This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perluniprops: specify exact chars in PerlSpace
[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:
ffe43484 12# my $addr = do { no overloading; pack 'J', $self; }
f998e60c
KW
13# with
14# my $addr = main::objaddr $self;
6c68572b 15# (or reverse commit 9b01bafde4b022706c3d6f947a0963f821b2e50b
051df77b
NC
16# that instituted the change to main::objaddr, and subsequent commits that
17# changed 0+$self to pack 'J', $self.)
6c68572b 18
cdcef19a 19my $start_time;
98dc9551 20BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
21 # get it as close as possible
22 $start_time= time;
23}
24
25
23e33b60 26require 5.010_001;
d73e5302 27use strict;
99870f4d 28use warnings;
cf25bb62 29use Carp;
bd9ebcfd 30use Config;
99870f4d
KW
31use File::Find;
32use File::Path;
d07a55ed 33use File::Spec;
99870f4d
KW
34use Text::Tabs;
35
36sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 37my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d
KW
38
39##########################################################################
40#
41# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
42# from the Unicode database files (lib/unicore/.../*.txt), It also generates
43# a pod file and a .t file
44#
45# The structure of this file is:
46# First these introductory comments; then
47# code needed for everywhere, such as debugging stuff; then
48# code to handle input parameters; then
49# data structures likely to be of external interest (some of which depend on
50# the input parameters, so follows them; then
51# more data structures and subroutine and package (class) definitions; then
52# the small actual loop to process the input files and finish up; then
53# a __DATA__ section, for the .t tests
54#
5f7264c7 55# This program works on all releases of Unicode through at least 6.0. The
99870f4d
KW
56# outputs have been scrutinized most intently for release 5.1. The others
57# have been checked for somewhat more than just sanity. It can handle all
58# existing Unicode character properties in those releases.
59#
99870f4d
KW
60# This program is mostly about Unicode character (or code point) properties.
61# A property describes some attribute or quality of a code point, like if it
62# is lowercase or not, its name, what version of Unicode it was first defined
63# in, or what its uppercase equivalent is. Unicode deals with these disparate
64# possibilities by making all properties into mappings from each code point
65# into some corresponding value. In the case of it being lowercase or not,
66# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
67# property maps each Unicode code point to a single value, called a "property
68# value". (Hence each Unicode property is a true mathematical function with
69# exactly one value per code point.)
70#
71# When using a property in a regular expression, what is desired isn't the
72# mapping of the code point to its property's value, but the reverse (or the
73# mathematical "inverse relation"): starting with the property value, "Does a
74# code point map to it?" These are written in a "compound" form:
75# \p{property=value}, e.g., \p{category=punctuation}. This program generates
76# files containing the lists of code points that map to each such regular
77# expression property value, one file per list
78#
79# There is also a single form shortcut that Perl adds for many of the commonly
80# used properties. This happens for all binary properties, plus script,
81# general_category, and block properties.
82#
83# Thus the outputs of this program are files. There are map files, mostly in
84# the 'To' directory; and there are list files for use in regular expression
85# matching, all in subdirectories of the 'lib' directory, with each
86# subdirectory being named for the property that the lists in it are for.
87# Bookkeeping, test, and documentation files are also generated.
88
89my $matches_directory = 'lib'; # Where match (\p{}) files go.
90my $map_directory = 'To'; # Where map files go.
91
92# DATA STRUCTURES
93#
94# The major data structures of this program are Property, of course, but also
95# Table. There are two kinds of tables, very similar to each other.
96# "Match_Table" is the data structure giving the list of code points that have
97# a particular property value, mentioned above. There is also a "Map_Table"
98# data structure which gives the property's mapping from code point to value.
99# There are two structures because the match tables need to be combined in
100# various ways, such as constructing unions, intersections, complements, etc.,
101# and the map ones don't. And there would be problems, perhaps subtle, if
102# a map table were inadvertently operated on in some of those ways.
103# The use of separate classes with operations defined on one but not the other
104# prevents accidentally confusing the two.
105#
106# At the heart of each table's data structure is a "Range_List", which is just
107# an ordered list of "Ranges", plus ancillary information, and methods to
108# operate on them. A Range is a compact way to store property information.
109# Each range has a starting code point, an ending code point, and a value that
110# is meant to apply to all the code points between the two end points,
111# inclusive. For a map table, this value is the property value for those
112# code points. Two such ranges could be written like this:
113# 0x41 .. 0x5A, 'Upper',
114# 0x61 .. 0x7A, 'Lower'
115#
116# Each range also has a type used as a convenience to classify the values.
117# Most ranges in this program will be Type 0, or normal, but there are some
118# ranges that have a non-zero type. These are used only in map tables, and
119# are for mappings that don't fit into the normal scheme of things. Mappings
120# that require a hash entry to communicate with utf8.c are one example;
121# another example is mappings for charnames.pm to use which indicate a name
122# that is algorithmically determinable from its code point (and vice-versa).
123# These are used to significantly compact these tables, instead of listing
124# each one of the tens of thousands individually.
125#
126# In a match table, the value of a range is irrelevant (and hence the type as
127# well, which will always be 0), and arbitrarily set to the null string.
128# Using the example above, there would be two match tables for those two
129# entries, one named Upper would contain the 0x41..0x5A range, and the other
130# named Lower would contain 0x61..0x7A.
131#
132# Actually, there are two types of range lists, "Range_Map" is the one
133# associated with map tables, and "Range_List" with match tables.
134# Again, this is so that methods can be defined on one and not the other so as
135# to prevent operating on them in incorrect ways.
136#
137# Eventually, most tables are written out to files to be read by utf8_heavy.pl
138# in the perl core. All tables could in theory be written, but some are
139# suppressed because there is no current practical use for them. It is easy
140# to change which get written by changing various lists that are near the top
141# of the actual code in this file. The table data structures contain enough
142# ancillary information to allow them to be treated as separate entities for
143# writing, such as the path to each one's file. There is a heading in each
144# map table that gives the format of its entries, and what the map is for all
145# the code points missing from it. (This allows tables to be more compact.)
678f13d5 146#
99870f4d
KW
147# The Property data structure contains one or more tables. All properties
148# contain a map table (except the $perl property which is a
149# pseudo-property containing only match tables), and any properties that
150# are usable in regular expression matches also contain various matching
151# tables, one for each value the property can have. A binary property can
152# have two values, True and False (or Y and N, which are preferred by Unicode
153# terminology). Thus each of these properties will have a map table that
154# takes every code point and maps it to Y or N (but having ranges cuts the
155# number of entries in that table way down), and two match tables, one
156# which has a list of all the code points that map to Y, and one for all the
157# code points that map to N. (For each of these, a third table is also
158# generated for the pseudo Perl property. It contains the identical code
159# points as the Y table, but can be written, not in the compound form, but in
160# a "single" form like \p{IsUppercase}.) Many properties are binary, but some
161# properties have several possible values, some have many, and properties like
162# Name have a different value for every named code point. Those will not,
163# unless the controlling lists are changed, have their match tables written
164# out. But all the ones which can be used in regular expression \p{} and \P{}
165# constructs will. Generally a property will have either its map table or its
166# match tables written but not both. Again, what gets written is controlled
dc85bd38
KW
167# by lists which can easily be changed. Properties have a 'Type', like
168# binary, or string, or enum depending on how many match tables there are and
169# the content of the maps. This 'Type' is different than a range 'Type', so
170# don't get confused by the two concepts having the same name.
678f13d5 171#
99870f4d
KW
172# For information about the Unicode properties, see Unicode's UAX44 document:
173
174my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
175
176# As stated earlier, this program will work on any release of Unicode so far.
177# Most obvious problems in earlier data have NOT been corrected except when
178# necessary to make Perl or this program work reasonably. For example, no
179# folding information was given in early releases, so this program uses the
180# substitute of lower case, just so that a regular expression with the /i
181# option will do something that actually gives the right results in many
182# cases. There are also a couple other corrections for version 1.1.5,
183# commented at the point they are made. As an example of corrections that
184# weren't made (but could be) is this statement from DerivedAge.txt: "The
185# supplementary private use code points and the non-character code points were
186# assigned in version 2.0, but not specifically listed in the UCD until
187# versions 3.0 and 3.1 respectively." (To be precise it was 3.0.1 not 3.0.0)
188# More information on Unicode version glitches is further down in these
189# introductory comments.
190#
5f7264c7
KW
191# This program works on all non-provisional properties as of 6.0, though the
192# files for some are suppressed from apparent lack of demand for them. You
193# can change which are output by changing lists in this program.
678f13d5 194#
dc85bd38 195# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
196# loose matchings rules (from Unicode TR18):
197#
198# The recommended names for UCD properties and property values are in
199# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
200# [PropValue]. There are both abbreviated names and longer, more
201# descriptive names. It is strongly recommended that both names be
202# recognized, and that loose matching of property names be used,
203# whereby the case distinctions, whitespace, hyphens, and underbar
204# are ignored.
205# The program still allows Fuzzy to override its determination of if loose
206# matching should be used, but it isn't currently used, as it is no longer
207# needed; the calculations it makes are good enough.
678f13d5 208#
99870f4d
KW
209# SUMMARY OF HOW IT WORKS:
210#
211# Process arguments
212#
213# A list is constructed containing each input file that is to be processed
214#
215# Each file on the list is processed in a loop, using the associated handler
216# code for each:
217# The PropertyAliases.txt and PropValueAliases.txt files are processed
218# first. These files name the properties and property values.
219# Objects are created of all the property and property value names
220# that the rest of the input should expect, including all synonyms.
221# The other input files give mappings from properties to property
222# values. That is, they list code points and say what the mapping
223# is under the given property. Some files give the mappings for
224# just one property; and some for many. This program goes through
225# each file and populates the properties from them. Some properties
226# are listed in more than one file, and Unicode has set up a
227# precedence as to which has priority if there is a conflict. Thus
228# the order of processing matters, and this program handles the
229# conflict possibility by processing the overriding input files
230# last, so that if necessary they replace earlier values.
231# After this is all done, the program creates the property mappings not
232# furnished by Unicode, but derivable from what it does give.
233# The tables of code points that match each property value in each
234# property that is accessible by regular expressions are created.
235# The Perl-defined properties are created and populated. Many of these
236# require data determined from the earlier steps
237# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 238# and Unicode are reconciled and warned about.
99870f4d
KW
239# All the properties are written to files
240# Any other files are written, and final warnings issued.
678f13d5 241#
99870f4d
KW
242# For clarity, a number of operators have been overloaded to work on tables:
243# ~ means invert (take all characters not in the set). The more
244# conventional '!' is not used because of the possibility of confusing
245# it with the actual boolean operation.
246# + means union
247# - means subtraction
248# & means intersection
249# The precedence of these is the order listed. Parentheses should be
250# copiously used. These are not a general scheme. The operations aren't
251# defined for a number of things, deliberately, to avoid getting into trouble.
252# Operations are done on references and affect the underlying structures, so
253# that the copy constructors for them have been overloaded to not return a new
254# clone, but the input object itself.
678f13d5 255#
99870f4d
KW
256# The bool operator is deliberately not overloaded to avoid confusion with
257# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
258#
259# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
260#
261# This program needs to be able to run under miniperl. Therefore, it uses a
262# minimum of other modules, and hence implements some things itself that could
263# be gotten from CPAN
264#
265# This program uses inputs published by the Unicode Consortium. These can
266# change incompatibly between releases without the Perl maintainers realizing
267# it. Therefore this program is now designed to try to flag these. It looks
268# at the directories where the inputs are, and flags any unrecognized files.
269# It keeps track of all the properties in the files it handles, and flags any
270# that it doesn't know how to handle. It also flags any input lines that
271# don't match the expected syntax, among other checks.
272#
273# It is also designed so if a new input file matches one of the known
274# templates, one hopefully just needs to add it to a list to have it
275# processed.
276#
277# As mentioned earlier, some properties are given in more than one file. In
278# particular, the files in the extracted directory are supposedly just
279# reformattings of the others. But they contain information not easily
280# derivable from the other files, including results for Unihan, which this
281# program doesn't ordinarily look at, and for unassigned code points. They
282# also have historically had errors or been incomplete. In an attempt to
283# create the best possible data, this program thus processes them first to
284# glean information missing from the other files; then processes those other
285# files to override any errors in the extracted ones. Much of the design was
286# driven by this need to store things and then possibly override them.
287#
288# It tries to keep fatal errors to a minimum, to generate something usable for
289# testing purposes. It always looks for files that could be inputs, and will
290# warn about any that it doesn't know how to handle (the -q option suppresses
291# the warning).
99870f4d
KW
292#
293# Why have files written out for binary 'N' matches?
294# For binary properties, if you know the mapping for either Y or N; the
678f13d5
KW
295# other is trivial to construct, so could be done at Perl run-time by just
296# complementing the result, instead of having a file for it. That is, if
297# someone types in \p{foo: N}, Perl could translate that to \P{foo: Y} and
298# not need a file. The problem is communicating to Perl that a given
299# property is binary. Perl can't figure it out from looking at the N (or
300# No), as some non-binary properties have these as property values. So
301# rather than inventing a way to communicate this info back to the core,
302# which would have required changes there as well, it was simpler just to
303# add the extra tables.
304#
305# Why is there more than one type of range?
306# This simplified things. There are some very specialized code points that
307# have to be handled specially for output, such as Hangul syllable names.
308# By creating a range type (done late in the development process), it
309# allowed this to be stored with the range, and overridden by other input.
310# Originally these were stored in another data structure, and it became a
311# mess trying to decide if a second file that was for the same property was
312# overriding the earlier one or not.
313#
314# Why are there two kinds of tables, match and map?
315# (And there is a base class shared by the two as well.) As stated above,
316# they actually are for different things. Development proceeded much more
317# smoothly when I (khw) realized the distinction. Map tables are used to
318# give the property value for every code point (actually every code point
319# that doesn't map to a default value). Match tables are used for regular
320# expression matches, and are essentially the inverse mapping. Separating
321# the two allows more specialized methods, and error checks so that one
322# can't just take the intersection of two map tables, for example, as that
323# is nonsensical.
99870f4d
KW
324#
325# There are no match tables generated for matches of the null string. These
c1739a4a 326# would look like qr/\p{JSN=}/ currently without modifying the regex code.
678f13d5
KW
327# Perhaps something like them could be added if necessary. The JSN does have
328# a real code point U+110B that maps to the null string, but it is a
329# contributory property, and therefore not output by default. And it's easily
330# handled so far by making the null string the default where it is a
331# possibility.
99870f4d 332#
23e33b60
KW
333# DEBUGGING
334#
678f13d5
KW
335# This program is written so it will run under miniperl. Occasionally changes
336# will cause an error where the backtrace doesn't work well under miniperl.
337# To diagnose the problem, you can instead run it under regular perl, if you
338# have one compiled.
339#
340# There is a good trace facility. To enable it, first sub DEBUG must be set
341# to return true. Then a line like
342#
343# local $to_trace = 1 if main::DEBUG;
344#
345# can be added to enable tracing in its lexical scope or until you insert
346# another line:
347#
348# local $to_trace = 0 if main::DEBUG;
349#
350# then use a line like "trace $a, @b, %c, ...;
351#
352# Some of the more complex subroutines already have trace statements in them.
353# Permanent trace statements should be like:
354#
355# trace ... if main::DEBUG && $to_trace;
356#
357# If there is just one or a few files that you're debugging, you can easily
358# cause most everything else to be skipped. Change the line
359#
360# my $debug_skip = 0;
361#
362# to 1, and every file whose object is in @input_file_objects and doesn't have
363# a, 'non_skip => 1,' in its constructor will be skipped.
364#
b4a0206c 365# To compare the output tables, it may be useful to specify the -annotate
c4019d52
KW
366# flag. This causes the tables to expand so there is one entry for each
367# non-algorithmically named code point giving, currently its name, and its
368# graphic representation if printable (and you have a font that knows about
369# it). This makes it easier to see what the particular code points are in
370# each output table. The tables are usable, but because they don't have
371# ranges (for the most part), a Perl using them will run slower. Non-named
372# code points are annotated with a description of their status, and contiguous
373# ones with the same description will be output as a range rather than
374# individually. Algorithmically named characters are also output as ranges,
375# except when there are just a few contiguous ones.
376#
99870f4d
KW
377# FUTURE ISSUES
378#
379# The program would break if Unicode were to change its names so that
380# interior white space, underscores, or dashes differences were significant
381# within property and property value names.
382#
383# It might be easier to use the xml versions of the UCD if this program ever
384# would need heavy revision, and the ability to handle old versions was not
385# required.
386#
387# There is the potential for name collisions, in that Perl has chosen names
388# that Unicode could decide it also likes. There have been such collisions in
389# the past, with mostly Perl deciding to adopt the Unicode definition of the
390# name. However in the 5.2 Unicode beta testing, there were a number of such
391# collisions, which were withdrawn before the final release, because of Perl's
392# and other's protests. These all involved new properties which began with
393# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
394# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
395# Unicode document, so they are unlikely to be used by Unicode for another
396# purpose. However, they might try something beginning with 'In', or use any
397# of the other Perl-defined properties. This program will warn you of name
398# collisions, and refuse to generate tables with them, but manual intervention
399# will be required in this event. One scheme that could be implemented, if
400# necessary, would be to have this program generate another file, or add a
401# field to mktables.lst that gives the date of first definition of a property.
402# Each new release of Unicode would use that file as a basis for the next
403# iteration. And the Perl synonym addition code could sort based on the age
404# of the property, so older properties get priority, and newer ones that clash
405# would be refused; hence existing code would not be impacted, and some other
406# synonym would have to be used for the new property. This is ugly, and
407# manual intervention would certainly be easier to do in the short run; lets
408# hope it never comes to this.
678f13d5 409#
99870f4d
KW
410# A NOTE ON UNIHAN
411#
412# This program can generate tables from the Unihan database. But it doesn't
413# by default, letting the CPAN module Unicode::Unihan handle them. Prior to
414# version 5.2, this database was in a single file, Unihan.txt. In 5.2 the
415# database was split into 8 different files, all beginning with the letters
416# 'Unihan'. This program will read those file(s) if present, but it needs to
417# know which of the many properties in the file(s) should have tables created
418# for them. It will create tables for any properties listed in
419# PropertyAliases.txt and PropValueAliases.txt, plus any listed in the
420# @cjk_properties array and the @cjk_property_values array. Thus, if a
421# property you want is not in those files of the release you are building
422# against, you must add it to those two arrays. Starting in 4.0, the
423# Unicode_Radical_Stroke was listed in those files, so if the Unihan database
424# is present in the directory, a table will be generated for that property.
425# In 5.2, several more properties were added. For your convenience, the two
5f7264c7 426# arrays are initialized with all the 6.0 listed properties that are also in
99870f4d
KW
427# earlier releases. But these are commented out. You can just uncomment the
428# ones you want, or use them as a template for adding entries for other
429# properties.
430#
431# You may need to adjust the entries to suit your purposes. setup_unihan(),
432# and filter_unihan_line() are the functions where this is done. This program
433# already does some adjusting to make the lines look more like the rest of the
434# Unicode DB; You can see what that is in filter_unihan_line()
435#
436# There is a bug in the 3.2 data file in which some values for the
437# kPrimaryNumeric property have commas and an unexpected comment. A filter
438# could be added for these; or for a particular installation, the Unihan.txt
439# file could be edited to fix them.
99870f4d 440#
678f13d5
KW
441# HOW TO ADD A FILE TO BE PROCESSED
442#
443# A new file from Unicode needs to have an object constructed for it in
444# @input_file_objects, probably at the end or at the end of the extracted
445# ones. The program should warn you if its name will clash with others on
446# restrictive file systems, like DOS. If so, figure out a better name, and
447# add lines to the README.perl file giving that. If the file is a character
448# property, it should be in the format that Unicode has by default
449# standardized for such files for the more recently introduced ones.
450# If so, the Input_file constructor for @input_file_objects can just be the
451# file name and release it first appeared in. If not, then it should be
452# possible to construct an each_line_handler() to massage the line into the
453# standardized form.
454#
455# For non-character properties, more code will be needed. You can look at
456# the existing entries for clues.
457#
458# UNICODE VERSIONS NOTES
459#
460# The Unicode UCD has had a number of errors in it over the versions. And
461# these remain, by policy, in the standard for that version. Therefore it is
462# risky to correct them, because code may be expecting the error. So this
463# program doesn't generally make changes, unless the error breaks the Perl
464# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
465# for U+1105, which causes real problems for the algorithms for Jamo
466# calculations, so it is changed here.
467#
468# But it isn't so clear cut as to what to do about concepts that are
469# introduced in a later release; should they extend back to earlier releases
470# where the concept just didn't exist? It was easier to do this than to not,
471# so that's what was done. For example, the default value for code points not
472# in the files for various properties was probably undefined until changed by
473# some version. No_Block for blocks is such an example. This program will
474# assign No_Block even in Unicode versions that didn't have it. This has the
475# benefit that code being written doesn't have to special case earlier
476# versions; and the detriment that it doesn't match the Standard precisely for
477# the affected versions.
478#
479# Here are some observations about some of the issues in early versions:
480#
6426c51b 481# The number of code points in \p{alpha} halved in 2.1.9. It turns out that
678f13d5
KW
482# the reason is that the CJK block starting at 4E00 was removed from PropList,
483# and was not put back in until 3.1.0
484#
485# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
486# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
487# reason is that 3.2 introduced U+205F=medium math space, which was not
488# classed as white space, but Perl figured out that it should have been. 4.0
489# reclassified it correctly.
490#
491# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
492# this was erroneously a synonym for 202. In 4.0, ATB became 202, and ATBL
493# was left with no code points, as all the ones that mapped to 202 stayed
494# mapped to 202. Thus if your program used the numeric name for the class,
495# it would not have been affected, but if it used the mnemonic, it would have
496# been.
497#
498# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that code
499# points which eventually came to have this script property value, instead
500# mapped to "Unknown". But in the next release all these code points were
501# moved to \p{sc=common} instead.
99870f4d
KW
502#
503# The default for missing code points for BidiClass is complicated. Starting
504# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
505# tries to do the best it can for earlier releases. It is done in
506# process_PropertyAliases()
507#
508##############################################################################
509
510my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
511 # and errors
512my $MAX_LINE_WIDTH = 78;
513
514# Debugging aid to skip most files so as to not be distracted by them when
515# concentrating on the ones being debugged. Add
516# non_skip => 1,
517# to the constructor for those files you want processed when you set this.
518# Files with a first version number of 0 are special: they are always
519# processed regardless of the state of this flag.
520my $debug_skip = 0;
521
522# Set to 1 to enable tracing.
523our $to_trace = 0;
524
525{ # Closure for trace: debugging aid
526 my $print_caller = 1; # ? Include calling subroutine name
527 my $main_with_colon = 'main::';
528 my $main_colon_length = length($main_with_colon);
529
530 sub trace {
531 return unless $to_trace; # Do nothing if global flag not set
532
533 my @input = @_;
534
535 local $DB::trace = 0;
536 $DB::trace = 0; # Quiet 'used only once' message
537
538 my $line_number;
539
540 # Loop looking up the stack to get the first non-trace caller
541 my $caller_line;
542 my $caller_name;
543 my $i = 0;
544 do {
545 $line_number = $caller_line;
546 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
547 $caller = $main_with_colon unless defined $caller;
548
549 $caller_name = $caller;
550
551 # get rid of pkg
552 $caller_name =~ s/.*:://;
553 if (substr($caller_name, 0, $main_colon_length)
554 eq $main_with_colon)
555 {
556 $caller_name = substr($caller_name, $main_colon_length);
557 }
558
559 } until ($caller_name ne 'trace');
560
561 # If the stack was empty, we were called from the top level
562 $caller_name = 'main' if ($caller_name eq ""
563 || $caller_name eq 'trace');
564
565 my $output = "";
566 foreach my $string (@input) {
567 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
568 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
569 $output .= simple_dumper($string);
570 }
571 else {
572 $string = "$string" if ref $string;
573 $string = $UNDEF unless defined $string;
574 chomp $string;
575 $string = '""' if $string eq "";
576 $output .= " " if $output ne ""
577 && $string ne ""
578 && substr($output, -1, 1) ne " "
579 && substr($string, 0, 1) ne " ";
580 $output .= $string;
581 }
582 }
583
99f78760
KW
584 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
585 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
586 print STDERR $output, "\n";
587 return;
588 }
589}
590
591# This is for a rarely used development feature that allows you to compare two
592# versions of the Unicode standard without having to deal with changes caused
593# by the code points introduced in the later verson. Change the 0 to a SINGLE
594# dotted Unicode release number (e.g. 2.1). Only code points introduced in
595# that release and earlier will be used; later ones are thrown away. You use
596# the version number of the earliest one you want to compare; then run this
597# program on directory structures containing each release, and compare the
598# outputs. These outputs will therefore include only the code points common
599# to both releases, and you can see the changes caused just by the underlying
600# release semantic changes. For versions earlier than 3.2, you must copy a
601# version of DAge.txt into the directory.
602my $string_compare_versions = DEBUG && 0; # e.g., v2.1;
603my $compare_versions = DEBUG
604 && $string_compare_versions
605 && pack "C*", split /\./, $string_compare_versions;
606
607sub uniques {
608 # Returns non-duplicated input values. From "Perl Best Practices:
609 # Encapsulated Cleverness". p. 455 in first edition.
610
611 my %seen;
0e407844
NC
612 # Arguably this breaks encapsulation, if the goal is to permit multiple
613 # distinct objects to stringify to the same value, and be interchangeable.
614 # However, for this program, no two objects stringify identically, and all
615 # lists passed to this function are either objects or strings. So this
616 # doesn't affect correctness, but it does give a couple of percent speedup.
617 no overloading;
99870f4d
KW
618 return grep { ! $seen{$_}++ } @_;
619}
620
621$0 = File::Spec->canonpath($0);
622
623my $make_test_script = 0; # ? Should we output a test script
624my $write_unchanged_files = 0; # ? Should we update the output files even if
625 # we don't think they have changed
626my $use_directory = ""; # ? Should we chdir somewhere.
627my $pod_directory; # input directory to store the pod file.
628my $pod_file = 'perluniprops';
629my $t_path; # Path to the .t test file
630my $file_list = 'mktables.lst'; # File to store input and output file names.
631 # This is used to speed up the build, by not
632 # executing the main body of the program if
633 # nothing on the list has changed since the
634 # previous build
635my $make_list = 1; # ? Should we write $file_list. Set to always
636 # make a list so that when the pumpking is
637 # preparing a release, s/he won't have to do
638 # special things
639my $glob_list = 0; # ? Should we try to include unknown .txt files
640 # in the input.
bd9ebcfd
KW
641my $output_range_counts = $debugging_build; # ? Should we include the number
642 # of code points in ranges in
643 # the output
558712cf 644my $annotate = 0; # ? Should character names be in the output
9ef2b94f 645
99870f4d
KW
646# Verbosity levels; 0 is quiet
647my $NORMAL_VERBOSITY = 1;
648my $PROGRESS = 2;
649my $VERBOSE = 3;
650
651my $verbosity = $NORMAL_VERBOSITY;
652
653# Process arguments
654while (@ARGV) {
cf25bb62
JH
655 my $arg = shift @ARGV;
656 if ($arg eq '-v') {
99870f4d
KW
657 $verbosity = $VERBOSE;
658 }
659 elsif ($arg eq '-p') {
660 $verbosity = $PROGRESS;
661 $| = 1; # Flush buffers as we go.
662 }
663 elsif ($arg eq '-q') {
664 $verbosity = 0;
665 }
666 elsif ($arg eq '-w') {
667 $write_unchanged_files = 1; # update the files even if havent changed
668 }
669 elsif ($arg eq '-check') {
6ae7e459
YO
670 my $this = shift @ARGV;
671 my $ok = shift @ARGV;
672 if ($this ne $ok) {
673 print "Skipping as check params are not the same.\n";
674 exit(0);
675 }
00a8df5c 676 }
99870f4d
KW
677 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
678 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
679 }
3df51b85
KW
680 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
681 {
99870f4d 682 $make_test_script = 1;
99870f4d
KW
683 }
684 elsif ($arg eq '-makelist') {
685 $make_list = 1;
686 }
687 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
688 -d $use_directory or croak "Unknown directory '$use_directory'";
689 }
690 elsif ($arg eq '-L') {
691
692 # Existence not tested until have chdir'd
693 $file_list = shift;
694 }
695 elsif ($arg eq '-globlist') {
696 $glob_list = 1;
697 }
698 elsif ($arg eq '-c') {
699 $output_range_counts = ! $output_range_counts
700 }
b4a0206c 701 elsif ($arg eq '-annotate') {
558712cf 702 $annotate = 1;
bd9ebcfd
KW
703 $debugging_build = 1;
704 $output_range_counts = 1;
9ef2b94f 705 }
99870f4d
KW
706 else {
707 my $with_c = 'with';
708 $with_c .= 'out' if $output_range_counts; # Complements the state
709 croak <<END;
710usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
711 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
712 [-check A B ]
713 -c : Output comments $with_c number of code points in ranges
714 -q : Quiet Mode: Only output serious warnings.
715 -p : Set verbosity level to normal plus show progress.
716 -v : Set Verbosity level high: Show progress and non-serious
717 warnings
718 -w : Write files regardless
719 -C dir : Change to this directory before proceeding. All relative paths
720 except those specified by the -P and -T options will be done
721 with respect to this directory.
722 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 723 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
724 -L filelist : Use alternate 'filelist' instead of standard one
725 -globlist : Take as input all non-Test *.txt files in current and sub
726 directories
3df51b85
KW
727 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
728 overrides -T
99870f4d 729 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 730 -annotate : Output an annotation for each character in the table files;
c4019d52
KW
731 useful for debugging mktables, looking at diffs; but is slow,
732 memory intensive; resulting tables are usable but slow and
733 very large.
99870f4d
KW
734 -check A B : Executes $0 only if A and B are the same
735END
736 }
737}
738
739# Stores the most-recently changed file. If none have changed, can skip the
740# build
aeab6150 741my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
742
743# Change directories now, because need to read 'version' early.
744if ($use_directory) {
3df51b85 745 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
746 $pod_directory = File::Spec->rel2abs($pod_directory);
747 }
3df51b85 748 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 749 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 750 }
99870f4d 751 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 752 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 753 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 754 }
3df51b85 755 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 756 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 757 }
00a8df5c
YO
758}
759
99870f4d
KW
760# Get Unicode version into regular and v-string. This is done now because
761# various tables below get populated based on it. These tables are populated
762# here to be near the top of the file, and so easily seeable by those needing
763# to modify things.
764open my $VERSION, "<", "version"
765 or croak "$0: can't open required file 'version': $!\n";
766my $string_version = <$VERSION>;
767close $VERSION;
768chomp $string_version;
769my $v_version = pack "C*", split /\./, $string_version; # v string
770
771# The following are the complete names of properties with property values that
772# are known to not match any code points in some versions of Unicode, but that
773# may change in the future so they should be matchable, hence an empty file is
774# generated for them.
775my @tables_that_may_be_empty = (
776 'Joining_Type=Left_Joining',
777 );
778push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
779push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
780push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
781 if $v_version ge v4.1.0;
782
783# The lists below are hashes, so the key is the item in the list, and the
784# value is the reason why it is in the list. This makes generation of
785# documentation easier.
786
787my %why_suppressed; # No file generated for these.
788
789# Files aren't generated for empty extraneous properties. This is arguable.
790# Extraneous properties generally come about because a property is no longer
791# used in a newer version of Unicode. If we generated a file without code
792# points, programs that used to work on that property will still execute
793# without errors. It just won't ever match (or will always match, with \P{}).
794# This means that the logic is now likely wrong. I (khw) think its better to
795# find this out by getting an error message. Just move them to the table
796# above to change this behavior
797my %why_suppress_if_empty_warn_if_not = (
798
799 # It is the only property that has ever officially been removed from the
800 # Standard. The database never contained any code points for it.
801 'Special_Case_Condition' => 'Obsolete',
802
803 # Apparently never official, but there were code points in some versions of
804 # old-style PropList.txt
805 'Non_Break' => 'Obsolete',
806);
807
808# These would normally go in the warn table just above, but they were changed
809# a long time before this program was written, so warnings about them are
810# moot.
811if ($v_version gt v3.2.0) {
812 push @tables_that_may_be_empty,
813 'Canonical_Combining_Class=Attached_Below_Left'
814}
815
5f7264c7 816# These are listed in the Property aliases file in 6.0, but Unihan is ignored
99870f4d
KW
817# unless explicitly added.
818if ($v_version ge v5.2.0) {
819 my $unihan = 'Unihan; remove from list if using Unihan';
ea25a9b2 820 foreach my $table (qw (
99870f4d
KW
821 kAccountingNumeric
822 kOtherNumeric
823 kPrimaryNumeric
824 kCompatibilityVariant
825 kIICore
826 kIRG_GSource
827 kIRG_HSource
828 kIRG_JSource
829 kIRG_KPSource
830 kIRG_MSource
831 kIRG_KSource
832 kIRG_TSource
833 kIRG_USource
834 kIRG_VSource
835 kRSUnicode
ea25a9b2 836 ))
99870f4d
KW
837 {
838 $why_suppress_if_empty_warn_if_not{$table} = $unihan;
839 }
ca12659b
NC
840}
841
272501f6
KW
842# Enum values for to_output_map() method in the Map_Table package.
843my $EXTERNAL_MAP = 1;
844my $INTERNAL_MAP = 2;
845
fcf1973c
KW
846# To override computed values for writing the map tables for these properties.
847# The default for enum map tables is to write them out, so that the Unicode
848# .txt files can be removed, but all the data to compute any property value
849# for any code point is available in a more compact form.
850my %global_to_output_map = (
851 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
852 # get stuck supporting it if things change. Sinc it is a STRING property,
853 # it normally would be listed in the pod, but INTERNAL_MAP suppresses
854 # that.
855 Unicode_1_Name => $INTERNAL_MAP,
856
857 Present_In => 0, # Suppress, as easily computed from Age
858 Canonical_Combining_Class => 0, # Duplicate of CombiningClass.pl
859 Block => 0, # Suppress, as Blocks.txt is retained.
860);
861
99870f4d
KW
862# Properties that this program ignores.
863my @unimplemented_properties = (
864'Unicode_Radical_Stroke' # Remove if changing to handle this one.
865);
d73e5302 866
99870f4d
KW
867# There are several types of obsolete properties defined by Unicode. These
868# must be hand-edited for every new Unicode release.
869my %why_deprecated; # Generates a deprecated warning message if used.
870my %why_stabilized; # Documentation only
871my %why_obsolete; # Documentation only
872
873{ # Closure
874 my $simple = 'Perl uses the more complete version of this property';
875 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
876
877 my $other_properties = 'other properties';
878 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 879 my $why_no_expand = "Deprecated by Unicode. These are characters that expand to more than one character in the specified normalization form, but whether they actually take up more bytes or not depends on the encoding being used. For example, a UTF-8 encoded character may expand to a different number of bytes than a UTF-32 encoded character.";
99870f4d
KW
880
881 %why_deprecated = (
5f7264c7 882 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
883 'Jamo_Short_Name' => $contributory,
884 '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',
885 'Other_Alphabetic' => $contributory,
886 'Other_Default_Ignorable_Code_Point' => $contributory,
887 'Other_Grapheme_Extend' => $contributory,
888 'Other_ID_Continue' => $contributory,
889 'Other_ID_Start' => $contributory,
890 'Other_Lowercase' => $contributory,
891 'Other_Math' => $contributory,
892 'Other_Uppercase' => $contributory,
893 );
894
895 %why_suppressed = (
5f7264c7 896 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
897 # contains the same information, but without the algorithmically
898 # determinable Hangul syllables'. This file is not published, so it's
899 # existence is not noted in the comment.
900 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize',
901
902 'ISO_Comment' => 'Apparently no demand for it, but can access it through Unicode::UCD::charinfo. Obsoleted, and code points for it removed in Unicode 5.2',
99870f4d
KW
903
904 'Simple_Case_Folding' => "$simple. Can access this through Unicode::UCD::casefold",
905 'Simple_Lowercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
906 'Simple_Titlecase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
907 'Simple_Uppercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
908
909 'Name' => "Accessible via 'use charnames;'",
910 'Name_Alias' => "Accessible via 'use charnames;'",
911
5f7264c7 912 FC_NFKC_Closure => 'Supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
913 Expands_On_NFC => $why_no_expand,
914 Expands_On_NFD => $why_no_expand,
915 Expands_On_NFKC => $why_no_expand,
916 Expands_On_NFKD => $why_no_expand,
917 );
918
919 # The following are suppressed because they were made contributory or
920 # deprecated by Unicode before Perl ever thought about supporting them.
921 foreach my $property ('Jamo_Short_Name', 'Grapheme_Link') {
922 $why_suppressed{$property} = $why_deprecated{$property};
923 }
cf25bb62 924
99870f4d
KW
925 # Customize the message for all the 'Other_' properties
926 foreach my $property (keys %why_deprecated) {
927 next if (my $main_property = $property) !~ s/^Other_//;
928 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
929 }
930}
931
932if ($v_version ge 4.0.0) {
933 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
934 if ($v_version ge 6.0.0) {
935 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
936 }
99870f4d 937}
5f7264c7 938if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 939 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7
KW
940 if ($v_version ge 6.0.0) {
941 $why_deprecated{'ISO_Comment'} = 'No longer needed for chart generation; otherwise not useful, and code points for it have been removed';
942 }
99870f4d
KW
943}
944
945# Probably obsolete forever
946if ($v_version ge v4.1.0) {
947 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common"';
948}
949
950# This program can create files for enumerated-like properties, such as
951# 'Numeric_Type'. This file would be the same format as for a string
952# property, with a mapping from code point to its value, so you could look up,
953# for example, the script a code point is in. But no one so far wants this
954# mapping, or they have found another way to get it since this is a new
955# feature. So no file is generated except if it is in this list.
956my @output_mapped_properties = split "\n", <<END;
957END
958
959# If you are using the Unihan database, you need to add the properties that
960# you want to extract from it to this table. For your convenience, the
5f7264c7 961# properties in the 6.0 PropertyAliases.txt file are listed, commented out
99870f4d
KW
962my @cjk_properties = split "\n", <<'END';
963#cjkAccountingNumeric; kAccountingNumeric
964#cjkOtherNumeric; kOtherNumeric
965#cjkPrimaryNumeric; kPrimaryNumeric
966#cjkCompatibilityVariant; kCompatibilityVariant
967#cjkIICore ; kIICore
968#cjkIRG_GSource; kIRG_GSource
969#cjkIRG_HSource; kIRG_HSource
970#cjkIRG_JSource; kIRG_JSource
971#cjkIRG_KPSource; kIRG_KPSource
972#cjkIRG_KSource; kIRG_KSource
973#cjkIRG_TSource; kIRG_TSource
974#cjkIRG_USource; kIRG_USource
975#cjkIRG_VSource; kIRG_VSource
976#cjkRSUnicode; kRSUnicode ; Unicode_Radical_Stroke; URS
977END
978
979# Similarly for the property values. For your convenience, the lines in the
5f7264c7 980# 6.0 PropertyAliases.txt file are listed. Just remove the first BUT NOT both
99870f4d
KW
981# '#' marks
982my @cjk_property_values = split "\n", <<'END';
983## @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
984## @missing: 0000..10FFFF; cjkCompatibilityVariant; <code point>
985## @missing: 0000..10FFFF; cjkIICore; <none>
986## @missing: 0000..10FFFF; cjkIRG_GSource; <none>
987## @missing: 0000..10FFFF; cjkIRG_HSource; <none>
988## @missing: 0000..10FFFF; cjkIRG_JSource; <none>
989## @missing: 0000..10FFFF; cjkIRG_KPSource; <none>
990## @missing: 0000..10FFFF; cjkIRG_KSource; <none>
991## @missing: 0000..10FFFF; cjkIRG_TSource; <none>
992## @missing: 0000..10FFFF; cjkIRG_USource; <none>
993## @missing: 0000..10FFFF; cjkIRG_VSource; <none>
994## @missing: 0000..10FFFF; cjkOtherNumeric; NaN
995## @missing: 0000..10FFFF; cjkPrimaryNumeric; NaN
996## @missing: 0000..10FFFF; cjkRSUnicode; <none>
997END
998
999# The input files don't list every code point. Those not listed are to be
1000# defaulted to some value. Below are hard-coded what those values are for
1001# non-binary properties as of 5.1. Starting in 5.0, there are
1002# machine-parsable comment lines in the files the give the defaults; so this
1003# list shouldn't have to be extended. The claim is that all missing entries
1004# for binary properties will default to 'N'. Unicode tried to change that in
1005# 5.2, but the beta period produced enough protest that they backed off.
1006#
1007# The defaults for the fields that appear in UnicodeData.txt in this hash must
1008# be in the form that it expects. The others may be synonyms.
1009my $CODE_POINT = '<code point>';
1010my %default_mapping = (
1011 Age => "Unassigned",
1012 # Bidi_Class => Complicated; set in code
1013 Bidi_Mirroring_Glyph => "",
1014 Block => 'No_Block',
1015 Canonical_Combining_Class => 0,
1016 Case_Folding => $CODE_POINT,
1017 Decomposition_Mapping => $CODE_POINT,
1018 Decomposition_Type => 'None',
1019 East_Asian_Width => "Neutral",
1020 FC_NFKC_Closure => $CODE_POINT,
1021 General_Category => 'Cn',
1022 Grapheme_Cluster_Break => 'Other',
1023 Hangul_Syllable_Type => 'NA',
1024 ISO_Comment => "",
1025 Jamo_Short_Name => "",
1026 Joining_Group => "No_Joining_Group",
1027 # Joining_Type => Complicated; set in code
1028 kIICore => 'N', # Is converted to binary
1029 #Line_Break => Complicated; set in code
1030 Lowercase_Mapping => $CODE_POINT,
1031 Name => "",
1032 Name_Alias => "",
1033 NFC_QC => 'Yes',
1034 NFD_QC => 'Yes',
1035 NFKC_QC => 'Yes',
1036 NFKD_QC => 'Yes',
1037 Numeric_Type => 'None',
1038 Numeric_Value => 'NaN',
1039 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1040 Sentence_Break => 'Other',
1041 Simple_Case_Folding => $CODE_POINT,
1042 Simple_Lowercase_Mapping => $CODE_POINT,
1043 Simple_Titlecase_Mapping => $CODE_POINT,
1044 Simple_Uppercase_Mapping => $CODE_POINT,
1045 Titlecase_Mapping => $CODE_POINT,
1046 Unicode_1_Name => "",
1047 Unicode_Radical_Stroke => "",
1048 Uppercase_Mapping => $CODE_POINT,
1049 Word_Break => 'Other',
1050);
1051
1052# Below are files that Unicode furnishes, but this program ignores, and why
1053my %ignored_files = (
1054 'CJKRadicals.txt' => 'Unihan data',
1055 'Index.txt' => 'An index, not actual data',
1056 'NamedSqProv.txt' => 'Not officially part of the Unicode standard; Append it to NamedSequences.txt if you want to process the contents.',
1057 'NamesList.txt' => 'Just adds commentary',
1058 'NormalizationCorrections.txt' => 'Data is already in other files.',
1059 'Props.txt' => 'Adds nothing to PropList.txt; only in very early releases',
1060 'ReadMe.txt' => 'Just comments',
1061 'README.TXT' => 'Just comments',
1062 'StandardizedVariants.txt' => 'Only for glyph changes, not a Unicode character property. Does not fit into current scheme where one code point is mapped',
5f7264c7
KW
1063 'EmojiSources.txt' => 'Not of general utility: for Japanese legacy cell-phone applications',
1064 'IndicMatraCategory.txt' => 'Provisional',
1065 'IndicSyllabicCategory.txt' => 'Provisional',
1066 'ScriptExtensions.txt' => 'Provisional',
99870f4d
KW
1067);
1068
678f13d5 1069### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1070
1071my $HEADER=<<"EOF";
1072# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85
KW
1073# This file is machine-generated by $0 from the Unicode
1074# database, Version $string_version. Any changes made here will be lost!
cf25bb62
JH
1075EOF
1076
b6922eda 1077my $INTERNAL_ONLY=<<"EOF";
99870f4d
KW
1078
1079# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
b6922eda 1080# This file is for internal use by the Perl program only. The format and even
99870f4d
KW
1081# the name or existence of this file are subject to change without notice.
1082# Don't use it directly.
1083EOF
1084
1085my $DEVELOPMENT_ONLY=<<"EOF";
1086# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1087# This file contains information artificially constrained to code points
1088# present in Unicode release $string_compare_versions.
1089# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1090# not be used for production.
b6922eda
KW
1091
1092EOF
1093
99870f4d
KW
1094my $LAST_UNICODE_CODEPOINT_STRING = "10FFFF";
1095my $LAST_UNICODE_CODEPOINT = hex $LAST_UNICODE_CODEPOINT_STRING;
1096my $MAX_UNICODE_CODEPOINTS = $LAST_UNICODE_CODEPOINT + 1;
1097
1098# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1099# two must be 10; if there are 5, the first must not be a 0. Written this way
1100# to decrease backtracking
1101my $code_point_re =
1102 qr/ \b (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1103
1104# This matches the beginning of the line in the Unicode db files that give the
1105# defaults for code points not listed (i.e., missing) in the file. The code
1106# depends on this ending with a semi-colon, so it can assume it is a valid
1107# field when the line is split() by semi-colons
1108my $missing_defaults_prefix =
1109 qr/^#\s+\@missing:\s+0000\.\.$LAST_UNICODE_CODEPOINT_STRING\s*;/;
1110
1111# Property types. Unicode has more types, but these are sufficient for our
1112# purposes.
1113my $UNKNOWN = -1; # initialized to illegal value
1114my $NON_STRING = 1; # Either binary or enum
1115my $BINARY = 2;
1116my $ENUM = 3; # Include catalog
1117my $STRING = 4; # Anything else: string or misc
1118
1119# Some input files have lines that give default values for code points not
1120# contained in the file. Sometimes these should be ignored.
1121my $NO_DEFAULTS = 0; # Must evaluate to false
1122my $NOT_IGNORED = 1;
1123my $IGNORED = 2;
1124
1125# Range types. Each range has a type. Most ranges are type 0, for normal,
1126# and will appear in the main body of the tables in the output files, but
1127# there are other types of ranges as well, listed below, that are specially
1128# handled. There are pseudo-types as well that will never be stored as a
1129# type, but will affect the calculation of the type.
1130
1131# 0 is for normal, non-specials
1132my $MULTI_CP = 1; # Sequence of more than code point
1133my $HANGUL_SYLLABLE = 2;
1134my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1135my $NULL = 4; # The map is to the null string; utf8.c can't
1136 # handle these, nor is there an accepted syntax
1137 # for them in \p{} constructs
f86864ac 1138my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1139 # otherwise be $MULTI_CP type are instead type 0
1140
1141# process_generic_property_file() can accept certain overrides in its input.
1142# Each of these must begin AND end with $CMD_DELIM.
1143my $CMD_DELIM = "\a";
1144my $REPLACE_CMD = 'replace'; # Override the Replace
1145my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1146
1147my $NO = 0;
1148my $YES = 1;
1149
1150# Values for the Replace argument to add_range.
1151# $NO # Don't replace; add only the code points not
1152 # already present.
1153my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1154 # the comments at the subroutine definition.
1155my $UNCONDITIONALLY = 2; # Replace without conditions.
1156my $MULTIPLE = 4; # Don't replace, but add a duplicate record if
1157 # already there
56343c78 1158my $CROAK = 5; # Die with an error if is already there
99870f4d
KW
1159
1160# Flags to give property statuses. The phrases are to remind maintainers that
1161# if the flag is changed, the indefinite article referring to it in the
1162# documentation may need to be as well.
1163my $NORMAL = "";
1164my $SUPPRESSED = 'z'; # The character should never actually be seen, since
1165 # it is suppressed
37e2e78e 1166my $PLACEHOLDER = 'P'; # Implies no pod entry generated
99870f4d
KW
1167my $DEPRECATED = 'D';
1168my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1169my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1170my $DISCOURAGED = 'X';
1171my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1172my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1173my $STRICTER = 'T';
1174my $a_bold_stricter = "a 'B<$STRICTER>'";
1175my $A_bold_stricter = "A 'B<$STRICTER>'";
1176my $STABILIZED = 'S';
1177my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1178my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1179my $OBSOLETE = 'O';
1180my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1181my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1182
1183my %status_past_participles = (
1184 $DISCOURAGED => 'discouraged',
1185 $SUPPRESSED => 'should never be generated',
1186 $STABILIZED => 'stabilized',
1187 $OBSOLETE => 'obsolete',
37e2e78e 1188 $DEPRECATED => 'deprecated',
99870f4d
KW
1189);
1190
f5817e0a
KW
1191# The format of the values of the tables:
1192my $EMPTY_FORMAT = "";
99870f4d
KW
1193my $BINARY_FORMAT = 'b';
1194my $DECIMAL_FORMAT = 'd';
1195my $FLOAT_FORMAT = 'f';
1196my $INTEGER_FORMAT = 'i';
1197my $HEX_FORMAT = 'x';
1198my $RATIONAL_FORMAT = 'r';
1199my $STRING_FORMAT = 's';
a14f3cb1 1200my $DECOMP_STRING_FORMAT = 'c';
99870f4d
KW
1201
1202my %map_table_formats = (
1203 $BINARY_FORMAT => 'binary',
1204 $DECIMAL_FORMAT => 'single decimal digit',
1205 $FLOAT_FORMAT => 'floating point number',
1206 $INTEGER_FORMAT => 'integer',
1207 $HEX_FORMAT => 'positive hex whole number; a code point',
1208 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1209 $STRING_FORMAT => 'string',
92f9d56c 1210 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
99870f4d
KW
1211);
1212
1213# Unicode didn't put such derived files in a separate directory at first.
1214my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1215my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1216my $AUXILIARY = 'auxiliary';
1217
1218# Hashes that will eventually go into Heavy.pl for the use of utf8_heavy.pl
1219my %loose_to_file_of; # loosely maps table names to their respective
1220 # files
1221my %stricter_to_file_of; # same; but for stricter mapping.
1222my %nv_floating_to_rational; # maps numeric values floating point numbers to
1223 # their rational equivalent
1224my %loose_property_name_of; # Loosely maps property names to standard form
1225
d867ccfb
KW
1226# Most properties are immune to caseless matching, otherwise you would get
1227# nonsensical results, as properties are a function of a code point, not
1228# everything that is caselessly equivalent to that code point. For example,
1229# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1230# be true because 's' and 'S' are equivalent caselessly. However,
1231# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1232# extend that concept to those very few properties that are like this. Each
1233# such property will match the full range caselessly. They are hard-coded in
1234# the program; it's not worth trying to make it general as it's extremely
1235# unlikely that they will ever change.
1236my %caseless_equivalent_to;
1237
99870f4d
KW
1238# These constants names and values were taken from the Unicode standard,
1239# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1240# syllables. The '_string' versions are so generated tables can retain the
1241# hex format, which is the more familiar value
1242my $SBase_string = "0xAC00";
1243my $SBase = CORE::hex $SBase_string;
1244my $LBase_string = "0x1100";
1245my $LBase = CORE::hex $LBase_string;
1246my $VBase_string = "0x1161";
1247my $VBase = CORE::hex $VBase_string;
1248my $TBase_string = "0x11A7";
1249my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1250my $SCount = 11172;
1251my $LCount = 19;
1252my $VCount = 21;
1253my $TCount = 28;
1254my $NCount = $VCount * $TCount;
1255
1256# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1257# with the above published constants.
1258my %Jamo;
1259my %Jamo_L; # Leading consonants
1260my %Jamo_V; # Vowels
1261my %Jamo_T; # Trailing consonants
1262
37e2e78e 1263my @backslash_X_tests; # List of tests read in for testing \X
99870f4d
KW
1264my @unhandled_properties; # Will contain a list of properties found in
1265 # the input that we didn't process.
f86864ac 1266my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1267 # listed in the pod
1268my @map_properties; # Properties that get map files written
1269my @named_sequences; # NamedSequences.txt contents.
1270my %potential_files; # Generated list of all .txt files in the directory
1271 # structure so we can warn if something is being
1272 # ignored.
1273my @files_actually_output; # List of files we generated.
1274my @more_Names; # Some code point names are compound; this is used
1275 # to store the extra components of them.
1276my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1277 # the minimum before we consider it equivalent to a
1278 # candidate rational
1279my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1280
1281# These store references to certain commonly used property objects
1282my $gc;
1283my $perl;
1284my $block;
3e20195b
KW
1285my $perl_charname;
1286my $print;
99870f4d
KW
1287
1288# Are there conflicting names because of beginning with 'In_', or 'Is_'
1289my $has_In_conflicts = 0;
1290my $has_Is_conflicts = 0;
1291
1292sub internal_file_to_platform ($) {
1293 # Convert our file paths which have '/' separators to those of the
1294 # platform.
1295
1296 my $file = shift;
1297 return undef unless defined $file;
1298
1299 return File::Spec->join(split '/', $file);
d07a55ed 1300}
5beb625e 1301
99870f4d
KW
1302sub file_exists ($) { # platform independent '-e'. This program internally
1303 # uses slash as a path separator.
1304 my $file = shift;
1305 return 0 if ! defined $file;
1306 return -e internal_file_to_platform($file);
1307}
5beb625e 1308
99870f4d 1309sub objaddr($) {
23e33b60
KW
1310 # Returns the address of the blessed input object.
1311 # It doesn't check for blessedness because that would do a string eval
1312 # every call, and the program is structured so that this is never called
1313 # for a non-blessed object.
99870f4d 1314
23e33b60 1315 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1316
1317 # Numifying a ref gives its address.
051df77b 1318 return pack 'J', $_[0];
99870f4d
KW
1319}
1320
558712cf 1321# These are used only if $annotate is true.
c4019d52
KW
1322# The entire range of Unicode characters is examined to populate these
1323# after all the input has been processed. But most can be skipped, as they
1324# have the same descriptive phrases, such as being unassigned
1325my @viacode; # Contains the 1 million character names
1326my @printable; # boolean: And are those characters printable?
1327my @annotate_char_type; # Contains a type of those characters, specifically
1328 # for the purposes of annotation.
1329my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1330 # name for the purposes of annotation. They map to the
c4019d52
KW
1331 # upper edge of the range, so that the end point can
1332 # be immediately found. This is used to skip ahead to
1333 # the end of a range, and avoid processing each
1334 # individual code point in it.
1335my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1336 # characters, but excluding those which are
1337 # also noncharacter code points
1338
1339# The annotation types are an extension of the regular range types, though
1340# some of the latter are folded into one. Make the new types negative to
1341# avoid conflicting with the regular types
1342my $SURROGATE_TYPE = -1;
1343my $UNASSIGNED_TYPE = -2;
1344my $PRIVATE_USE_TYPE = -3;
1345my $NONCHARACTER_TYPE = -4;
1346my $CONTROL_TYPE = -5;
1347my $UNKNOWN_TYPE = -6; # Used only if there is a bug in this program
1348
1349sub populate_char_info ($) {
558712cf 1350 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1351 # input code point's info that are needed for outputting more detailed
1352 # comments. If calling context wants a return, it is the end point of
1353 # any contiguous range of characters that share essentially the same info
1354
1355 my $i = shift;
1356 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1357
1358 $viacode[$i] = $perl_charname->value_of($i) || "";
1359
1360 # A character is generally printable if Unicode says it is,
1361 # but below we make sure that most Unicode general category 'C' types
1362 # aren't.
1363 $printable[$i] = $print->contains($i);
1364
1365 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1366
1367 # Only these two regular types are treated specially for annotations
1368 # purposes
1369 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1370 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1371
1372 # Give a generic name to all code points that don't have a real name.
1373 # We output ranges, if applicable, for these. Also calculate the end
1374 # point of the range.
1375 my $end;
1376 if (! $viacode[$i]) {
1377 if ($gc-> table('Surrogate')->contains($i)) {
1378 $viacode[$i] = 'Surrogate';
1379 $annotate_char_type[$i] = $SURROGATE_TYPE;
1380 $printable[$i] = 0;
1381 $end = $gc->table('Surrogate')->containing_range($i)->end;
1382 }
1383 elsif ($gc-> table('Private_use')->contains($i)) {
1384 $viacode[$i] = 'Private Use';
1385 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1386 $printable[$i] = 0;
1387 $end = $gc->table('Private_Use')->containing_range($i)->end;
1388 }
1389 elsif (Property::property_ref('Noncharacter_Code_Point')-> table('Y')->
1390 contains($i))
1391 {
1392 $viacode[$i] = 'Noncharacter';
1393 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1394 $printable[$i] = 0;
1395 $end = property_ref('Noncharacter_Code_Point')->table('Y')->
1396 containing_range($i)->end;
1397 }
1398 elsif ($gc-> table('Control')->contains($i)) {
1399 $viacode[$i] = 'Control';
1400 $annotate_char_type[$i] = $CONTROL_TYPE;
1401 $printable[$i] = 0;
1402 $end = 0x81 if $i == 0x80; # Hard-code this one known case
1403 }
1404 elsif ($gc-> table('Unassigned')->contains($i)) {
1405 $viacode[$i] = 'Unassigned, block=' . $block-> value_of($i);
1406 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1407 $printable[$i] = 0;
1408
1409 # Because we name the unassigned by the blocks they are in, it
1410 # can't go past the end of that block, and it also can't go past
1411 # the unassigned range it is in. The special table makes sure
1412 # that the non-characters, which are unassigned, are separated
1413 # out.
1414 $end = min($block->containing_range($i)->end,
1415 $unassigned_sans_noncharacters-> containing_range($i)->
1416 end);
13ca76ff
KW
1417 }
1418 else {
1419 Carp::my_carp_bug("Can't figure out how to annotate "
1420 . sprintf("U+%04X", $i)
1421 . ". Proceeding anyway.");
c4019d52
KW
1422 $viacode[$i] = 'UNKNOWN';
1423 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1424 $printable[$i] = 0;
1425 }
1426 }
1427
1428 # Here, has a name, but if it's one in which the code point number is
1429 # appended to the name, do that.
1430 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1431 $viacode[$i] .= sprintf("-%04X", $i);
1432 $end = $perl_charname->containing_range($i)->end;
1433 }
1434
1435 # And here, has a name, but if it's a hangul syllable one, replace it with
1436 # the correct name from the Unicode algorithm
1437 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1438 use integer;
1439 my $SIndex = $i - $SBase;
1440 my $L = $LBase + $SIndex / $NCount;
1441 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1442 my $T = $TBase + $SIndex % $TCount;
1443 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1444 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1445 $end = $perl_charname->containing_range($i)->end;
1446 }
1447
1448 return if ! defined wantarray;
1449 return $i if ! defined $end; # If not a range, return the input
1450
1451 # Save this whole range so can find the end point quickly
1452 $annotate_ranges->add_map($i, $end, $end);
1453
1454 return $end;
1455}
1456
23e33b60
KW
1457# Commented code below should work on Perl 5.8.
1458## This 'require' doesn't necessarily work in miniperl, and even if it does,
1459## the native perl version of it (which is what would operate under miniperl)
1460## is extremely slow, as it does a string eval every call.
1461#my $has_fast_scalar_util = $\18 !~ /miniperl/
1462# && defined eval "require Scalar::Util";
1463#
1464#sub objaddr($) {
1465# # Returns the address of the blessed input object. Uses the XS version if
1466# # available. It doesn't check for blessedness because that would do a
1467# # string eval every call, and the program is structured so that this is
1468# # never called for a non-blessed object.
1469#
1470# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1471#
1472# # Check at least that is a ref.
1473# my $pkg = ref($_[0]) or return undef;
1474#
1475# # Change to a fake package to defeat any overloaded stringify
1476# bless $_[0], 'main::Fake';
1477#
1478# # Numifying a ref gives its address.
051df77b 1479# my $addr = pack 'J', $_[0];
23e33b60
KW
1480#
1481# # Return to original class
1482# bless $_[0], $pkg;
1483# return $addr;
1484#}
1485
99870f4d
KW
1486sub max ($$) {
1487 my $a = shift;
1488 my $b = shift;
1489 return $a if $a >= $b;
1490 return $b;
1491}
1492
1493sub min ($$) {
1494 my $a = shift;
1495 my $b = shift;
1496 return $a if $a <= $b;
1497 return $b;
1498}
1499
1500sub clarify_number ($) {
1501 # This returns the input number with underscores inserted every 3 digits
1502 # in large (5 digits or more) numbers. Input must be entirely digits, not
1503 # checked.
1504
1505 my $number = shift;
1506 my $pos = length($number) - 3;
1507 return $number if $pos <= 1;
1508 while ($pos > 0) {
1509 substr($number, $pos, 0) = '_';
1510 $pos -= 3;
5beb625e 1511 }
99870f4d 1512 return $number;
99598c8c
JH
1513}
1514
12ac2576 1515
99870f4d 1516package Carp;
7ebf06b3 1517
99870f4d
KW
1518# These routines give a uniform treatment of messages in this program. They
1519# are placed in the Carp package to cause the stack trace to not include them,
1520# although an alternative would be to use another package and set @CARP_NOT
1521# for it.
12ac2576 1522
99870f4d 1523our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1524
99f78760
KW
1525# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1526# and overload trying to load Scalar:Util under miniperl. See
1527# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1528undef $overload::VERSION;
1529
99870f4d
KW
1530sub my_carp {
1531 my $message = shift || "";
1532 my $nofold = shift || 0;
7ebf06b3 1533
99870f4d
KW
1534 if ($message) {
1535 $message = main::join_lines($message);
1536 $message =~ s/^$0: *//; # Remove initial program name
1537 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1538 $message = "\n$0: $message;";
12ac2576 1539
99870f4d
KW
1540 # Fold the message with program name, semi-colon end punctuation
1541 # (which looks good with the message that carp appends to it), and a
1542 # hanging indent for continuation lines.
1543 $message = main::simple_fold($message, "", 4) unless $nofold;
1544 $message =~ s/\n$//; # Remove the trailing nl so what carp
1545 # appends is to the same line
1546 }
12ac2576 1547
99870f4d 1548 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1549
99870f4d
KW
1550 carp $message;
1551 return;
1552}
7ebf06b3 1553
99870f4d
KW
1554sub my_carp_bug {
1555 # This is called when it is clear that the problem is caused by a bug in
1556 # this program.
7ebf06b3 1557
99870f4d
KW
1558 my $message = shift;
1559 $message =~ s/^$0: *//;
1560 $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");
1561 carp $message;
1562 return;
1563}
7ebf06b3 1564
99870f4d
KW
1565sub carp_too_few_args {
1566 if (@_ != 2) {
1567 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1568 return;
12ac2576 1569 }
7ebf06b3 1570
99870f4d
KW
1571 my $args_ref = shift;
1572 my $count = shift;
7ebf06b3 1573
99870f4d
KW
1574 my_carp_bug("Need at least $count arguments to "
1575 . (caller 1)[3]
1576 . ". Instead got: '"
1577 . join ', ', @$args_ref
1578 . "'. No action taken.");
1579 return;
12ac2576
JP
1580}
1581
99870f4d
KW
1582sub carp_extra_args {
1583 my $args_ref = shift;
1584 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1585
99870f4d
KW
1586 unless (ref $args_ref) {
1587 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1588 return;
1589 }
1590 my ($package, $file, $line) = caller;
1591 my $subroutine = (caller 1)[3];
cf25bb62 1592
99870f4d
KW
1593 my $list;
1594 if (ref $args_ref eq 'HASH') {
1595 foreach my $key (keys %$args_ref) {
1596 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1597 }
99870f4d 1598 $list = join ', ', each %{$args_ref};
cf25bb62 1599 }
99870f4d
KW
1600 elsif (ref $args_ref eq 'ARRAY') {
1601 foreach my $arg (@$args_ref) {
1602 $arg = $UNDEF unless defined $arg;
1603 }
1604 $list = join ', ', @$args_ref;
1605 }
1606 else {
1607 my_carp_bug("Can't cope with ref "
1608 . ref($args_ref)
1609 . " . argument to 'carp_extra_args'. Not checking arguments.");
1610 return;
1611 }
1612
1613 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1614 return;
d73e5302
JH
1615}
1616
99870f4d
KW
1617package main;
1618
1619{ # Closure
1620
1621 # This program uses the inside-out method for objects, as recommended in
1622 # "Perl Best Practices". This closure aids in generating those. There
1623 # are two routines. setup_package() is called once per package to set
1624 # things up, and then set_access() is called for each hash representing a
1625 # field in the object. These routines arrange for the object to be
1626 # properly destroyed when no longer used, and for standard accessor
1627 # functions to be generated. If you need more complex accessors, just
1628 # write your own and leave those accesses out of the call to set_access().
1629 # More details below.
1630
1631 my %constructor_fields; # fields that are to be used in constructors; see
1632 # below
1633
1634 # The values of this hash will be the package names as keys to other
1635 # hashes containing the name of each field in the package as keys, and
1636 # references to their respective hashes as values.
1637 my %package_fields;
1638
1639 sub setup_package {
1640 # Sets up the package, creating standard DESTROY and dump methods
1641 # (unless already defined). The dump method is used in debugging by
1642 # simple_dumper().
1643 # The optional parameters are:
1644 # a) a reference to a hash, that gets populated by later
1645 # set_access() calls with one of the accesses being
1646 # 'constructor'. The caller can then refer to this, but it is
1647 # not otherwise used by these two routines.
1648 # b) a reference to a callback routine to call during destruction
1649 # of the object, before any fields are actually destroyed
1650
1651 my %args = @_;
1652 my $constructor_ref = delete $args{'Constructor_Fields'};
1653 my $destroy_callback = delete $args{'Destroy_Callback'};
1654 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1655
1656 my %fields;
1657 my $package = (caller)[0];
1658
1659 $package_fields{$package} = \%fields;
1660 $constructor_fields{$package} = $constructor_ref;
1661
1662 unless ($package->can('DESTROY')) {
1663 my $destroy_name = "${package}::DESTROY";
1664 no strict "refs";
1665
1666 # Use typeglob to give the anonymous subroutine the name we want
1667 *$destroy_name = sub {
1668 my $self = shift;
ffe43484 1669 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1670
1671 $self->$destroy_callback if $destroy_callback;
1672 foreach my $field (keys %{$package_fields{$package}}) {
1673 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1674 delete $package_fields{$package}{$field}{$addr};
1675 }
1676 return;
1677 }
1678 }
1679
1680 unless ($package->can('dump')) {
1681 my $dump_name = "${package}::dump";
1682 no strict "refs";
1683 *$dump_name = sub {
1684 my $self = shift;
1685 return dump_inside_out($self, $package_fields{$package}, @_);
1686 }
1687 }
1688 return;
1689 }
1690
1691 sub set_access {
1692 # Arrange for the input field to be garbage collected when no longer
1693 # needed. Also, creates standard accessor functions for the field
1694 # based on the optional parameters-- none if none of these parameters:
1695 # 'addable' creates an 'add_NAME()' accessor function.
1696 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1697 # function.
1698 # 'settable' creates a 'set_NAME()' accessor function.
1699 # 'constructor' doesn't create an accessor function, but adds the
1700 # field to the hash that was previously passed to
1701 # setup_package();
1702 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1703 # 'add' etc. all mean 'addable'.
1704 # The read accessor function will work on both array and scalar
1705 # values. If another accessor in the parameter list is 'a', the read
1706 # access assumes an array. You can also force it to be array access
1707 # by specifying 'readable_array' instead of 'readable'
1708 #
1709 # A sort-of 'protected' access can be set-up by preceding the addable,
1710 # readable or settable with some initial portion of 'protected_' (but,
1711 # the underscore is required), like 'p_a', 'pro_set', etc. The
1712 # "protection" is only by convention. All that happens is that the
1713 # accessor functions' names begin with an underscore. So instead of
1714 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1715 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1716 # end of each package, and then storing the __LINE__ ranges and
1717 # checking them on every accessor. But that is way overkill.)
1718
1719 # We create anonymous subroutines as the accessors and then use
1720 # typeglobs to assign them to the proper package and name
1721
1722 my $name = shift; # Name of the field
1723 my $field = shift; # Reference to the inside-out hash containing the
1724 # field
1725
1726 my $package = (caller)[0];
1727
1728 if (! exists $package_fields{$package}) {
1729 croak "$0: Must call 'setup_package' before 'set_access'";
1730 }
d73e5302 1731
99870f4d
KW
1732 # Stash the field so DESTROY can get it.
1733 $package_fields{$package}{$name} = $field;
cf25bb62 1734
99870f4d
KW
1735 # Remaining arguments are the accessors. For each...
1736 foreach my $access (@_) {
1737 my $access = lc $access;
cf25bb62 1738
99870f4d 1739 my $protected = "";
cf25bb62 1740
99870f4d
KW
1741 # Match the input as far as it goes.
1742 if ($access =~ /^(p[^_]*)_/) {
1743 $protected = $1;
1744 if (substr('protected_', 0, length $protected)
1745 eq $protected)
1746 {
1747
1748 # Add 1 for the underscore not included in $protected
1749 $access = substr($access, length($protected) + 1);
1750 $protected = '_';
1751 }
1752 else {
1753 $protected = "";
1754 }
1755 }
1756
1757 if (substr('addable', 0, length $access) eq $access) {
1758 my $subname = "${package}::${protected}add_$name";
1759 no strict "refs";
1760
1761 # add_ accessor. Don't add if already there, which we
1762 # determine using 'eq' for scalars and '==' otherwise.
1763 *$subname = sub {
1764 use strict "refs";
1765 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1766 my $self = shift;
1767 my $value = shift;
ffe43484 1768 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1769 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1770 if (ref $value) {
f998e60c 1771 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1772 }
1773 else {
f998e60c 1774 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1775 }
f998e60c 1776 push @{$field->{$addr}}, $value;
99870f4d
KW
1777 return;
1778 }
1779 }
1780 elsif (substr('constructor', 0, length $access) eq $access) {
1781 if ($protected) {
1782 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1783 }
1784 else {
1785 $constructor_fields{$package}{$name} = $field;
1786 }
1787 }
1788 elsif (substr('readable_array', 0, length $access) eq $access) {
1789
1790 # Here has read access. If one of the other parameters for
1791 # access is array, or this one specifies array (by being more
1792 # than just 'readable_'), then create a subroutine that
1793 # assumes the data is an array. Otherwise just a scalar
1794 my $subname = "${package}::${protected}$name";
1795 if (grep { /^a/i } @_
1796 or length($access) > length('readable_'))
1797 {
1798 no strict "refs";
1799 *$subname = sub {
1800 use strict "refs";
23e33b60 1801 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 1802 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
1803 if (ref $field->{$addr} ne 'ARRAY') {
1804 my $type = ref $field->{$addr};
1805 $type = 'scalar' unless $type;
1806 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
1807 return;
1808 }
1809 return scalar @{$field->{$addr}} unless wantarray;
1810
1811 # Make a copy; had problems with caller modifying the
1812 # original otherwise
1813 my @return = @{$field->{$addr}};
1814 return @return;
1815 }
1816 }
1817 else {
1818
1819 # Here not an array value, a simpler function.
1820 no strict "refs";
1821 *$subname = sub {
1822 use strict "refs";
23e33b60 1823 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 1824 no overloading;
051df77b 1825 return $field->{pack 'J', $_[0]};
99870f4d
KW
1826 }
1827 }
1828 }
1829 elsif (substr('settable', 0, length $access) eq $access) {
1830 my $subname = "${package}::${protected}set_$name";
1831 no strict "refs";
1832 *$subname = sub {
1833 use strict "refs";
23e33b60
KW
1834 if (main::DEBUG) {
1835 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
1836 Carp::carp_extra_args(\@_) if @_ > 2;
1837 }
1838 # $self is $_[0]; $value is $_[1]
f998e60c 1839 no overloading;
051df77b 1840 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
1841 return;
1842 }
1843 }
1844 else {
1845 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
1846 }
cf25bb62 1847 }
99870f4d 1848 return;
cf25bb62 1849 }
99870f4d
KW
1850}
1851
1852package Input_file;
1853
1854# All input files use this object, which stores various attributes about them,
1855# and provides for convenient, uniform handling. The run method wraps the
1856# processing. It handles all the bookkeeping of opening, reading, and closing
1857# the file, returning only significant input lines.
1858#
1859# Each object gets a handler which processes the body of the file, and is
1860# called by run(). Most should use the generic, default handler, which has
1861# code scrubbed to handle things you might not expect. A handler should
1862# basically be a while(next_line()) {...} loop.
1863#
1864# You can also set up handlers to
1865# 1) call before the first line is read for pre processing
1866# 2) call to adjust each line of the input before the main handler gets them
1867# 3) call upon EOF before the main handler exits its loop
1868# 4) call at the end for post processing
1869#
1870# $_ is used to store the input line, and is to be filtered by the
1871# each_line_handler()s. So, if the format of the line is not in the desired
1872# format for the main handler, these are used to do that adjusting. They can
1873# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
1874# so the $_ output of one is used as the input to the next. None of the other
1875# handlers are stackable, but could easily be changed to be so.
1876#
1877# Most of the handlers can call insert_lines() or insert_adjusted_lines()
1878# which insert the parameters as lines to be processed before the next input
1879# file line is read. This allows the EOF handler to flush buffers, for
1880# example. The difference between the two routines is that the lines inserted
1881# by insert_lines() are subjected to the each_line_handler()s. (So if you
1882# called it from such a handler, you would get infinite recursion.) Lines
1883# inserted by insert_adjusted_lines() go directly to the main handler without
1884# any adjustments. If the post-processing handler calls any of these, there
1885# will be no effect. Some error checking for these conditions could be added,
1886# but it hasn't been done.
1887#
1888# carp_bad_line() should be called to warn of bad input lines, which clears $_
1889# to prevent further processing of the line. This routine will output the
1890# message as a warning once, and then keep a count of the lines that have the
1891# same message, and output that count at the end of the file's processing.
1892# This keeps the number of messages down to a manageable amount.
1893#
1894# get_missings() should be called to retrieve any @missing input lines.
1895# Messages will be raised if this isn't done if the options aren't to ignore
1896# missings.
1897
1898sub trace { return main::trace(@_); }
1899
99870f4d
KW
1900{ # Closure
1901 # Keep track of fields that are to be put into the constructor.
1902 my %constructor_fields;
1903
1904 main::setup_package(Constructor_Fields => \%constructor_fields);
1905
1906 my %file; # Input file name, required
1907 main::set_access('file', \%file, qw{ c r });
1908
1909 my %first_released; # Unicode version file was first released in, required
1910 main::set_access('first_released', \%first_released, qw{ c r });
1911
1912 my %handler; # Subroutine to process the input file, defaults to
1913 # 'process_generic_property_file'
1914 main::set_access('handler', \%handler, qw{ c });
1915
1916 my %property;
1917 # name of property this file is for. defaults to none, meaning not
1918 # applicable, or is otherwise determinable, for example, from each line.
1919 main::set_access('property', \%property, qw{ c });
1920
1921 my %optional;
1922 # If this is true, the file is optional. If not present, no warning is
1923 # output. If it is present, the string given by this parameter is
1924 # evaluated, and if false the file is not processed.
1925 main::set_access('optional', \%optional, 'c', 'r');
1926
1927 my %non_skip;
1928 # This is used for debugging, to skip processing of all but a few input
1929 # files. Add 'non_skip => 1' to the constructor for those files you want
1930 # processed when you set the $debug_skip global.
1931 main::set_access('non_skip', \%non_skip, 'c');
1932
37e2e78e
KW
1933 my %skip;
1934 # This is used to skip processing of this input file semi-permanently.
1935 # It is used for files that we aren't planning to process anytime soon,
1936 # but want to allow to be in the directory and not raise a message that we
1937 # are not handling. Mostly for test files. This is in contrast to the
1938 # non_skip element, which is supposed to be used very temporarily for
1939 # debugging. Sets 'optional' to 1
1940 main::set_access('skip', \%skip, 'c');
1941
99870f4d
KW
1942 my %each_line_handler;
1943 # list of subroutines to look at and filter each non-comment line in the
1944 # file. defaults to none. The subroutines are called in order, each is
1945 # to adjust $_ for the next one, and the final one adjusts it for
1946 # 'handler'
1947 main::set_access('each_line_handler', \%each_line_handler, 'c');
1948
1949 my %has_missings_defaults;
1950 # ? Are there lines in the file giving default values for code points
1951 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
1952 # the norm, but IGNORED means it has such lines, but the handler doesn't
1953 # use them. Having these three states allows us to catch changes to the
1954 # UCD that this program should track
1955 main::set_access('has_missings_defaults',
1956 \%has_missings_defaults, qw{ c r });
1957
1958 my %pre_handler;
1959 # Subroutine to call before doing anything else in the file. If undef, no
1960 # such handler is called.
1961 main::set_access('pre_handler', \%pre_handler, qw{ c });
1962
1963 my %eof_handler;
1964 # Subroutine to call upon getting an EOF on the input file, but before
1965 # that is returned to the main handler. This is to allow buffers to be
1966 # flushed. The handler is expected to call insert_lines() or
1967 # insert_adjusted() with the buffered material
1968 main::set_access('eof_handler', \%eof_handler, qw{ c r });
1969
1970 my %post_handler;
1971 # Subroutine to call after all the lines of the file are read in and
1972 # processed. If undef, no such handler is called.
1973 main::set_access('post_handler', \%post_handler, qw{ c });
1974
1975 my %progress_message;
1976 # Message to print to display progress in lieu of the standard one
1977 main::set_access('progress_message', \%progress_message, qw{ c });
1978
1979 my %handle;
1980 # cache open file handle, internal. Is undef if file hasn't been
1981 # processed at all, empty if has;
1982 main::set_access('handle', \%handle);
1983
1984 my %added_lines;
1985 # cache of lines added virtually to the file, internal
1986 main::set_access('added_lines', \%added_lines);
1987
1988 my %errors;
1989 # cache of errors found, internal
1990 main::set_access('errors', \%errors);
1991
1992 my %missings;
1993 # storage of '@missing' defaults lines
1994 main::set_access('missings', \%missings);
1995
1996 sub new {
1997 my $class = shift;
1998
1999 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2000 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2001
2002 # Set defaults
2003 $handler{$addr} = \&main::process_generic_property_file;
2004 $non_skip{$addr} = 0;
37e2e78e 2005 $skip{$addr} = 0;
99870f4d
KW
2006 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2007 $handle{$addr} = undef;
2008 $added_lines{$addr} = [ ];
2009 $each_line_handler{$addr} = [ ];
2010 $errors{$addr} = { };
2011 $missings{$addr} = [ ];
2012
2013 # Two positional parameters.
99f78760 2014 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2015 $file{$addr} = main::internal_file_to_platform(shift);
2016 $first_released{$addr} = shift;
2017
2018 # The rest of the arguments are key => value pairs
2019 # %constructor_fields has been set up earlier to list all possible
2020 # ones. Either set or push, depending on how the default has been set
2021 # up just above.
2022 my %args = @_;
2023 foreach my $key (keys %args) {
2024 my $argument = $args{$key};
2025
2026 # Note that the fields are the lower case of the constructor keys
2027 my $hash = $constructor_fields{lc $key};
2028 if (! defined $hash) {
2029 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2030 next;
2031 }
2032 if (ref $hash->{$addr} eq 'ARRAY') {
2033 if (ref $argument eq 'ARRAY') {
2034 foreach my $argument (@{$argument}) {
2035 next if ! defined $argument;
2036 push @{$hash->{$addr}}, $argument;
2037 }
2038 }
2039 else {
2040 push @{$hash->{$addr}}, $argument if defined $argument;
2041 }
2042 }
2043 else {
2044 $hash->{$addr} = $argument;
2045 }
2046 delete $args{$key};
2047 };
2048
2049 # If the file has a property for it, it means that the property is not
2050 # listed in the file's entries. So add a handler to the list of line
2051 # handlers to insert the property name into the lines, to provide a
2052 # uniform interface to the final processing subroutine.
2053 # the final code doesn't have to worry about that.
2054 if ($property{$addr}) {
2055 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
2056 }
2057
2058 if ($non_skip{$addr} && ! $debug_skip && $verbosity) {
2059 print "Warning: " . __PACKAGE__ . " constructor for $file{$addr} has useless 'non_skip' in it\n";
a3a8c5f0 2060 }
99870f4d 2061
37e2e78e
KW
2062 $optional{$addr} = 1 if $skip{$addr};
2063
99870f4d 2064 return $self;
d73e5302
JH
2065 }
2066
cf25bb62 2067
99870f4d
KW
2068 use overload
2069 fallback => 0,
2070 qw("") => "_operator_stringify",
2071 "." => \&main::_operator_dot,
2072 ;
cf25bb62 2073
99870f4d
KW
2074 sub _operator_stringify {
2075 my $self = shift;
cf25bb62 2076
99870f4d 2077 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2078 }
d73e5302 2079
99870f4d
KW
2080 # flag to make sure extracted files are processed early
2081 my $seen_non_extracted_non_age = 0;
d73e5302 2082
99870f4d
KW
2083 sub run {
2084 # Process the input object $self. This opens and closes the file and
2085 # calls all the handlers for it. Currently, this can only be called
2086 # once per file, as it destroy's the EOF handler
d73e5302 2087
99870f4d
KW
2088 my $self = shift;
2089 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2090
ffe43484 2091 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2092
99870f4d 2093 my $file = $file{$addr};
d73e5302 2094
99870f4d
KW
2095 # Don't process if not expecting this file (because released later
2096 # than this Unicode version), and isn't there. This means if someone
2097 # copies it into an earlier version's directory, we will go ahead and
2098 # process it.
2099 return if $first_released{$addr} gt $v_version && ! -e $file;
2100
2101 # If in debugging mode and this file doesn't have the non-skip
2102 # flag set, and isn't one of the critical files, skip it.
2103 if ($debug_skip
2104 && $first_released{$addr} ne v0
2105 && ! $non_skip{$addr})
2106 {
2107 print "Skipping $file in debugging\n" if $verbosity;
2108 return;
2109 }
2110
2111 # File could be optional
37e2e78e 2112 if ($optional{$addr}) {
99870f4d
KW
2113 return unless -e $file;
2114 my $result = eval $optional{$addr};
2115 if (! defined $result) {
2116 Carp::my_carp_bug("Got '$@' when tried to eval $optional{$addr}. $file Skipped.");
2117 return;
2118 }
2119 if (! $result) {
2120 if ($verbosity) {
2121 print STDERR "Skipping processing input file '$file' because '$optional{$addr}' is not true\n";
2122 }
2123 return;
2124 }
2125 }
2126
2127 if (! defined $file || ! -e $file) {
2128
2129 # If the file doesn't exist, see if have internal data for it
2130 # (based on first_released being 0).
2131 if ($first_released{$addr} eq v0) {
2132 $handle{$addr} = 'pretend_is_open';
2133 }
2134 else {
2135 if (! $optional{$addr} # File could be optional
2136 && $v_version ge $first_released{$addr})
2137 {
2138 print STDERR "Skipping processing input file '$file' because not found\n" if $v_version ge $first_released{$addr};
2139 }
2140 return;
2141 }
2142 }
2143 else {
2144
37e2e78e
KW
2145 # Here, the file exists. Some platforms may change the case of
2146 # its name
99870f4d 2147 if ($seen_non_extracted_non_age) {
517956bf 2148 if ($file =~ /$EXTRACTED/i) {
99870f4d 2149 Carp::my_carp_bug(join_lines(<<END
99f78760 2150$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2151anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2152have subtle problems
2153END
2154 ));
2155 }
2156 }
2157 elsif ($EXTRACTED_DIR
2158 && $first_released{$addr} ne v0
517956bf
CB
2159 && $file !~ /$EXTRACTED/i
2160 && lc($file) ne 'dage.txt')
99870f4d
KW
2161 {
2162 # We don't set this (by the 'if' above) if we have no
2163 # extracted directory, so if running on an early version,
2164 # this test won't work. Not worth worrying about.
2165 $seen_non_extracted_non_age = 1;
2166 }
2167
2168 # And mark the file as having being processed, and warn if it
2169 # isn't a file we are expecting. As we process the files,
2170 # they are deleted from the hash, so any that remain at the
2171 # end of the program are files that we didn't process.
517956bf
CB
2172 my $fkey = File::Spec->rel2abs($file);
2173 my $expecting = delete $potential_files{$fkey};
2174 $expecting = delete $potential_files{lc($fkey)} unless defined $expecting;
678f13d5
KW
2175 Carp::my_carp("Was not expecting '$file'.") if
2176 ! $expecting
99870f4d
KW
2177 && ! defined $handle{$addr};
2178
37e2e78e
KW
2179 # Having deleted from expected files, we can quit if not to do
2180 # anything. Don't print progress unless really want verbosity
2181 if ($skip{$addr}) {
2182 print "Skipping $file.\n" if $verbosity >= $VERBOSE;
2183 return;
2184 }
2185
99870f4d
KW
2186 # Open the file, converting the slashes used in this program
2187 # into the proper form for the OS
2188 my $file_handle;
2189 if (not open $file_handle, "<", $file) {
2190 Carp::my_carp("Can't open $file. Skipping: $!");
2191 return 0;
2192 }
2193 $handle{$addr} = $file_handle; # Cache the open file handle
2194 }
2195
2196 if ($verbosity >= $PROGRESS) {
2197 if ($progress_message{$addr}) {
2198 print "$progress_message{$addr}\n";
2199 }
2200 else {
2201 # If using a virtual file, say so.
2202 print "Processing ", (-e $file)
2203 ? $file
2204 : "substitute $file",
2205 "\n";
2206 }
2207 }
2208
2209
2210 # Call any special handler for before the file.
2211 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2212
2213 # Then the main handler
2214 &{$handler{$addr}}($self);
2215
2216 # Then any special post-file handler.
2217 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2218
2219 # If any errors have been accumulated, output the counts (as the first
2220 # error message in each class was output when it was encountered).
2221 if ($errors{$addr}) {
2222 my $total = 0;
2223 my $types = 0;
2224 foreach my $error (keys %{$errors{$addr}}) {
2225 $total += $errors{$addr}->{$error};
2226 delete $errors{$addr}->{$error};
2227 $types++;
2228 }
2229 if ($total > 1) {
2230 my $message
2231 = "A total of $total lines had errors in $file. ";
2232
2233 $message .= ($types == 1)
2234 ? '(Only the first one was displayed.)'
2235 : '(Only the first of each type was displayed.)';
2236 Carp::my_carp($message);
2237 }
2238 }
2239
2240 if (@{$missings{$addr}}) {
2241 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2242 }
2243
2244 # If a real file handle, close it.
2245 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2246 ref $handle{$addr};
2247 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2248 # the file, as opposed to undef
2249 return;
2250 }
2251
2252 sub next_line {
2253 # Sets $_ to be the next logical input line, if any. Returns non-zero
2254 # if such a line exists. 'logical' means that any lines that have
2255 # been added via insert_lines() will be returned in $_ before the file
2256 # is read again.
2257
2258 my $self = shift;
2259 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2260
ffe43484 2261 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2262
2263 # Here the file is open (or if the handle is not a ref, is an open
2264 # 'virtual' file). Get the next line; any inserted lines get priority
2265 # over the file itself.
2266 my $adjusted;
2267
2268 LINE:
2269 while (1) { # Loop until find non-comment, non-empty line
2270 #local $to_trace = 1 if main::DEBUG;
2271 my $inserted_ref = shift @{$added_lines{$addr}};
2272 if (defined $inserted_ref) {
2273 ($adjusted, $_) = @{$inserted_ref};
2274 trace $adjusted, $_ if main::DEBUG && $to_trace;
2275 return 1 if $adjusted;
2276 }
2277 else {
2278 last if ! ref $handle{$addr}; # Don't read unless is real file
2279 last if ! defined ($_ = readline $handle{$addr});
2280 }
2281 chomp;
2282 trace $_ if main::DEBUG && $to_trace;
2283
2284 # See if this line is the comment line that defines what property
2285 # value that code points that are not listed in the file should
2286 # have. The format or existence of these lines is not guaranteed
2287 # by Unicode since they are comments, but the documentation says
2288 # that this was added for machine-readability, so probably won't
2289 # change. This works starting in Unicode Version 5.0. They look
2290 # like:
2291 #
2292 # @missing: 0000..10FFFF; Not_Reordered
2293 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2294 # @missing: 0000..10FFFF; ; NaN
2295 #
2296 # Save the line for a later get_missings() call.
2297 if (/$missing_defaults_prefix/) {
2298 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2299 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2300 }
2301 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2302 my @defaults = split /\s* ; \s*/x, $_;
2303
2304 # The first field is the @missing, which ends in a
2305 # semi-colon, so can safely shift.
2306 shift @defaults;
2307
2308 # Some of these lines may have empty field placeholders
2309 # which get in the way. An example is:
2310 # @missing: 0000..10FFFF; ; NaN
2311 # Remove them. Process starting from the top so the
2312 # splice doesn't affect things still to be looked at.
2313 for (my $i = @defaults - 1; $i >= 0; $i--) {
2314 next if $defaults[$i] ne "";
2315 splice @defaults, $i, 1;
2316 }
2317
2318 # What's left should be just the property (maybe) and the
2319 # default. Having only one element means it doesn't have
2320 # the property.
2321 my $default;
2322 my $property;
2323 if (@defaults >= 1) {
2324 if (@defaults == 1) {
2325 $default = $defaults[0];
2326 }
2327 else {
2328 $property = $defaults[0];
2329 $default = $defaults[1];
2330 }
2331 }
2332
2333 if (@defaults < 1
2334 || @defaults > 2
2335 || ($default =~ /^</
2336 && $default !~ /^<code *point>$/i
2337 && $default !~ /^<none>$/i))
2338 {
2339 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2340 }
2341 else {
2342
2343 # If the property is missing from the line, it should
2344 # be the one for the whole file
2345 $property = $property{$addr} if ! defined $property;
2346
2347 # Change <none> to the null string, which is what it
2348 # really means. If the default is the code point
2349 # itself, set it to <code point>, which is what
2350 # Unicode uses (but sometimes they've forgotten the
2351 # space)
2352 if ($default =~ /^<none>$/i) {
2353 $default = "";
2354 }
2355 elsif ($default =~ /^<code *point>$/i) {
2356 $default = $CODE_POINT;
2357 }
2358
2359 # Store them as a sub-arrays with both components.
2360 push @{$missings{$addr}}, [ $default, $property ];
2361 }
2362 }
2363
2364 # There is nothing for the caller to process on this comment
2365 # line.
2366 next;
2367 }
2368
2369 # Remove comments and trailing space, and skip this line if the
2370 # result is empty
2371 s/#.*//;
2372 s/\s+$//;
2373 next if /^$/;
2374
2375 # Call any handlers for this line, and skip further processing of
2376 # the line if the handler sets the line to null.
2377 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2378 &{$sub_ref}($self);
2379 next LINE if /^$/;
2380 }
2381
2382 # Here the line is ok. return success.
2383 return 1;
2384 } # End of looping through lines.
2385
2386 # If there is an EOF handler, call it (only once) and if it generates
2387 # more lines to process go back in the loop to handle them.
2388 if ($eof_handler{$addr}) {
2389 &{$eof_handler{$addr}}($self);
2390 $eof_handler{$addr} = ""; # Currently only get one shot at it.
2391 goto LINE if $added_lines{$addr};
2392 }
2393
2394 # Return failure -- no more lines.
2395 return 0;
2396
2397 }
2398
2399# Not currently used, not fully tested.
2400# sub peek {
2401# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
2402# # record. Not callable from an each_line_handler(), nor does it call
2403# # an each_line_handler() on the line.
2404#
2405# my $self = shift;
ffe43484 2406# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2407#
2408# foreach my $inserted_ref (@{$added_lines{$addr}}) {
2409# my ($adjusted, $line) = @{$inserted_ref};
2410# next if $adjusted;
2411#
2412# # Remove comments and trailing space, and return a non-empty
2413# # resulting line
2414# $line =~ s/#.*//;
2415# $line =~ s/\s+$//;
2416# return $line if $line ne "";
2417# }
2418#
2419# return if ! ref $handle{$addr}; # Don't read unless is real file
2420# while (1) { # Loop until find non-comment, non-empty line
2421# local $to_trace = 1 if main::DEBUG;
2422# trace $_ if main::DEBUG && $to_trace;
2423# return if ! defined (my $line = readline $handle{$addr});
2424# chomp $line;
2425# push @{$added_lines{$addr}}, [ 0, $line ];
2426#
2427# $line =~ s/#.*//;
2428# $line =~ s/\s+$//;
2429# return $line if $line ne "";
2430# }
2431#
2432# return;
2433# }
2434
2435
2436 sub insert_lines {
2437 # Lines can be inserted so that it looks like they were in the input
2438 # file at the place it was when this routine is called. See also
2439 # insert_adjusted_lines(). Lines inserted via this routine go through
2440 # any each_line_handler()
2441
2442 my $self = shift;
2443
2444 # Each inserted line is an array, with the first element being 0 to
2445 # indicate that this line hasn't been adjusted, and needs to be
2446 # processed.
f998e60c 2447 no overloading;
051df77b 2448 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
2449 return;
2450 }
2451
2452 sub insert_adjusted_lines {
2453 # Lines can be inserted so that it looks like they were in the input
2454 # file at the place it was when this routine is called. See also
2455 # insert_lines(). Lines inserted via this routine are already fully
2456 # adjusted, ready to be processed; each_line_handler()s handlers will
2457 # not be called. This means this is not a completely general
2458 # facility, as only the last each_line_handler on the stack should
2459 # call this. It could be made more general, by passing to each of the
2460 # line_handlers their position on the stack, which they would pass on
2461 # to this routine, and that would replace the boolean first element in
2462 # the anonymous array pushed here, so that the next_line routine could
2463 # use that to call only those handlers whose index is after it on the
2464 # stack. But this is overkill for what is needed now.
2465
2466 my $self = shift;
2467 trace $_[0] if main::DEBUG && $to_trace;
2468
2469 # Each inserted line is an array, with the first element being 1 to
2470 # indicate that this line has been adjusted
f998e60c 2471 no overloading;
051df77b 2472 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
2473 return;
2474 }
2475
2476 sub get_missings {
2477 # Returns the stored up @missings lines' values, and clears the list.
2478 # The values are in an array, consisting of the default in the first
2479 # element, and the property in the 2nd. However, since these lines
2480 # can be stacked up, the return is an array of all these arrays.
2481
2482 my $self = shift;
2483 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2484
ffe43484 2485 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2486
2487 # If not accepting a list return, just return the first one.
2488 return shift @{$missings{$addr}} unless wantarray;
2489
2490 my @return = @{$missings{$addr}};
2491 undef @{$missings{$addr}};
2492 return @return;
2493 }
2494
2495 sub _insert_property_into_line {
2496 # Add a property field to $_, if this file requires it.
2497
f998e60c 2498 my $self = shift;
ffe43484 2499 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 2500 my $property = $property{$addr};
99870f4d
KW
2501 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2502
2503 $_ =~ s/(;|$)/; $property$1/;
2504 return;
2505 }
2506
2507 sub carp_bad_line {
2508 # Output consistent error messages, using either a generic one, or the
2509 # one given by the optional parameter. To avoid gazillions of the
2510 # same message in case the syntax of a file is way off, this routine
2511 # only outputs the first instance of each message, incrementing a
2512 # count so the totals can be output at the end of the file.
2513
2514 my $self = shift;
2515 my $message = shift;
2516 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2517
ffe43484 2518 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2519
2520 $message = 'Unexpected line' unless $message;
2521
2522 # No trailing punctuation so as to fit with our addenda.
2523 $message =~ s/[.:;,]$//;
2524
2525 # If haven't seen this exact message before, output it now. Otherwise
2526 # increment the count of how many times it has occurred
2527 unless ($errors{$addr}->{$message}) {
2528 Carp::my_carp("$message in '$_' in "
f998e60c 2529 . $file{$addr}
99870f4d
KW
2530 . " at line $.. Skipping this line;");
2531 $errors{$addr}->{$message} = 1;
2532 }
2533 else {
2534 $errors{$addr}->{$message}++;
2535 }
2536
2537 # Clear the line to prevent any further (meaningful) processing of it.
2538 $_ = "";
2539
2540 return;
2541 }
2542} # End closure
2543
2544package Multi_Default;
2545
2546# Certain properties in early versions of Unicode had more than one possible
2547# default for code points missing from the files. In these cases, one
2548# default applies to everything left over after all the others are applied,
2549# and for each of the others, there is a description of which class of code
2550# points applies to it. This object helps implement this by storing the
2551# defaults, and for all but that final default, an eval string that generates
2552# the class that it applies to.
2553
2554
2555{ # Closure
2556
2557 main::setup_package();
2558
2559 my %class_defaults;
2560 # The defaults structure for the classes
2561 main::set_access('class_defaults', \%class_defaults);
2562
2563 my %other_default;
2564 # The default that applies to everything left over.
2565 main::set_access('other_default', \%other_default, 'r');
2566
2567
2568 sub new {
2569 # The constructor is called with default => eval pairs, terminated by
2570 # the left-over default. e.g.
2571 # Multi_Default->new(
2572 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
2573 # - 0x200D',
2574 # 'R' => 'some other expression that evaluates to code points',
2575 # .
2576 # .
2577 # .
2578 # 'U'));
2579
2580 my $class = shift;
2581
2582 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 2583 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2584
2585 while (@_ > 1) {
2586 my $default = shift;
2587 my $eval = shift;
2588 $class_defaults{$addr}->{$default} = $eval;
2589 }
2590
2591 $other_default{$addr} = shift;
2592
2593 return $self;
2594 }
2595
2596 sub get_next_defaults {
2597 # Iterates and returns the next class of defaults.
2598 my $self = shift;
2599 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2600
ffe43484 2601 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2602
2603 return each %{$class_defaults{$addr}};
2604 }
2605}
2606
2607package Alias;
2608
2609# An alias is one of the names that a table goes by. This class defines them
2610# including some attributes. Everything is currently setup in the
2611# constructor.
2612
2613
2614{ # Closure
2615
2616 main::setup_package();
2617
2618 my %name;
2619 main::set_access('name', \%name, 'r');
2620
2621 my %loose_match;
2622 # Determined by the constructor code if this name should match loosely or
2623 # not. The constructor parameters can override this, but it isn't fully
2624 # implemented, as should have ability to override Unicode one's via
2625 # something like a set_loose_match()
2626 main::set_access('loose_match', \%loose_match, 'r');
2627
2628 my %make_pod_entry;
2629 # Some aliases should not get their own entries because they are covered
2630 # by a wild-card, and some we want to discourage use of. Binary
2631 main::set_access('make_pod_entry', \%make_pod_entry, 'r');
2632
2633 my %status;
2634 # Aliases have a status, like deprecated, or even suppressed (which means
2635 # they don't appear in documentation). Enum
2636 main::set_access('status', \%status, 'r');
2637
2638 my %externally_ok;
2639 # Similarly, some aliases should not be considered as usable ones for
2640 # external use, such as file names, or we don't want documentation to
2641 # recommend them. Boolean
2642 main::set_access('externally_ok', \%externally_ok, 'r');
2643
2644 sub new {
2645 my $class = shift;
2646
2647 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2648 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2649
2650 $name{$addr} = shift;
2651 $loose_match{$addr} = shift;
2652 $make_pod_entry{$addr} = shift;
2653 $externally_ok{$addr} = shift;
2654 $status{$addr} = shift;
2655
2656 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2657
2658 # Null names are never ok externally
2659 $externally_ok{$addr} = 0 if $name{$addr} eq "";
2660
2661 return $self;
2662 }
2663}
2664
2665package Range;
2666
2667# A range is the basic unit for storing code points, and is described in the
2668# comments at the beginning of the program. Each range has a starting code
2669# point; an ending code point (not less than the starting one); a value
2670# that applies to every code point in between the two end-points, inclusive;
2671# and an enum type that applies to the value. The type is for the user's
2672# convenience, and has no meaning here, except that a non-zero type is
2673# considered to not obey the normal Unicode rules for having standard forms.
2674#
2675# The same structure is used for both map and match tables, even though in the
2676# latter, the value (and hence type) is irrelevant and could be used as a
2677# comment. In map tables, the value is what all the code points in the range
2678# map to. Type 0 values have the standardized version of the value stored as
2679# well, so as to not have to recalculate it a lot.
2680
2681sub trace { return main::trace(@_); }
2682
2683{ # Closure
2684
2685 main::setup_package();
2686
2687 my %start;
2688 main::set_access('start', \%start, 'r', 's');
2689
2690 my %end;
2691 main::set_access('end', \%end, 'r', 's');
2692
2693 my %value;
2694 main::set_access('value', \%value, 'r');
2695
2696 my %type;
2697 main::set_access('type', \%type, 'r');
2698
2699 my %standard_form;
2700 # The value in internal standard form. Defined only if the type is 0.
2701 main::set_access('standard_form', \%standard_form);
2702
2703 # Note that if these fields change, the dump() method should as well
2704
2705 sub new {
2706 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
2707 my $class = shift;
2708
2709 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2710 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2711
2712 $start{$addr} = shift;
2713 $end{$addr} = shift;
2714
2715 my %args = @_;
2716
2717 my $value = delete $args{'Value'}; # Can be 0
2718 $value = "" unless defined $value;
2719 $value{$addr} = $value;
2720
2721 $type{$addr} = delete $args{'Type'} || 0;
2722
2723 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2724
2725 if (! $type{$addr}) {
2726 $standard_form{$addr} = main::standardize($value);
2727 }
2728
2729 return $self;
2730 }
2731
2732 use overload
2733 fallback => 0,
2734 qw("") => "_operator_stringify",
2735 "." => \&main::_operator_dot,
2736 ;
2737
2738 sub _operator_stringify {
2739 my $self = shift;
ffe43484 2740 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2741
2742 # Output it like '0041..0065 (value)'
2743 my $return = sprintf("%04X", $start{$addr})
2744 . '..'
2745 . sprintf("%04X", $end{$addr});
2746 my $value = $value{$addr};
2747 my $type = $type{$addr};
2748 $return .= ' (';
2749 $return .= "$value";
2750 $return .= ", Type=$type" if $type != 0;
2751 $return .= ')';
2752
2753 return $return;
2754 }
2755
2756 sub standard_form {
2757 # The standard form is the value itself if the standard form is
2758 # undefined (that is if the value is special)
2759
2760 my $self = shift;
2761 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2762
ffe43484 2763 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2764
2765 return $standard_form{$addr} if defined $standard_form{$addr};
2766 return $value{$addr};
2767 }
2768
2769 sub dump {
2770 # Human, not machine readable. For machine readable, comment out this
2771 # entire routine and let the standard one take effect.
2772 my $self = shift;
2773 my $indent = shift;
2774 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2775
ffe43484 2776 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2777
2778 my $return = $indent
2779 . sprintf("%04X", $start{$addr})
2780 . '..'
2781 . sprintf("%04X", $end{$addr})
2782 . " '$value{$addr}';";
2783 if (! defined $standard_form{$addr}) {
2784 $return .= "(type=$type{$addr})";
2785 }
2786 elsif ($standard_form{$addr} ne $value{$addr}) {
2787 $return .= "(standard '$standard_form{$addr}')";
2788 }
2789 return $return;
2790 }
2791} # End closure
2792
2793package _Range_List_Base;
2794
2795# Base class for range lists. A range list is simply an ordered list of
2796# ranges, so that the ranges with the lowest starting numbers are first in it.
2797#
2798# When a new range is added that is adjacent to an existing range that has the
2799# same value and type, it merges with it to form a larger range.
2800#
2801# Ranges generally do not overlap, except that there can be multiple entries
2802# of single code point ranges. This is because of NameAliases.txt.
2803#
2804# In this program, there is a standard value such that if two different
2805# values, have the same standard value, they are considered equivalent. This
2806# value was chosen so that it gives correct results on Unicode data
2807
2808# There are a number of methods to manipulate range lists, and some operators
2809# are overloaded to handle them.
2810
99870f4d
KW
2811sub trace { return main::trace(@_); }
2812
2813{ # Closure
2814
2815 our $addr;
2816
2817 main::setup_package();
2818
2819 my %ranges;
2820 # The list of ranges
2821 main::set_access('ranges', \%ranges, 'readable_array');
2822
2823 my %max;
2824 # The highest code point in the list. This was originally a method, but
2825 # actual measurements said it was used a lot.
2826 main::set_access('max', \%max, 'r');
2827
2828 my %each_range_iterator;
2829 # Iterator position for each_range()
2830 main::set_access('each_range_iterator', \%each_range_iterator);
2831
2832 my %owner_name_of;
2833 # Name of parent this is attached to, if any. Solely for better error
2834 # messages.
2835 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
2836
2837 my %_search_ranges_cache;
2838 # A cache of the previous result from _search_ranges(), for better
2839 # performance
2840 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
2841
2842 sub new {
2843 my $class = shift;
2844 my %args = @_;
2845
2846 # Optional initialization data for the range list.
2847 my $initialize = delete $args{'Initialize'};
2848
2849 my $self;
2850
2851 # Use _union() to initialize. _union() returns an object of this
2852 # class, which means that it will call this constructor recursively.
2853 # But it won't have this $initialize parameter so that it won't
2854 # infinitely loop on this.
2855 return _union($class, $initialize, %args) if defined $initialize;
2856
2857 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2858 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2859
2860 # Optional parent object, only for debug info.
2861 $owner_name_of{$addr} = delete $args{'Owner'};
2862 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
2863
2864 # Stringify, in case it is an object.
2865 $owner_name_of{$addr} = "$owner_name_of{$addr}";
2866
2867 # This is used only for error messages, and so a colon is added
2868 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
2869
2870 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2871
2872 # Max is initialized to a negative value that isn't adjacent to 0,
2873 # for simpler tests
2874 $max{$addr} = -2;
2875
2876 $_search_ranges_cache{$addr} = 0;
2877 $ranges{$addr} = [];
2878
2879 return $self;
2880 }
2881
2882 use overload
2883 fallback => 0,
2884 qw("") => "_operator_stringify",
2885 "." => \&main::_operator_dot,
2886 ;
2887
2888 sub _operator_stringify {
2889 my $self = shift;
ffe43484 2890 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2891
2892 return "Range_List attached to '$owner_name_of{$addr}'"
2893 if $owner_name_of{$addr};
2894 return "anonymous Range_List " . \$self;
2895 }
2896
2897 sub _union {
2898 # Returns the union of the input code points. It can be called as
2899 # either a constructor or a method. If called as a method, the result
2900 # will be a new() instance of the calling object, containing the union
2901 # of that object with the other parameter's code points; if called as
2902 # a constructor, the first parameter gives the class the new object
2903 # should be, and the second parameter gives the code points to go into
2904 # it.
2905 # In either case, there are two parameters looked at by this routine;
2906 # any additional parameters are passed to the new() constructor.
2907 #
2908 # The code points can come in the form of some object that contains
2909 # ranges, and has a conventionally named method to access them; or
2910 # they can be an array of individual code points (as integers); or
2911 # just a single code point.
2912 #
2913 # If they are ranges, this routine doesn't make any effort to preserve
2914 # the range values of one input over the other. Therefore this base
2915 # class should not allow _union to be called from other than
2916 # initialization code, so as to prevent two tables from being added
2917 # together where the range values matter. The general form of this
2918 # routine therefore belongs in a derived class, but it was moved here
2919 # to avoid duplication of code. The failure to overload this in this
2920 # class keeps it safe.
2921 #
2922
2923 my $self;
2924 my @args; # Arguments to pass to the constructor
2925
2926 my $class = shift;
2927
2928 # If a method call, will start the union with the object itself, and
2929 # the class of the new object will be the same as self.
2930 if (ref $class) {
2931 $self = $class;
2932 $class = ref $self;
2933 push @args, $self;
2934 }
2935
2936 # Add the other required parameter.
2937 push @args, shift;
2938 # Rest of parameters are passed on to the constructor
2939
2940 # Accumulate all records from both lists.
2941 my @records;
2942 for my $arg (@args) {
2943 #local $to_trace = 0 if main::DEBUG;
2944 trace "argument = $arg" if main::DEBUG && $to_trace;
2945 if (! defined $arg) {
2946 my $message = "";
2947 if (defined $self) {
f998e60c 2948 no overloading;
051df77b 2949 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2950 }
2951 Carp::my_carp_bug($message .= "Undefined argument to _union. No union done.");
2952 return;
2953 }
2954 $arg = [ $arg ] if ! ref $arg;
2955 my $type = ref $arg;
2956 if ($type eq 'ARRAY') {
2957 foreach my $element (@$arg) {
2958 push @records, Range->new($element, $element);
2959 }
2960 }
2961 elsif ($arg->isa('Range')) {
2962 push @records, $arg;
2963 }
2964 elsif ($arg->can('ranges')) {
2965 push @records, $arg->ranges;
2966 }
2967 else {
2968 my $message = "";
2969 if (defined $self) {
f998e60c 2970 no overloading;
051df77b 2971 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2972 }
2973 Carp::my_carp_bug($message . "Cannot take the union of a $type. No union done.");
2974 return;
2975 }
2976 }
2977
2978 # Sort with the range containing the lowest ordinal first, but if
2979 # two ranges start at the same code point, sort with the bigger range
2980 # of the two first, because it takes fewer cycles.
2981 @records = sort { ($a->start <=> $b->start)
2982 or
2983 # if b is shorter than a, b->end will be
2984 # less than a->end, and we want to select
2985 # a, so want to return -1
2986 ($b->end <=> $a->end)
2987 } @records;
2988
2989 my $new = $class->new(@_);
2990
2991 # Fold in records so long as they add new information.
2992 for my $set (@records) {
2993 my $start = $set->start;
2994 my $end = $set->end;
2995 my $value = $set->value;
2996 if ($start > $new->max) {
2997 $new->_add_delete('+', $start, $end, $value);
2998 }
2999 elsif ($end > $new->max) {
3000 $new->_add_delete('+', $new->max +1, $end, $value);
3001 }
3002 }
3003
3004 return $new;
3005 }
3006
3007 sub range_count { # Return the number of ranges in the range list
3008 my $self = shift;
3009 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3010
f998e60c 3011 no overloading;
051df77b 3012 return scalar @{$ranges{pack 'J', $self}};
99870f4d
KW
3013 }
3014
3015 sub min {
3016 # Returns the minimum code point currently in the range list, or if
3017 # the range list is empty, 2 beyond the max possible. This is a
3018 # method because used so rarely, that not worth saving between calls,
3019 # and having to worry about changing it as ranges are added and
3020 # deleted.
3021
3022 my $self = shift;
3023 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3024
ffe43484 3025 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3026
3027 # If the range list is empty, return a large value that isn't adjacent
3028 # to any that could be in the range list, for simpler tests
3029 return $LAST_UNICODE_CODEPOINT + 2 unless scalar @{$ranges{$addr}};
3030 return $ranges{$addr}->[0]->start;
3031 }
3032
3033 sub contains {
3034 # Boolean: Is argument in the range list? If so returns $i such that:
3035 # range[$i]->end < $codepoint <= range[$i+1]->end
3036 # which is one beyond what you want; this is so that the 0th range
3037 # doesn't return false
3038 my $self = shift;
3039 my $codepoint = shift;
3040 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3041
99870f4d
KW
3042 my $i = $self->_search_ranges($codepoint);
3043 return 0 unless defined $i;
3044
3045 # The search returns $i, such that
3046 # range[$i-1]->end < $codepoint <= range[$i]->end
3047 # So is in the table if and only iff it is at least the start position
3048 # of range $i.
f998e60c 3049 no overloading;
051df77b 3050 return 0 if $ranges{pack 'J', $self}->[$i]->start > $codepoint;
99870f4d
KW
3051 return $i + 1;
3052 }
3053
2f7a8815
KW
3054 sub containing_range {
3055 # Returns the range object that contains the code point, undef if none
3056
3057 my $self = shift;
3058 my $codepoint = shift;
3059 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3060
3061 my $i = $self->contains($codepoint);
3062 return unless $i;
3063
3064 # contains() returns 1 beyond where we should look
3065 no overloading;
3066 return $ranges{pack 'J', $self}->[$i-1];
3067 }
3068
99870f4d
KW
3069 sub value_of {
3070 # Returns the value associated with the code point, undef if none
3071
3072 my $self = shift;
3073 my $codepoint = shift;
3074 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3075
d69c231b
KW
3076 my $range = $self->containing_range($codepoint);
3077 return unless defined $range;
99870f4d 3078
d69c231b 3079 return $range->value;
99870f4d
KW
3080 }
3081
0a9dbafc
KW
3082 sub type_of {
3083 # Returns the type of the range containing the code point, undef if
3084 # the code point is not in the table
3085
3086 my $self = shift;
3087 my $codepoint = shift;
3088 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3089
3090 my $range = $self->containing_range($codepoint);
3091 return unless defined $range;
3092
3093 return $range->type;
3094 }
3095
99870f4d
KW
3096 sub _search_ranges {
3097 # Find the range in the list which contains a code point, or where it
3098 # should go if were to add it. That is, it returns $i, such that:
3099 # range[$i-1]->end < $codepoint <= range[$i]->end
3100 # Returns undef if no such $i is possible (e.g. at end of table), or
3101 # if there is an error.
3102
3103 my $self = shift;
3104 my $code_point = shift;
3105 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3106
ffe43484 3107 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3108
3109 return if $code_point > $max{$addr};
3110 my $r = $ranges{$addr}; # The current list of ranges
3111 my $range_list_size = scalar @$r;
3112 my $i;
3113
3114 use integer; # want integer division
3115
3116 # Use the cached result as the starting guess for this one, because,
3117 # an experiment on 5.1 showed that 90% of the time the cache was the
3118 # same as the result on the next call (and 7% it was one less).
3119 $i = $_search_ranges_cache{$addr};
3120 $i = 0 if $i >= $range_list_size; # Reset if no longer valid (prob.
3121 # from an intervening deletion
3122 #local $to_trace = 1 if main::DEBUG;
3123 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);
3124 return $i if $code_point <= $r->[$i]->end
3125 && ($i == 0 || $r->[$i-1]->end < $code_point);
3126
3127 # Here the cache doesn't yield the correct $i. Try adding 1.
3128 if ($i < $range_list_size - 1
3129 && $r->[$i]->end < $code_point &&
3130 $code_point <= $r->[$i+1]->end)
3131 {
3132 $i++;
3133 trace "next \$i is correct: $i" if main::DEBUG && $to_trace;
3134 $_search_ranges_cache{$addr} = $i;
3135 return $i;
3136 }
3137
3138 # Here, adding 1 also didn't work. We do a binary search to
3139 # find the correct position, starting with current $i
3140 my $lower = 0;
3141 my $upper = $range_list_size - 1;
3142 while (1) {
3143 trace "top of loop i=$i:", sprintf("%04X", $r->[$lower]->start), "[$lower] .. ", sprintf("%04X", $r->[$i]->start), "[$i] .. ", sprintf("%04X", $r->[$upper]->start), "[$upper]" if main::DEBUG && $to_trace;
3144
3145 if ($code_point <= $r->[$i]->end) {
3146
3147 # Here we have met the upper constraint. We can quit if we
3148 # also meet the lower one.
3149 last if $i == 0 || $r->[$i-1]->end < $code_point;
3150
3151 $upper = $i; # Still too high.
3152
3153 }
3154 else {
3155
3156 # Here, $r[$i]->end < $code_point, so look higher up.
3157 $lower = $i;
3158 }
3159
3160 # Split search domain in half to try again.
3161 my $temp = ($upper + $lower) / 2;
3162
3163 # No point in continuing unless $i changes for next time
3164 # in the loop.
3165 if ($temp == $i) {
3166
3167 # We can't reach the highest element because of the averaging.
3168 # So if one below the upper edge, force it there and try one
3169 # more time.
3170 if ($i == $range_list_size - 2) {
3171
3172 trace "Forcing to upper edge" if main::DEBUG && $to_trace;
3173 $i = $range_list_size - 1;
3174
3175 # Change $lower as well so if fails next time through,
3176 # taking the average will yield the same $i, and we will
3177 # quit with the error message just below.
3178 $lower = $i;
3179 next;
3180 }
3181 Carp::my_carp_bug("$owner_name_of{$addr}Can't find where the range ought to go. No action taken.");
3182 return;
3183 }
3184 $i = $temp;
3185 } # End of while loop
3186
3187 if (main::DEBUG && $to_trace) {
3188 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i;
3189 trace "i= [ $i ]", $r->[$i];
3190 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < $range_list_size - 1;
3191 }
3192
3193 # Here we have found the offset. Cache it as a starting point for the
3194 # next call.
3195 $_search_ranges_cache{$addr} = $i;
3196 return $i;
3197 }
3198
3199 sub _add_delete {
3200 # Add, replace or delete ranges to or from a list. The $type
3201 # parameter gives which:
3202 # '+' => insert or replace a range, returning a list of any changed
3203 # ranges.
3204 # '-' => delete a range, returning a list of any deleted ranges.
3205 #
3206 # The next three parameters give respectively the start, end, and
3207 # value associated with the range. 'value' should be null unless the
3208 # operation is '+';
3209 #
3210 # The range list is kept sorted so that the range with the lowest
3211 # starting position is first in the list, and generally, adjacent
c1739a4a 3212 # ranges with the same values are merged into a single larger one (see
99870f4d
KW
3213 # exceptions below).
3214 #
c1739a4a 3215 # There are more parameters; all are key => value pairs:
99870f4d
KW
3216 # Type gives the type of the value. It is only valid for '+'.
3217 # All ranges have types; if this parameter is omitted, 0 is
3218 # assumed. Ranges with type 0 are assumed to obey the
3219 # Unicode rules for casing, etc; ranges with other types are
3220 # not. Otherwise, the type is arbitrary, for the caller's
3221 # convenience, and looked at only by this routine to keep
3222 # adjacent ranges of different types from being merged into
3223 # a single larger range, and when Replace =>
3224 # $IF_NOT_EQUIVALENT is specified (see just below).
3225 # Replace determines what to do if the range list already contains
3226 # ranges which coincide with all or portions of the input
3227 # range. It is only valid for '+':
3228 # => $NO means that the new value is not to replace
3229 # any existing ones, but any empty gaps of the
3230 # range list coinciding with the input range
3231 # will be filled in with the new value.
3232 # => $UNCONDITIONALLY means to replace the existing values with
3233 # this one unconditionally. However, if the
3234 # new and old values are identical, the
3235 # replacement is skipped to save cycles
3236 # => $IF_NOT_EQUIVALENT means to replace the existing values
3237 # with this one if they are not equivalent.
3238 # Ranges are equivalent if their types are the
c1739a4a 3239 # same, and they are the same string; or if
99870f4d
KW
3240 # both are type 0 ranges, if their Unicode
3241 # standard forms are identical. In this last
3242 # case, the routine chooses the more "modern"
3243 # one to use. This is because some of the
3244 # older files are formatted with values that
3245 # are, for example, ALL CAPs, whereas the
3246 # derived files have a more modern style,
3247 # which looks better. By looking for this
3248 # style when the pre-existing and replacement
3249 # standard forms are the same, we can move to
3250 # the modern style
3251 # => $MULTIPLE means that if this range duplicates an
3252 # existing one, but has a different value,
3253 # don't replace the existing one, but insert
3254 # this, one so that the same range can occur
53d84487
KW
3255 # multiple times. They are stored LIFO, so
3256 # that the final one inserted is the first one
3257 # returned in an ordered search of the table.
99870f4d
KW
3258 # => anything else is the same as => $IF_NOT_EQUIVALENT
3259 #
c1739a4a
KW
3260 # "same value" means identical for non-type-0 ranges, and it means
3261 # having the same standard forms for type-0 ranges.
99870f4d
KW
3262
3263 return Carp::carp_too_few_args(\@_, 5) if main::DEBUG && @_ < 5;
3264
3265 my $self = shift;
3266 my $operation = shift; # '+' for add/replace; '-' for delete;
3267 my $start = shift;
3268 my $end = shift;
3269 my $value = shift;
3270
3271 my %args = @_;
3272
3273 $value = "" if not defined $value; # warning: $value can be "0"
3274
3275 my $replace = delete $args{'Replace'};
3276 $replace = $IF_NOT_EQUIVALENT unless defined $replace;
3277
3278 my $type = delete $args{'Type'};
3279 $type = 0 unless defined $type;
3280
3281 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3282
ffe43484 3283 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3284
3285 if ($operation ne '+' && $operation ne '-') {
3286 Carp::my_carp_bug("$owner_name_of{$addr}First parameter to _add_delete must be '+' or '-'. No action taken.");
3287 return;
3288 }
3289 unless (defined $start && defined $end) {
3290 Carp::my_carp_bug("$owner_name_of{$addr}Undefined start and/or end to _add_delete. No action taken.");
3291 return;
3292 }
3293 unless ($end >= $start) {
3294 Carp::my_carp_bug("$owner_name_of{$addr}End of range (" . sprintf("%04X", $end) . ") must not be before start (" . sprintf("%04X", $start) . "). No action taken.");
3295 return;
3296 }
3297 #local $to_trace = 1 if main::DEBUG;
3298
3299 if ($operation eq '-') {
3300 if ($replace != $IF_NOT_EQUIVALENT) {
3301 Carp::my_carp_bug("$owner_name_of{$addr}Replace => \$IF_NOT_EQUIVALENT is required when deleting a range from a range list. Assuming Replace => \$IF_NOT_EQUIVALENT.");
3302 $replace = $IF_NOT_EQUIVALENT;
3303 }
3304 if ($type) {
3305 Carp::my_carp_bug("$owner_name_of{$addr}Type => 0 is required when deleting a range from a range list. Assuming Type => 0.");
3306 $type = 0;
3307 }
3308 if ($value ne "") {
3309 Carp::my_carp_bug("$owner_name_of{$addr}Value => \"\" is required when deleting a range from a range list. Assuming Value => \"\".");
3310 $value = "";
3311 }
3312 }
3313
3314 my $r = $ranges{$addr}; # The current list of ranges
3315 my $range_list_size = scalar @$r; # And its size
3316 my $max = $max{$addr}; # The current high code point in
3317 # the list of ranges
3318
3319 # Do a special case requiring fewer machine cycles when the new range
3320 # starts after the current highest point. The Unicode input data is
3321 # structured so this is common.
3322 if ($start > $max) {
3323
3324 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) type=$type" if main::DEBUG && $to_trace;
3325 return if $operation eq '-'; # Deleting a non-existing range is a
3326 # no-op
3327
3328 # If the new range doesn't logically extend the current final one
3329 # in the range list, create a new range at the end of the range
3330 # list. (max cleverly is initialized to a negative number not
3331 # adjacent to 0 if the range list is empty, so even adding a range
3332 # to an empty range list starting at 0 will have this 'if'
3333 # succeed.)
3334 if ($start > $max + 1 # non-adjacent means can't extend.
3335 || @{$r}[-1]->value ne $value # values differ, can't extend.
3336 || @{$r}[-1]->type != $type # types differ, can't extend.
3337 ) {
3338 push @$r, Range->new($start, $end,
3339 Value => $value,
3340 Type => $type);
3341 }
3342 else {
3343
3344 # Here, the new range starts just after the current highest in
3345 # the range list, and they have the same type and value.
3346 # Extend the current range to incorporate the new one.
3347 @{$r}[-1]->set_end($end);
3348 }
3349
3350 # This becomes the new maximum.
3351 $max{$addr} = $end;
3352
3353 return;
3354 }
3355 #local $to_trace = 0 if main::DEBUG;
3356
3357 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) replace=$replace" if main::DEBUG && $to_trace;
3358
3359 # Here, the input range isn't after the whole rest of the range list.
3360 # Most likely 'splice' will be needed. The rest of the routine finds
3361 # the needed splice parameters, and if necessary, does the splice.
3362 # First, find the offset parameter needed by the splice function for
3363 # the input range. Note that the input range may span multiple
3364 # existing ones, but we'll worry about that later. For now, just find
3365 # the beginning. If the input range is to be inserted starting in a
3366 # position not currently in the range list, it must (obviously) come
3367 # just after the range below it, and just before the range above it.
3368 # Slightly less obviously, it will occupy the position currently
3369 # occupied by the range that is to come after it. More formally, we
3370 # are looking for the position, $i, in the array of ranges, such that:
3371 #
3372 # r[$i-1]->start <= r[$i-1]->end < $start < r[$i]->start <= r[$i]->end
3373 #
3374 # (The ordered relationships within existing ranges are also shown in
3375 # the equation above). However, if the start of the input range is
3376 # within an existing range, the splice offset should point to that
3377 # existing range's position in the list; that is $i satisfies a
3378 # somewhat different equation, namely:
3379 #
3380 #r[$i-1]->start <= r[$i-1]->end < r[$i]->start <= $start <= r[$i]->end
3381 #
3382 # More briefly, $start can come before or after r[$i]->start, and at
3383 # this point, we don't know which it will be. However, these
3384 # two equations share these constraints:
3385 #
3386 # r[$i-1]->end < $start <= r[$i]->end
3387 #
3388 # And that is good enough to find $i.
3389
3390 my $i = $self->_search_ranges($start);
3391 if (! defined $i) {
3392 Carp::my_carp_bug("Searching $self for range beginning with $start unexpectedly returned undefined. Operation '$operation' not performed");
3393 return;
3394 }
3395
3396 # The search function returns $i such that:
3397 #
3398 # r[$i-1]->end < $start <= r[$i]->end
3399 #
3400 # That means that $i points to the first range in the range list
3401 # that could possibly be affected by this operation. We still don't
3402 # know if the start of the input range is within r[$i], or if it
3403 # points to empty space between r[$i-1] and r[$i].
3404 trace "[$i] is the beginning splice point. Existing range there is ", $r->[$i] if main::DEBUG && $to_trace;
3405
3406 # Special case the insertion of data that is not to replace any
3407 # existing data.
3408 if ($replace == $NO) { # If $NO, has to be operation '+'
3409 #local $to_trace = 1 if main::DEBUG;
3410 trace "Doesn't replace" if main::DEBUG && $to_trace;
3411
3412 # Here, the new range is to take effect only on those code points
3413 # that aren't already in an existing range. This can be done by
3414 # looking through the existing range list and finding the gaps in
3415 # the ranges that this new range affects, and then calling this
3416 # function recursively on each of those gaps, leaving untouched
3417 # anything already in the list. Gather up a list of the changed
3418 # gaps first so that changes to the internal state as new ranges
3419 # are added won't be a problem.
3420 my @gap_list;
3421
3422 # First, if the starting point of the input range is outside an
3423 # existing one, there is a gap from there to the beginning of the
3424 # existing range -- add a span to fill the part that this new
3425 # range occupies
3426 if ($start < $r->[$i]->start) {
3427 push @gap_list, Range->new($start,
3428 main::min($end,
3429 $r->[$i]->start - 1),
3430 Type => $type);
3431 trace "gap before $r->[$i] [$i], will add", $gap_list[-1] if main::DEBUG && $to_trace;
3432 }
3433
3434 # Then look through the range list for other gaps until we reach
3435 # the highest range affected by the input one.
3436 my $j;
3437 for ($j = $i+1; $j < $range_list_size; $j++) {
3438 trace "j=[$j]", $r->[$j] if main::DEBUG && $to_trace;
3439 last if $end < $r->[$j]->start;
3440
3441 # If there is a gap between when this range starts and the
3442 # previous one ends, add a span to fill it. Note that just
3443 # because there are two ranges doesn't mean there is a
3444 # non-zero gap between them. It could be that they have
3445 # different values or types
3446 if ($r->[$j-1]->end + 1 != $r->[$j]->start) {
3447 push @gap_list,
3448 Range->new($r->[$j-1]->end + 1,
3449 $r->[$j]->start - 1,
3450 Type => $type);
3451 trace "gap between $r->[$j-1] and $r->[$j] [$j], will add: $gap_list[-1]" if main::DEBUG && $to_trace;
3452 }
3453 }
3454
3455 # Here, we have either found an existing range in the range list,
3456 # beyond the area affected by the input one, or we fell off the
3457 # end of the loop because the input range affects the whole rest
3458 # of the range list. In either case, $j is 1 higher than the
3459 # highest affected range. If $j == $i, it means that there are no
3460 # affected ranges, that the entire insertion is in the gap between
3461 # r[$i-1], and r[$i], which we already have taken care of before
3462 # the loop.
3463 # On the other hand, if there are affected ranges, it might be
3464 # that there is a gap that needs filling after the final such
3465 # range to the end of the input range
3466 if ($r->[$j-1]->end < $end) {
3467 push @gap_list, Range->new(main::max($start,
3468 $r->[$j-1]->end + 1),
3469 $end,
3470 Type => $type);
3471 trace "gap after $r->[$j-1], will add $gap_list[-1]" if main::DEBUG && $to_trace;
3472 }
3473
3474 # Call recursively to fill in all the gaps.
3475 foreach my $gap (@gap_list) {
3476 $self->_add_delete($operation,
3477 $gap->start,
3478 $gap->end,
3479 $value,
3480 Type => $type);
3481 }
3482
3483 return;
3484 }
3485
53d84487
KW
3486 # Here, we have taken care of the case where $replace is $NO.
3487 # Remember that here, r[$i-1]->end < $start <= r[$i]->end
3488 # If inserting a multiple record, this is where it goes, before the
3489 # first (if any) existing one. This implies an insertion, and no
3490 # change to any existing ranges. Note that $i can be -1 if this new
3491 # range doesn't actually duplicate any existing, and comes at the
3492 # beginning of the list.
3493 if ($replace == $MULTIPLE) {
3494
3495 if ($start != $end) {
3496 Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple record when the range ($start..$end) contains more than one code point. No action taken.");
3497 return;
3498 }
3499
3500 # Don't add an exact duplicate, as it isn't really a multiple
3501 if ($end >= $r->[$i]->start) {
3502 if ($r->[$i]->start != $r->[$i]->end) {
3503 Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple record when the other range ($r->[$i]) contains more than one code point. No action taken.");
3504 return;
3505 }
3506 return if $value eq $r->[$i]->value && $type eq $r->[$i]->type;
3507 }
3508
3509 trace "Adding multiple record at $i with $start..$end, $value" if main::DEBUG && $to_trace;
3510 my @return = splice @$r,
3511 $i,
3512 0,
3513 Range->new($start,
3514 $end,
3515 Value => $value,
3516 Type => $type);
3517 if (main::DEBUG && $to_trace) {
3518 trace "After splice:";
3519 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3520 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3521 trace "i =[", $i, "]", $r->[$i] if $i >= 0;
3522 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3523 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3524 trace 'i+3=[', $i+3, ']', $r->[$i+3] if $i < @$r - 3;
3525 }
3526 return @return;
3527 }
3528
3529 # Here, we have taken care of $NO and $MULTIPLE replaces. This leaves
3530 # delete, insert, and replace either unconditionally or if not
3531 # equivalent. $i still points to the first potential affected range.
3532 # Now find the highest range affected, which will determine the length
3533 # parameter to splice. (The input range can span multiple existing
3534 # ones.) If this isn't a deletion, while we are looking through the
3535 # range list, see also if this is a replacement rather than a clean
3536 # insertion; that is if it will change the values of at least one
3537 # existing range. Start off assuming it is an insert, until find it
3538 # isn't.
3539 my $clean_insert = $operation eq '+';
99870f4d
KW
3540 my $j; # This will point to the highest affected range
3541
3542 # For non-zero types, the standard form is the value itself;
3543 my $standard_form = ($type) ? $value : main::standardize($value);
3544
3545 for ($j = $i; $j < $range_list_size; $j++) {
3546 trace "Looking for highest affected range; the one at $j is ", $r->[$j] if main::DEBUG && $to_trace;
3547
3548 # If find a range that it doesn't overlap into, we can stop
3549 # searching
3550 last if $end < $r->[$j]->start;
3551
969a34cc
KW
3552 # Here, overlaps the range at $j. If the values don't match,
3553 # and so far we think this is a clean insertion, it becomes a
3554 # non-clean insertion, i.e., a 'change' or 'replace' instead.
3555 if ($clean_insert) {
99870f4d 3556 if ($r->[$j]->standard_form ne $standard_form) {
969a34cc 3557 $clean_insert = 0;
56343c78
KW
3558 if ($replace == $CROAK) {
3559 main::croak("The range to add "
3560 . sprintf("%04X", $start)
3561 . '-'
3562 . sprintf("%04X", $end)
3563 . " with value '$value' overlaps an existing range $r->[$j]");
3564 }
99870f4d
KW
3565 }
3566 else {
3567
3568 # Here, the two values are essentially the same. If the
3569 # two are actually identical, replacing wouldn't change
3570 # anything so skip it.
3571 my $pre_existing = $r->[$j]->value;
3572 if ($pre_existing ne $value) {
3573
3574 # Here the new and old standardized values are the
3575 # same, but the non-standardized values aren't. If
3576 # replacing unconditionally, then replace
3577 if( $replace == $UNCONDITIONALLY) {
969a34cc 3578 $clean_insert = 0;
99870f4d
KW
3579 }
3580 else {
3581
3582 # Here, are replacing conditionally. Decide to
3583 # replace or not based on which appears to look
3584 # the "nicest". If one is mixed case and the
3585 # other isn't, choose the mixed case one.
3586 my $new_mixed = $value =~ /[A-Z]/
3587 && $value =~ /[a-z]/;
3588 my $old_mixed = $pre_existing =~ /[A-Z]/
3589 && $pre_existing =~ /[a-z]/;
3590
3591 if ($old_mixed != $new_mixed) {
969a34cc 3592 $clean_insert = 0 if $new_mixed;
99870f4d 3593 if (main::DEBUG && $to_trace) {
969a34cc
KW
3594 if ($clean_insert) {
3595 trace "Retaining $pre_existing over $value";
99870f4d
KW
3596 }
3597 else {
969a34cc 3598 trace "Replacing $pre_existing with $value";
99870f4d
KW
3599 }
3600 }
3601 }
3602 else {
3603
3604 # Here casing wasn't different between the two.
3605 # If one has hyphens or underscores and the
3606 # other doesn't, choose the one with the
3607 # punctuation.
3608 my $new_punct = $value =~ /[-_]/;
3609 my $old_punct = $pre_existing =~ /[-_]/;
3610
3611 if ($old_punct != $new_punct) {
969a34cc 3612 $clean_insert = 0 if $new_punct;
99870f4d 3613 if (main::DEBUG && $to_trace) {
969a34cc
KW
3614 if ($clean_insert) {
3615 trace "Retaining $pre_existing over $value";
99870f4d
KW
3616 }
3617 else {
969a34cc 3618 trace "Replacing $pre_existing with $value";
99870f4d
KW
3619 }
3620 }
3621 } # else existing one is just as "good";
3622 # retain it to save cycles.
3623 }
3624 }
3625 }
3626 }
3627 }
3628 } # End of loop looking for highest affected range.
3629
3630 # Here, $j points to one beyond the highest range that this insertion
3631 # affects (hence to beyond the range list if that range is the final
3632 # one in the range list).
3633
3634 # The splice length is all the affected ranges. Get it before
3635 # subtracting, for efficiency, so we don't have to later add 1.
3636 my $length = $j - $i;
3637
3638 $j--; # $j now points to the highest affected range.
3639 trace "Final affected range is $j: $r->[$j]" if main::DEBUG && $to_trace;
3640
99870f4d
KW
3641 # Here, have taken care of $NO and $MULTIPLE replaces.
3642 # $j points to the highest affected range. But it can be < $i or even
3643 # -1. These happen only if the insertion is entirely in the gap
3644 # between r[$i-1] and r[$i]. Here's why: j < i means that the j loop
3645 # above exited first time through with $end < $r->[$i]->start. (And
3646 # then we subtracted one from j) This implies also that $start <
3647 # $r->[$i]->start, but we know from above that $r->[$i-1]->end <
3648 # $start, so the entire input range is in the gap.
3649 if ($j < $i) {
3650
3651 # Here the entire input range is in the gap before $i.
3652
3653 if (main::DEBUG && $to_trace) {
3654 if ($i) {
3655 trace "Entire range is between $r->[$i-1] and $r->[$i]";
3656 }
3657 else {
3658 trace "Entire range is before $r->[$i]";
3659 }
3660 }
3661 return if $operation ne '+'; # Deletion of a non-existent range is
3662 # a no-op
3663 }
3664 else {
3665
969a34cc
KW
3666 # Here part of the input range is not in the gap before $i. Thus,
3667 # there is at least one affected one, and $j points to the highest
3668 # such one.
99870f4d
KW
3669
3670 # At this point, here is the situation:
3671 # This is not an insertion of a multiple, nor of tentative ($NO)
3672 # data.
3673 # $i points to the first element in the current range list that
3674 # may be affected by this operation. In fact, we know
3675 # that the range at $i is affected because we are in
3676 # the else branch of this 'if'
3677 # $j points to the highest affected range.
3678 # In other words,
3679 # r[$i-1]->end < $start <= r[$i]->end
3680 # And:
3681 # r[$i-1]->end < $start <= $end <= r[$j]->end
3682 #
3683 # Also:
969a34cc
KW
3684 # $clean_insert is a boolean which is set true if and only if
3685 # this is a "clean insertion", i.e., not a change nor a
3686 # deletion (multiple was handled above).
99870f4d
KW
3687
3688 # We now have enough information to decide if this call is a no-op
969a34cc
KW
3689 # or not. It is a no-op if this is an insertion of already
3690 # existing data.
99870f4d 3691
969a34cc 3692 if (main::DEBUG && $to_trace && $clean_insert
99870f4d
KW
3693 && $i == $j
3694 && $start >= $r->[$i]->start)
3695 {
3696 trace "no-op";
3697 }
969a34cc 3698 return if $clean_insert
99870f4d
KW
3699 && $i == $j # more than one affected range => not no-op
3700
3701 # Here, r[$i-1]->end < $start <= $end <= r[$i]->end
3702 # Further, $start and/or $end is >= r[$i]->start
3703 # The test below hence guarantees that
3704 # r[$i]->start < $start <= $end <= r[$i]->end
3705 # This means the input range is contained entirely in
3706 # the one at $i, so is a no-op
3707 && $start >= $r->[$i]->start;
3708 }
3709
3710 # Here, we know that some action will have to be taken. We have
3711 # calculated the offset and length (though adjustments may be needed)
3712 # for the splice. Now start constructing the replacement list.
3713 my @replacement;
3714 my $splice_start = $i;
3715
3716 my $extends_below;
3717 my $extends_above;
3718
3719 # See if should extend any adjacent ranges.
3720 if ($operation eq '-') { # Don't extend deletions
3721 $extends_below = $extends_above = 0;
3722 }
3723 else { # Here, should extend any adjacent ranges. See if there are
3724 # any.
3725 $extends_below = ($i > 0
3726 # can't extend unless adjacent
3727 && $r->[$i-1]->end == $start -1
3728 # can't extend unless are same standard value
3729 && $r->[$i-1]->standard_form eq $standard_form
3730 # can't extend unless share type
3731 && $r->[$i-1]->type == $type);
3732 $extends_above = ($j+1 < $range_list_size
3733 && $r->[$j+1]->start == $end +1
3734 && $r->[$j+1]->standard_form eq $standard_form
23822bda 3735 && $r->[$j+1]->type == $type);
99870f4d
KW
3736 }
3737 if ($extends_below && $extends_above) { # Adds to both
3738 $splice_start--; # start replace at element below
3739 $length += 2; # will replace on both sides
3740 trace "Extends both below and above ranges" if main::DEBUG && $to_trace;
3741
3742 # The result will fill in any gap, replacing both sides, and
3743 # create one large range.
3744 @replacement = Range->new($r->[$i-1]->start,
3745 $r->[$j+1]->end,
3746 Value => $value,
3747 Type => $type);
3748 }
3749 else {
3750
3751 # Here we know that the result won't just be the conglomeration of
3752 # a new range with both its adjacent neighbors. But it could
3753 # extend one of them.
3754
3755 if ($extends_below) {
3756
3757 # Here the new element adds to the one below, but not to the
3758 # one above. If inserting, and only to that one range, can
3759 # just change its ending to include the new one.
969a34cc 3760 if ($length == 0 && $clean_insert) {
99870f4d
KW
3761 $r->[$i-1]->set_end($end);
3762 trace "inserted range extends range to below so it is now $r->[$i-1]" if main::DEBUG && $to_trace;
3763 return;
3764 }
3765 else {
3766 trace "Changing inserted range to start at ", sprintf("%04X", $r->[$i-1]->start), " instead of ", sprintf("%04X", $start) if main::DEBUG && $to_trace;
3767 $splice_start--; # start replace at element below
3768 $length++; # will replace the element below
3769 $start = $r->[$i-1]->start;
3770 }
3771 }
3772 elsif ($extends_above) {
3773
3774 # Here the new element adds to the one above, but not below.
3775 # Mirror the code above
969a34cc 3776 if ($length == 0 && $clean_insert) {
99870f4d
KW
3777 $r->[$j+1]->set_start($start);
3778 trace "inserted range extends range to above so it is now $r->[$j+1]" if main::DEBUG && $to_trace;
3779 return;
3780 }
3781 else {
3782 trace "Changing inserted range to end at ", sprintf("%04X", $r->[$j+1]->end), " instead of ", sprintf("%04X", $end) if main::DEBUG && $to_trace;
3783 $length++; # will replace the element above
3784 $end = $r->[$j+1]->end;
3785 }
3786 }
3787
3788 trace "Range at $i is $r->[$i]" if main::DEBUG && $to_trace;
3789
3790 # Finally, here we know there will have to be a splice.
3791 # If the change or delete affects only the highest portion of the
3792 # first affected range, the range will have to be split. The
3793 # splice will remove the whole range, but will replace it by a new
3794 # range containing just the unaffected part. So, in this case,
3795 # add to the replacement list just this unaffected portion.
3796 if (! $extends_below
3797 && $start > $r->[$i]->start && $start <= $r->[$i]->end)
3798 {
3799 push @replacement,
3800 Range->new($r->[$i]->start,
3801 $start - 1,
3802 Value => $r->[$i]->value,
3803 Type => $r->[$i]->type);
3804 }
3805
3806 # In the case of an insert or change, but not a delete, we have to
3807 # put in the new stuff; this comes next.
3808 if ($operation eq '+') {
3809 push @replacement, Range->new($start,
3810 $end,
3811 Value => $value,
3812 Type => $type);
3813 }
3814
3815 trace "Range at $j is $r->[$j]" if main::DEBUG && $to_trace && $j != $i;
3816 #trace "$end >=", $r->[$j]->start, " && $end <", $r->[$j]->end if main::DEBUG && $to_trace;
3817
3818 # And finally, if we're changing or deleting only a portion of the
3819 # highest affected range, it must be split, as the lowest one was.
3820 if (! $extends_above
3821 && $j >= 0 # Remember that j can be -1 if before first
3822 # current element
3823 && $end >= $r->[$j]->start
3824 && $end < $r->[$j]->end)
3825 {
3826 push @replacement,
3827 Range->new($end + 1,
3828 $r->[$j]->end,
3829 Value => $r->[$j]->value,
3830 Type => $r->[$j]->type);
3831 }
3832 }
3833
3834 # And do the splice, as calculated above
3835 if (main::DEBUG && $to_trace) {
3836 trace "replacing $length element(s) at $i with ";
3837 foreach my $replacement (@replacement) {
3838 trace " $replacement";
3839 }
3840 trace "Before splice:";
3841 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3842 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3843 trace "i =[", $i, "]", $r->[$i];
3844 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3845 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3846 }
3847
3848 my @return = splice @$r, $splice_start, $length, @replacement;
3849
3850 if (main::DEBUG && $to_trace) {
3851 trace "After splice:";
3852 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3853 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3854 trace "i =[", $i, "]", $r->[$i];
3855 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3856 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
e6451557 3857 trace "removed ", @return if @return;
99870f4d
KW
3858 }
3859
3860 # An actual deletion could have changed the maximum in the list.
3861 # There was no deletion if the splice didn't return something, but
3862 # otherwise recalculate it. This is done too rarely to worry about
3863 # performance.
3864 if ($operation eq '-' && @return) {
3865 $max{$addr} = $r->[-1]->end;
3866 }
3867 return @return;
3868 }
3869
3870 sub reset_each_range { # reset the iterator for each_range();
3871 my $self = shift;
3872 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3873
f998e60c 3874 no overloading;
051df77b 3875 undef $each_range_iterator{pack 'J', $self};
99870f4d
KW
3876 return;
3877 }
3878
3879 sub each_range {
3880 # Iterate over each range in a range list. Results are undefined if
3881 # the range list is changed during the iteration.
3882
3883 my $self = shift;
3884 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3885
ffe43484 3886 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3887
3888 return if $self->is_empty;
3889
3890 $each_range_iterator{$addr} = -1
3891 if ! defined $each_range_iterator{$addr};
3892 $each_range_iterator{$addr}++;
3893 return $ranges{$addr}->[$each_range_iterator{$addr}]
3894 if $each_range_iterator{$addr} < @{$ranges{$addr}};
3895 undef $each_range_iterator{$addr};
3896 return;
3897 }
3898
3899 sub count { # Returns count of code points in range list
3900 my $self = shift;
3901 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3902
ffe43484 3903 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3904
3905 my $count = 0;
3906 foreach my $range (@{$ranges{$addr}}) {
3907 $count += $range->end - $range->start + 1;
3908 }
3909 return $count;
3910 }
3911
3912 sub delete_range { # Delete a range
3913 my $self = shift;
3914 my $start = shift;
3915 my $end = shift;
3916
3917 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3918
3919 return $self->_add_delete('-', $start, $end, "");
3920 }
3921
3922 sub is_empty { # Returns boolean as to if a range list is empty
3923 my $self = shift;
3924 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3925
f998e60c 3926 no overloading;
051df77b 3927 return scalar @{$ranges{pack 'J', $self}} == 0;
99870f4d
KW
3928 }
3929
3930 sub hash {
3931 # Quickly returns a scalar suitable for separating tables into