start <
3395 # $r->[$i]->start, but we know from above that $r->[$i-1]->end <
3396 # $start, so the entire input range is in the gap.
3397 if ($j < $i) {
3398
3399 # Here the entire input range is in the gap before $i.
3400
3401 if (main::DEBUG && $to_trace) {
3402 if ($i) {
3403 trace "Entire range is between $r->[$i-1] and $r->[$i]";
3404 }
3405 else {
3406 trace "Entire range is before $r->[$i]";
3407 }
3408 }
3409 return if $operation ne '+'; # Deletion of a non-existent range is
3410 # a no-op
3411 }
3412 else {
3413
3414 # Here the entire input range is not in the gap before $i. There
3415 # is an affected one, and $j points to the highest such one.
3416
3417 # At this point, here is the situation:
3418 # This is not an insertion of a multiple, nor of tentative ($NO)
3419 # data.
3420 # $i points to the first element in the current range list that
3421 # may be affected by this operation. In fact, we know
3422 # that the range at $i is affected because we are in
3423 # the else branch of this 'if'
3424 # $j points to the highest affected range.
3425 # In other words,
3426 # r[$i-1]->end < $start <= r[$i]->end
3427 # And:
3428 # r[$i-1]->end < $start <= $end <= r[$j]->end
3429 #
3430 # Also:
3431 # $cdm is a boolean which is set true if and only if this is a
3432 # change or deletion (multiple was handled above). In
3433 # other words, it could be renamed to be just $cd.
3434
3435 # We now have enough information to decide if this call is a no-op
3436 # or not. It is a no-op if it is a deletion of a non-existent
3437 # range, or an insertion of already existing data.
3438
3439 if (main::DEBUG && $to_trace && ! $cdm
3440 && $i == $j
3441 && $start >= $r->[$i]->start)
3442 {
3443 trace "no-op";
3444 }
3445 return if ! $cdm # change or delete => not no-op
3446 && $i == $j # more than one affected range => not no-op
3447
3448 # Here, r[$i-1]->end < $start <= $end <= r[$i]->end
3449 # Further, $start and/or $end is >= r[$i]->start
3450 # The test below hence guarantees that
3451 # r[$i]->start < $start <= $end <= r[$i]->end
3452 # This means the input range is contained entirely in
3453 # the one at $i, so is a no-op
3454 && $start >= $r->[$i]->start;
3455 }
3456
3457 # Here, we know that some action will have to be taken. We have
3458 # calculated the offset and length (though adjustments may be needed)
3459 # for the splice. Now start constructing the replacement list.
3460 my @replacement;
3461 my $splice_start = $i;
3462
3463 my $extends_below;
3464 my $extends_above;
3465
3466 # See if should extend any adjacent ranges.
3467 if ($operation eq '-') { # Don't extend deletions
3468 $extends_below = $extends_above = 0;
3469 }
3470 else { # Here, should extend any adjacent ranges. See if there are
3471 # any.
3472 $extends_below = ($i > 0
3473 # can't extend unless adjacent
3474 && $r->[$i-1]->end == $start -1
3475 # can't extend unless are same standard value
3476 && $r->[$i-1]->standard_form eq $standard_form
3477 # can't extend unless share type
3478 && $r->[$i-1]->type == $type);
3479 $extends_above = ($j+1 < $range_list_size
3480 && $r->[$j+1]->start == $end +1
3481 && $r->[$j+1]->standard_form eq $standard_form
3482 && $r->[$j-1]->type == $type);
3483 }
3484 if ($extends_below && $extends_above) { # Adds to both
3485 $splice_start--; # start replace at element below
3486 $length += 2; # will replace on both sides
3487 trace "Extends both below and above ranges" if main::DEBUG && $to_trace;
3488
3489 # The result will fill in any gap, replacing both sides, and
3490 # create one large range.
3491 @replacement = Range->new($r->[$i-1]->start,
3492 $r->[$j+1]->end,
3493 Value => $value,
3494 Type => $type);
3495 }
3496 else {
3497
3498 # Here we know that the result won't just be the conglomeration of
3499 # a new range with both its adjacent neighbors. But it could
3500 # extend one of them.
3501
3502 if ($extends_below) {
3503
3504 # Here the new element adds to the one below, but not to the
3505 # one above. If inserting, and only to that one range, can
3506 # just change its ending to include the new one.
3507 if ($length == 0 && ! $cdm) {
3508 $r->[$i-1]->set_end($end);
3509 trace "inserted range extends range to below so it is now $r->[$i-1]" if main::DEBUG && $to_trace;
3510 return;
3511 }
3512 else {
3513 trace "Changing inserted range to start at ", sprintf("%04X", $r->[$i-1]->start), " instead of ", sprintf("%04X", $start) if main::DEBUG && $to_trace;
3514 $splice_start--; # start replace at element below
3515 $length++; # will replace the element below
3516 $start = $r->[$i-1]->start;
3517 }
3518 }
3519 elsif ($extends_above) {
3520
3521 # Here the new element adds to the one above, but not below.
3522 # Mirror the code above
3523 if ($length == 0 && ! $cdm) {
3524 $r->[$j+1]->set_start($start);
3525 trace "inserted range extends range to above so it is now $r->[$j+1]" if main::DEBUG && $to_trace;
3526 return;
3527 }
3528 else {
3529 trace "Changing inserted range to end at ", sprintf("%04X", $r->[$j+1]->end), " instead of ", sprintf("%04X", $end) if main::DEBUG && $to_trace;
3530 $length++; # will replace the element above
3531 $end = $r->[$j+1]->end;
3532 }
3533 }
3534
3535 trace "Range at $i is $r->[$i]" if main::DEBUG && $to_trace;
3536
3537 # Finally, here we know there will have to be a splice.
3538 # If the change or delete affects only the highest portion of the
3539 # first affected range, the range will have to be split. The
3540 # splice will remove the whole range, but will replace it by a new
3541 # range containing just the unaffected part. So, in this case,
3542 # add to the replacement list just this unaffected portion.
3543 if (! $extends_below
3544 && $start > $r->[$i]->start && $start <= $r->[$i]->end)
3545 {
3546 push @replacement,
3547 Range->new($r->[$i]->start,
3548 $start - 1,
3549 Value => $r->[$i]->value,
3550 Type => $r->[$i]->type);
3551 }
3552
3553 # In the case of an insert or change, but not a delete, we have to
3554 # put in the new stuff; this comes next.
3555 if ($operation eq '+') {
3556 push @replacement, Range->new($start,
3557 $end,
3558 Value => $value,
3559 Type => $type);
3560 }
3561
3562 trace "Range at $j is $r->[$j]" if main::DEBUG && $to_trace && $j != $i;
3563 #trace "$end >=", $r->[$j]->start, " && $end <", $r->[$j]->end if main::DEBUG && $to_trace;
3564
3565 # And finally, if we're changing or deleting only a portion of the
3566 # highest affected range, it must be split, as the lowest one was.
3567 if (! $extends_above
3568 && $j >= 0 # Remember that j can be -1 if before first
3569 # current element
3570 && $end >= $r->[$j]->start
3571 && $end < $r->[$j]->end)
3572 {
3573 push @replacement,
3574 Range->new($end + 1,
3575 $r->[$j]->end,
3576 Value => $r->[$j]->value,
3577 Type => $r->[$j]->type);
3578 }
3579 }
3580
3581 # And do the splice, as calculated above
3582 if (main::DEBUG && $to_trace) {
3583 trace "replacing $length element(s) at $i with ";
3584 foreach my $replacement (@replacement) {
3585 trace " $replacement";
3586 }
3587 trace "Before splice:";
3588 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3589 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3590 trace "i =[", $i, "]", $r->[$i];
3591 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3592 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3593 }
3594
3595 my @return = splice @$r, $splice_start, $length, @replacement;
3596
3597 if (main::DEBUG && $to_trace) {
3598 trace "After splice:";
3599 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3600 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3601 trace "i =[", $i, "]", $r->[$i];
3602 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3603 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3604 trace "removed @return";
3605 }
3606
3607 # An actual deletion could have changed the maximum in the list.
3608 # There was no deletion if the splice didn't return something, but
3609 # otherwise recalculate it. This is done too rarely to worry about
3610 # performance.
3611 if ($operation eq '-' && @return) {
3612 $max{$addr} = $r->[-1]->end;
3613 }
3614 return @return;
3615 }
3616
3617 sub reset_each_range { # reset the iterator for each_range();
3618 my $self = shift;
3619 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3620
f998e60c
KW
3621 no overloading;
3622 undef $each_range_iterator{0+$self};
99870f4d
KW
3623 return;
3624 }
3625
3626 sub each_range {
3627 # Iterate over each range in a range list. Results are undefined if
3628 # the range list is changed during the iteration.
3629
3630 my $self = shift;
3631 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3632
f998e60c 3633 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
3634
3635 return if $self->is_empty;
3636
3637 $each_range_iterator{$addr} = -1
3638 if ! defined $each_range_iterator{$addr};
3639 $each_range_iterator{$addr}++;
3640 return $ranges{$addr}->[$each_range_iterator{$addr}]
3641 if $each_range_iterator{$addr} < @{$ranges{$addr}};
3642 undef $each_range_iterator{$addr};
3643 return;
3644 }
3645
3646 sub count { # Returns count of code points in range list
3647 my $self = shift;
3648 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3649
f998e60c 3650 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
3651
3652 my $count = 0;
3653 foreach my $range (@{$ranges{$addr}}) {
3654 $count += $range->end - $range->start + 1;
3655 }
3656 return $count;
3657 }
3658
3659 sub delete_range { # Delete a range
3660 my $self = shift;
3661 my $start = shift;
3662 my $end = shift;
3663
3664 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3665
3666 return $self->_add_delete('-', $start, $end, "");
3667 }
3668
3669 sub is_empty { # Returns boolean as to if a range list is empty
3670 my $self = shift;
3671 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3672
f998e60c
KW
3673 no overloading;
3674 return scalar @{$ranges{0+$self}} == 0;
99870f4d
KW
3675 }
3676
3677 sub hash {
3678 # Quickly returns a scalar suitable for separating tables into
3679 # buckets, i.e. it is a hash function of the contents of a table, so
3680 # there are relatively few conflicts.
3681
3682 my $self = shift;
3683 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3684
f998e60c 3685 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
3686
3687 # These are quickly computable. Return looks like 'min..max;count'
3688 return $self->min . "..$max{$addr};" . scalar @{$ranges{$addr}};
3689 }
3690} # End closure for _Range_List_Base
3691
3692package Range_List;
3693use base '_Range_List_Base';
3694
3695# A Range_List is a range list for match tables; i.e. the range values are
3696# not significant. Thus a number of operations can be safely added to it,
3697# such as inversion, intersection. Note that union is also an unsafe
3698# operation when range values are cared about, and that method is in the base
3699# class, not here. But things are set up so that that method is callable only
3700# during initialization. Only in this derived class, is there an operation
3701# that combines two tables. A Range_Map can thus be used to initialize a
3702# Range_List, and its mappings will be in the list, but are not significant to
3703# this class.
3704
3705sub trace { return main::trace(@_); }
3706
3707{ # Closure
3708
3709 use overload
3710 fallback => 0,
3711 '+' => sub { my $self = shift;
3712 my $other = shift;
3713
3714 return $self->_union($other)
3715 },
3716 '&' => sub { my $self = shift;
3717 my $other = shift;
3718
3719 return $self->_intersect($other, 0);
3720 },
3721 '~' => "_invert",
3722 '-' => "_subtract",
3723 ;
3724
3725 sub _invert {
3726 # Returns a new Range_List that gives all code points not in $self.
3727
3728 my $self = shift;
3729
3730 my $new = Range_List->new;
3731
3732 # Go through each range in the table, finding the gaps between them
3733 my $max = -1; # Set so no gap before range beginning at 0
3734 for my $range ($self->ranges) {
3735 my $start = $range->start;
3736 my $end = $range->end;
3737
3738 # If there is a gap before this range, the inverse will contain
3739 # that gap.
3740 if ($start > $max + 1) {
3741 $new->add_range($max + 1, $start - 1);
3742 }
3743 $max = $end;
3744 }
3745
3746 # And finally, add the gap from the end of the table to the max
3747 # possible code point
3748 if ($max < $LAST_UNICODE_CODEPOINT) {
3749 $new->add_range($max + 1, $LAST_UNICODE_CODEPOINT);
3750 }
3751 return $new;
3752 }
3753
3754 sub _subtract {
3755 # Returns a new Range_List with the argument deleted from it. The
3756 # argument can be a single code point, a range, or something that has
3757 # a range, with the _range_list() method on it returning them
3758
3759 my $self = shift;
3760 my $other = shift;
3761 my $reversed = shift;
3762 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3763
3764 if ($reversed) {
3765 Carp::my_carp_bug("Can't cope with a "
3766 . __PACKAGE__
3767 . " being the second parameter in a '-'. Subtraction ignored.");
3768 return $self;
3769 }
3770
3771 my $new = Range_List->new(Initialize => $self);
3772
3773 if (! ref $other) { # Single code point
3774 $new->delete_range($other, $other);
3775 }
3776 elsif ($other->isa('Range')) {
3777 $new->delete_range($other->start, $other->end);
3778 }
3779 elsif ($other->can('_range_list')) {
3780 foreach my $range ($other->_range_list->ranges) {
3781 $new->delete_range($range->start, $range->end);
3782 }
3783 }
3784 else {
3785 Carp::my_carp_bug("Can't cope with a "
3786 . ref($other)
3787 . " argument to '-'. Subtraction ignored."
3788 );
3789 return $self;
3790 }
3791
3792 return $new;
3793 }
3794
3795 sub _intersect {
3796 # Returns either a boolean giving whether the two inputs' range lists
3797 # intersect (overlap), or a new Range_List containing the intersection
3798 # of the two lists. The optional final parameter being true indicates
3799 # to do the check instead of the intersection.
3800
3801 my $a_object = shift;
3802 my $b_object = shift;
3803 my $check_if_overlapping = shift;
3804 $check_if_overlapping = 0 unless defined $check_if_overlapping;
3805 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3806
3807 if (! defined $b_object) {
3808 my $message = "";
3809 $message .= $a_object->_owner_name_of if defined $a_object;
3810 Carp::my_carp_bug($message .= "Called with undefined value. Intersection not done.");
3811 return;
3812 }
3813
3814 # a & b = !(!a | !b), or in our terminology = ~ ( ~a + -b )
3815 # Thus the intersection could be much more simply be written:
3816 # return ~(~$a_object + ~$b_object);
3817 # But, this is slower, and when taking the inverse of a large
3818 # range_size_1 table, back when such tables were always stored that
3819 # way, it became prohibitively slow, hence the code was changed to the
3820 # below
3821
3822 if ($b_object->isa('Range')) {
3823 $b_object = Range_List->new(Initialize => $b_object,
3824 Owner => $a_object->_owner_name_of);
3825 }
3826 $b_object = $b_object->_range_list if $b_object->can('_range_list');
3827
3828 my @a_ranges = $a_object->ranges;
3829 my @b_ranges = $b_object->ranges;
3830
3831 #local $to_trace = 1 if main::DEBUG;
3832 trace "intersecting $a_object with ", scalar @a_ranges, "ranges and $b_object with", scalar @b_ranges, " ranges" if main::DEBUG && $to_trace;
3833
3834 # Start with the first range in each list
3835 my $a_i = 0;
3836 my $range_a = $a_ranges[$a_i];
3837 my $b_i = 0;
3838 my $range_b = $b_ranges[$b_i];
3839
3840 my $new = __PACKAGE__->new(Owner => $a_object->_owner_name_of)
3841 if ! $check_if_overlapping;
3842
3843 # If either list is empty, there is no intersection and no overlap
3844 if (! defined $range_a || ! defined $range_b) {
3845 return $check_if_overlapping ? 0 : $new;
3846 }
3847 trace "range_a[$a_i]=$range_a; range_b[$b_i]=$range_b" if main::DEBUG && $to_trace;
3848
3849 # Otherwise, must calculate the intersection/overlap. Start with the
3850 # very first code point in each list
3851 my $a = $range_a->start;
3852 my $b = $range_b->start;
3853
3854 # Loop through all the ranges of each list; in each iteration, $a and
3855 # $b are the current code points in their respective lists
3856 while (1) {
3857
3858 # If $a and $b are the same code point, ...
3859 if ($a == $b) {
3860
3861 # it means the lists overlap. If just checking for overlap
3862 # know the answer now,
3863 return 1 if $check_if_overlapping;
3864
3865 # The intersection includes this code point plus anything else
3866 # common to both current ranges.
3867 my $start = $a;
3868 my $end = main::min($range_a->end, $range_b->end);
3869 if (! $check_if_overlapping) {
3870 trace "adding intersection range ", sprintf("%04X", $start) . ".." . sprintf("%04X", $end) if main::DEBUG && $to_trace;
3871 $new->add_range($start, $end);
3872 }
3873
3874 # Skip ahead to the end of the current intersect
3875 $a = $b = $end;
3876
3877 # If the current intersect ends at the end of either range (as
3878 # it must for at least one of them), the next possible one
3879 # will be the beginning code point in it's list's next range.
3880 if ($a == $range_a->end) {
3881 $range_a = $a_ranges[++$a_i];
3882 last unless defined $range_a;
3883 $a = $range_a->start;
3884 }
3885 if ($b == $range_b->end) {
3886 $range_b = $b_ranges[++$b_i];
3887 last unless defined $range_b;
3888 $b = $range_b->start;
3889 }
3890
3891 trace "range_a[$a_i]=$range_a; range_b[$b_i]=$range_b" if main::DEBUG && $to_trace;
3892 }
3893 elsif ($a < $b) {
3894
3895 # Not equal, but if the range containing $a encompasses $b,
3896 # change $a to be the middle of the range where it does equal
3897 # $b, so the next iteration will get the intersection
3898 if ($range_a->end >= $b) {
3899 $a = $b;
3900 }
3901 else {
3902
3903 # Here, the current range containing $a is entirely below
3904 # $b. Go try to find a range that could contain $b.
3905 $a_i = $a_object->_search_ranges($b);
3906
3907 # If no range found, quit.
3908 last unless defined $a_i;
3909
3910 # The search returns $a_i, such that
3911 # range_a[$a_i-1]->end < $b <= range_a[$a_i]->end
3912 # Set $a to the beginning of this new range, and repeat.
3913 $range_a = $a_ranges[$a_i];
3914 $a = $range_a->start;
3915 }
3916 }
3917 else { # Here, $b < $a.
3918
3919 # Mirror image code to the leg just above
3920 if ($range_b->end >= $a) {
3921 $b = $a;
3922 }
3923 else {
3924 $b_i = $b_object->_search_ranges($a);
3925 last unless defined $b_i;
3926 $range_b = $b_ranges[$b_i];
3927 $b = $range_b->start;
3928 }
3929 }
3930 } # End of looping through ranges.
3931
3932 # Intersection fully computed, or now know that there is no overlap
3933 return $check_if_overlapping ? 0 : $new;
3934 }
3935
3936 sub overlaps {
3937 # Returns boolean giving whether the two arguments overlap somewhere
3938
3939 my $self = shift;
3940 my $other = shift;
3941 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3942
3943 return $self->_intersect($other, 1);
3944 }
3945
3946 sub add_range {
3947 # Add a range to the list.
3948
3949 my $self = shift;
3950 my $start = shift;
3951 my $end = shift;
3952 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3953
3954 return $self->_add_delete('+', $start, $end, "");
3955 }
3956
99870f4d
KW
3957 sub is_code_point_usable {
3958 # This used only for making the test script. See if the input
3959 # proposed trial code point is one that Perl will handle. If second
3960 # parameter is 0, it won't select some code points for various
3961 # reasons, noted below.
3962
3963 my $code = shift;
3964 my $try_hard = shift;
3965 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3966
3967 return 0 if $code < 0; # Never use a negative
3968
99870f4d
KW
3969 # shun null. I'm (khw) not sure why this was done, but NULL would be
3970 # the character very frequently used.
3971 return $try_hard if $code == 0x0000;
3972
3973 return 0 if $try_hard; # XXX Temporary until fix utf8.c
3974
3975 # shun non-character code points.
3976 return $try_hard if $code >= 0xFDD0 && $code <= 0xFDEF;
3977 return $try_hard if ($code & 0xFFFE) == 0xFFFE; # includes FFFF
3978
3979 return $try_hard if $code > $LAST_UNICODE_CODEPOINT; # keep in range
3980 return $try_hard if $code >= 0xD800 && $code <= 0xDFFF; # no surrogate
3981
3982 return 1;
3983 }
3984
3985 sub get_valid_code_point {
3986 # Return a code point that's part of the range list. Returns nothing
3987 # if the table is empty or we can't find a suitable code point. This
3988 # used only for making the test script.
3989
3990 my $self = shift;
3991 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3992
f998e60c 3993 my $addr; { no overloading; $addr = 0+$self; }
99870f4d
KW
3994
3995 # On first pass, don't choose less desirable code points; if no good
3996 # one is found, repeat, allowing a less desirable one to be selected.
3997 for my $try_hard (0, 1) {
3998
3999 # Look through all the ranges for a usable code point.
4000 for my $set ($self->ranges) {
4001
4002 # Try the edge cases first, starting with the end point of the
4003 # range.
4004 my $end = $set->end;
4005 return $end if is_code_point_usable($end, $try_hard);
4006
4007 # End point didn't, work. Start at the beginning and try
4008 # every one until find one that does work.
4009 for my $trial ($set->start .. $end - 1) {
4010 return $trial if is_code_point_usable($trial, $try_hard);
4011 }
4012 }
4013 }
4014 return (); # If none found, give up.
4015 }
4016
4017 sub get_invalid_code_point {
4018 # Return a code point that's not part of the table. Returns nothing
4019 # if the table covers all code points or a suitable code point can't
4020 # be found. This used only for making the test script.
4021
4022 my $self = shift;
4023 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
4024
4025 # Just find a valid code point of the inverse, if any.
4026 return Range_List->new(Initialize => ~ $self)->get_valid_code_point;
4027 }
4028} # end closure for Range_List
4029
4030package Range_Map;
4031use base '_Range_List_Base';
4032
4033# A Range_Map is a range list in which the range values (called maps) are
4034# significant, and hence shouldn't be manipulated by our other code, which
4035# could be ambiguous or lose things. For example, in taking the union of two
4036# lists, which share code points, but which have differing values, which one
4037# has precedence in the union?
4038# It turns out that these operations aren't really necessary for map tables,
4039# and so this class was created to make sure they aren't accidentally
4040# applied to them.
4041
4042{ # Closure
4043
4044 sub add_map {
4045 # Add a range containing a mapping value to the list
4046
4047 my $self = shift;
4048 # Rest of parameters passed on
4049
4050 return $self->_add_delete('+', @_);
4051 }
4052
4053 sub add_duplicate {
4054 # Adds entry to a range list which can duplicate an existing entry
4055
4056 my $self = shift;
4057 my $code_point = shift;
4058