This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
mktables: Empty tables are complements of \p{Any}
[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
1c2e8cca 593# by the code points introduced in the later version. Change the 0 to a SINGLE
99870f4d
KW
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
8c32d378
KW
1100# to decrease backtracking. The first one allows the code point to be at the
1101# end of a word, but to work properly, the word shouldn't end with a valid hex
1102# character. The second one won't match a code point at the end of a word,
1103# and doesn't have the run-on issue
1104my $run_on_code_point_re =
1105 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1106my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d
KW
1107
1108# This matches the beginning of the line in the Unicode db files that give the
1109# defaults for code points not listed (i.e., missing) in the file. The code
1110# depends on this ending with a semi-colon, so it can assume it is a valid
1111# field when the line is split() by semi-colons
1112my $missing_defaults_prefix =
1113 qr/^#\s+\@missing:\s+0000\.\.$LAST_UNICODE_CODEPOINT_STRING\s*;/;
1114
1115# Property types. Unicode has more types, but these are sufficient for our
1116# purposes.
1117my $UNKNOWN = -1; # initialized to illegal value
1118my $NON_STRING = 1; # Either binary or enum
1119my $BINARY = 2;
1120my $ENUM = 3; # Include catalog
1121my $STRING = 4; # Anything else: string or misc
1122
1123# Some input files have lines that give default values for code points not
1124# contained in the file. Sometimes these should be ignored.
1125my $NO_DEFAULTS = 0; # Must evaluate to false
1126my $NOT_IGNORED = 1;
1127my $IGNORED = 2;
1128
1129# Range types. Each range has a type. Most ranges are type 0, for normal,
1130# and will appear in the main body of the tables in the output files, but
1131# there are other types of ranges as well, listed below, that are specially
1132# handled. There are pseudo-types as well that will never be stored as a
1133# type, but will affect the calculation of the type.
1134
1135# 0 is for normal, non-specials
1136my $MULTI_CP = 1; # Sequence of more than code point
1137my $HANGUL_SYLLABLE = 2;
1138my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1139my $NULL = 4; # The map is to the null string; utf8.c can't
1140 # handle these, nor is there an accepted syntax
1141 # for them in \p{} constructs
f86864ac 1142my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1143 # otherwise be $MULTI_CP type are instead type 0
1144
1145# process_generic_property_file() can accept certain overrides in its input.
1146# Each of these must begin AND end with $CMD_DELIM.
1147my $CMD_DELIM = "\a";
1148my $REPLACE_CMD = 'replace'; # Override the Replace
1149my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1150
1151my $NO = 0;
1152my $YES = 1;
1153
1154# Values for the Replace argument to add_range.
1155# $NO # Don't replace; add only the code points not
1156 # already present.
1157my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1158 # the comments at the subroutine definition.
1159my $UNCONDITIONALLY = 2; # Replace without conditions.
1160my $MULTIPLE = 4; # Don't replace, but add a duplicate record if
1161 # already there
56343c78 1162my $CROAK = 5; # Die with an error if is already there
99870f4d
KW
1163
1164# Flags to give property statuses. The phrases are to remind maintainers that
1165# if the flag is changed, the indefinite article referring to it in the
1166# documentation may need to be as well.
1167my $NORMAL = "";
1168my $SUPPRESSED = 'z'; # The character should never actually be seen, since
1169 # it is suppressed
37e2e78e 1170my $PLACEHOLDER = 'P'; # Implies no pod entry generated
99870f4d
KW
1171my $DEPRECATED = 'D';
1172my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1173my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1174my $DISCOURAGED = 'X';
1175my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1176my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1177my $STRICTER = 'T';
1178my $a_bold_stricter = "a 'B<$STRICTER>'";
1179my $A_bold_stricter = "A 'B<$STRICTER>'";
1180my $STABILIZED = 'S';
1181my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1182my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1183my $OBSOLETE = 'O';
1184my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1185my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1186
1187my %status_past_participles = (
1188 $DISCOURAGED => 'discouraged',
1189 $SUPPRESSED => 'should never be generated',
1190 $STABILIZED => 'stabilized',
1191 $OBSOLETE => 'obsolete',
37e2e78e 1192 $DEPRECATED => 'deprecated',
99870f4d
KW
1193);
1194
f5817e0a
KW
1195# The format of the values of the tables:
1196my $EMPTY_FORMAT = "";
99870f4d
KW
1197my $BINARY_FORMAT = 'b';
1198my $DECIMAL_FORMAT = 'd';
1199my $FLOAT_FORMAT = 'f';
1200my $INTEGER_FORMAT = 'i';
1201my $HEX_FORMAT = 'x';
1202my $RATIONAL_FORMAT = 'r';
1203my $STRING_FORMAT = 's';
a14f3cb1 1204my $DECOMP_STRING_FORMAT = 'c';
99870f4d
KW
1205
1206my %map_table_formats = (
1207 $BINARY_FORMAT => 'binary',
1208 $DECIMAL_FORMAT => 'single decimal digit',
1209 $FLOAT_FORMAT => 'floating point number',
1210 $INTEGER_FORMAT => 'integer',
1211 $HEX_FORMAT => 'positive hex whole number; a code point',
1212 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1213 $STRING_FORMAT => 'string',
92f9d56c 1214 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
99870f4d
KW
1215);
1216
1217# Unicode didn't put such derived files in a separate directory at first.
1218my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1219my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1220my $AUXILIARY = 'auxiliary';
1221
1222# Hashes that will eventually go into Heavy.pl for the use of utf8_heavy.pl
1223my %loose_to_file_of; # loosely maps table names to their respective
1224 # files
1225my %stricter_to_file_of; # same; but for stricter mapping.
1226my %nv_floating_to_rational; # maps numeric values floating point numbers to
1227 # their rational equivalent
1228my %loose_property_name_of; # Loosely maps property names to standard form
1229
d867ccfb
KW
1230# Most properties are immune to caseless matching, otherwise you would get
1231# nonsensical results, as properties are a function of a code point, not
1232# everything that is caselessly equivalent to that code point. For example,
1233# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1234# be true because 's' and 'S' are equivalent caselessly. However,
1235# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1236# extend that concept to those very few properties that are like this. Each
1237# such property will match the full range caselessly. They are hard-coded in
1238# the program; it's not worth trying to make it general as it's extremely
1239# unlikely that they will ever change.
1240my %caseless_equivalent_to;
1241
99870f4d
KW
1242# These constants names and values were taken from the Unicode standard,
1243# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1244# syllables. The '_string' versions are so generated tables can retain the
1245# hex format, which is the more familiar value
1246my $SBase_string = "0xAC00";
1247my $SBase = CORE::hex $SBase_string;
1248my $LBase_string = "0x1100";
1249my $LBase = CORE::hex $LBase_string;
1250my $VBase_string = "0x1161";
1251my $VBase = CORE::hex $VBase_string;
1252my $TBase_string = "0x11A7";
1253my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1254my $SCount = 11172;
1255my $LCount = 19;
1256my $VCount = 21;
1257my $TCount = 28;
1258my $NCount = $VCount * $TCount;
1259
1260# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1261# with the above published constants.
1262my %Jamo;
1263my %Jamo_L; # Leading consonants
1264my %Jamo_V; # Vowels
1265my %Jamo_T; # Trailing consonants
1266
37e2e78e 1267my @backslash_X_tests; # List of tests read in for testing \X
99870f4d
KW
1268my @unhandled_properties; # Will contain a list of properties found in
1269 # the input that we didn't process.
f86864ac 1270my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1271 # listed in the pod
1272my @map_properties; # Properties that get map files written
1273my @named_sequences; # NamedSequences.txt contents.
1274my %potential_files; # Generated list of all .txt files in the directory
1275 # structure so we can warn if something is being
1276 # ignored.
1277my @files_actually_output; # List of files we generated.
1278my @more_Names; # Some code point names are compound; this is used
1279 # to store the extra components of them.
1280my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1281 # the minimum before we consider it equivalent to a
1282 # candidate rational
1283my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1284
1285# These store references to certain commonly used property objects
1286my $gc;
1287my $perl;
1288my $block;
3e20195b
KW
1289my $perl_charname;
1290my $print;
7fc6cb55 1291my $Any;
99870f4d
KW
1292
1293# Are there conflicting names because of beginning with 'In_', or 'Is_'
1294my $has_In_conflicts = 0;
1295my $has_Is_conflicts = 0;
1296
1297sub internal_file_to_platform ($) {
1298 # Convert our file paths which have '/' separators to those of the
1299 # platform.
1300
1301 my $file = shift;
1302 return undef unless defined $file;
1303
1304 return File::Spec->join(split '/', $file);
d07a55ed 1305}
5beb625e 1306
99870f4d
KW
1307sub file_exists ($) { # platform independent '-e'. This program internally
1308 # uses slash as a path separator.
1309 my $file = shift;
1310 return 0 if ! defined $file;
1311 return -e internal_file_to_platform($file);
1312}
5beb625e 1313
99870f4d 1314sub objaddr($) {
23e33b60
KW
1315 # Returns the address of the blessed input object.
1316 # It doesn't check for blessedness because that would do a string eval
1317 # every call, and the program is structured so that this is never called
1318 # for a non-blessed object.
99870f4d 1319
23e33b60 1320 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1321
1322 # Numifying a ref gives its address.
051df77b 1323 return pack 'J', $_[0];
99870f4d
KW
1324}
1325
558712cf 1326# These are used only if $annotate is true.
c4019d52
KW
1327# The entire range of Unicode characters is examined to populate these
1328# after all the input has been processed. But most can be skipped, as they
1329# have the same descriptive phrases, such as being unassigned
1330my @viacode; # Contains the 1 million character names
1331my @printable; # boolean: And are those characters printable?
1332my @annotate_char_type; # Contains a type of those characters, specifically
1333 # for the purposes of annotation.
1334my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1335 # name for the purposes of annotation. They map to the
c4019d52
KW
1336 # upper edge of the range, so that the end point can
1337 # be immediately found. This is used to skip ahead to
1338 # the end of a range, and avoid processing each
1339 # individual code point in it.
1340my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1341 # characters, but excluding those which are
1342 # also noncharacter code points
1343
1344# The annotation types are an extension of the regular range types, though
1345# some of the latter are folded into one. Make the new types negative to
1346# avoid conflicting with the regular types
1347my $SURROGATE_TYPE = -1;
1348my $UNASSIGNED_TYPE = -2;
1349my $PRIVATE_USE_TYPE = -3;
1350my $NONCHARACTER_TYPE = -4;
1351my $CONTROL_TYPE = -5;
1352my $UNKNOWN_TYPE = -6; # Used only if there is a bug in this program
1353
1354sub populate_char_info ($) {
558712cf 1355 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1356 # input code point's info that are needed for outputting more detailed
1357 # comments. If calling context wants a return, it is the end point of
1358 # any contiguous range of characters that share essentially the same info
1359
1360 my $i = shift;
1361 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1362
1363 $viacode[$i] = $perl_charname->value_of($i) || "";
1364
1365 # A character is generally printable if Unicode says it is,
1366 # but below we make sure that most Unicode general category 'C' types
1367 # aren't.
1368 $printable[$i] = $print->contains($i);
1369
1370 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1371
1372 # Only these two regular types are treated specially for annotations
1373 # purposes
1374 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1375 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1376
1377 # Give a generic name to all code points that don't have a real name.
1378 # We output ranges, if applicable, for these. Also calculate the end
1379 # point of the range.
1380 my $end;
1381 if (! $viacode[$i]) {
1382 if ($gc-> table('Surrogate')->contains($i)) {
1383 $viacode[$i] = 'Surrogate';
1384 $annotate_char_type[$i] = $SURROGATE_TYPE;
1385 $printable[$i] = 0;
1386 $end = $gc->table('Surrogate')->containing_range($i)->end;
1387 }
1388 elsif ($gc-> table('Private_use')->contains($i)) {
1389 $viacode[$i] = 'Private Use';
1390 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1391 $printable[$i] = 0;
1392 $end = $gc->table('Private_Use')->containing_range($i)->end;
1393 }
1394 elsif (Property::property_ref('Noncharacter_Code_Point')-> table('Y')->
1395 contains($i))
1396 {
1397 $viacode[$i] = 'Noncharacter';
1398 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1399 $printable[$i] = 0;
1400 $end = property_ref('Noncharacter_Code_Point')->table('Y')->
1401 containing_range($i)->end;
1402 }
1403 elsif ($gc-> table('Control')->contains($i)) {
1404 $viacode[$i] = 'Control';
1405 $annotate_char_type[$i] = $CONTROL_TYPE;
1406 $printable[$i] = 0;
1407 $end = 0x81 if $i == 0x80; # Hard-code this one known case
1408 }
1409 elsif ($gc-> table('Unassigned')->contains($i)) {
1410 $viacode[$i] = 'Unassigned, block=' . $block-> value_of($i);
1411 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1412 $printable[$i] = 0;
1413
1414 # Because we name the unassigned by the blocks they are in, it
1415 # can't go past the end of that block, and it also can't go past
1416 # the unassigned range it is in. The special table makes sure
1417 # that the non-characters, which are unassigned, are separated
1418 # out.
1419 $end = min($block->containing_range($i)->end,
1420 $unassigned_sans_noncharacters-> containing_range($i)->
1421 end);
13ca76ff
KW
1422 }
1423 else {
1424 Carp::my_carp_bug("Can't figure out how to annotate "
1425 . sprintf("U+%04X", $i)
1426 . ". Proceeding anyway.");
c4019d52
KW
1427 $viacode[$i] = 'UNKNOWN';
1428 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1429 $printable[$i] = 0;
1430 }
1431 }
1432
1433 # Here, has a name, but if it's one in which the code point number is
1434 # appended to the name, do that.
1435 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1436 $viacode[$i] .= sprintf("-%04X", $i);
1437 $end = $perl_charname->containing_range($i)->end;
1438 }
1439
1440 # And here, has a name, but if it's a hangul syllable one, replace it with
1441 # the correct name from the Unicode algorithm
1442 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1443 use integer;
1444 my $SIndex = $i - $SBase;
1445 my $L = $LBase + $SIndex / $NCount;
1446 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1447 my $T = $TBase + $SIndex % $TCount;
1448 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1449 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1450 $end = $perl_charname->containing_range($i)->end;
1451 }
1452
1453 return if ! defined wantarray;
1454 return $i if ! defined $end; # If not a range, return the input
1455
1456 # Save this whole range so can find the end point quickly
1457 $annotate_ranges->add_map($i, $end, $end);
1458
1459 return $end;
1460}
1461
23e33b60
KW
1462# Commented code below should work on Perl 5.8.
1463## This 'require' doesn't necessarily work in miniperl, and even if it does,
1464## the native perl version of it (which is what would operate under miniperl)
1465## is extremely slow, as it does a string eval every call.
1466#my $has_fast_scalar_util = $\18 !~ /miniperl/
1467# && defined eval "require Scalar::Util";
1468#
1469#sub objaddr($) {
1470# # Returns the address of the blessed input object. Uses the XS version if
1471# # available. It doesn't check for blessedness because that would do a
1472# # string eval every call, and the program is structured so that this is
1473# # never called for a non-blessed object.
1474#
1475# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1476#
1477# # Check at least that is a ref.
1478# my $pkg = ref($_[0]) or return undef;
1479#
1480# # Change to a fake package to defeat any overloaded stringify
1481# bless $_[0], 'main::Fake';
1482#
1483# # Numifying a ref gives its address.
051df77b 1484# my $addr = pack 'J', $_[0];
23e33b60
KW
1485#
1486# # Return to original class
1487# bless $_[0], $pkg;
1488# return $addr;
1489#}
1490
99870f4d
KW
1491sub max ($$) {
1492 my $a = shift;
1493 my $b = shift;
1494 return $a if $a >= $b;
1495 return $b;
1496}
1497
1498sub min ($$) {
1499 my $a = shift;
1500 my $b = shift;
1501 return $a if $a <= $b;
1502 return $b;
1503}
1504
1505sub clarify_number ($) {
1506 # This returns the input number with underscores inserted every 3 digits
1507 # in large (5 digits or more) numbers. Input must be entirely digits, not
1508 # checked.
1509
1510 my $number = shift;
1511 my $pos = length($number) - 3;
1512 return $number if $pos <= 1;
1513 while ($pos > 0) {
1514 substr($number, $pos, 0) = '_';
1515 $pos -= 3;
5beb625e 1516 }
99870f4d 1517 return $number;
99598c8c
JH
1518}
1519
12ac2576 1520
99870f4d 1521package Carp;
7ebf06b3 1522
99870f4d
KW
1523# These routines give a uniform treatment of messages in this program. They
1524# are placed in the Carp package to cause the stack trace to not include them,
1525# although an alternative would be to use another package and set @CARP_NOT
1526# for it.
12ac2576 1527
99870f4d 1528our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1529
99f78760
KW
1530# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1531# and overload trying to load Scalar:Util under miniperl. See
1532# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1533undef $overload::VERSION;
1534
99870f4d
KW
1535sub my_carp {
1536 my $message = shift || "";
1537 my $nofold = shift || 0;
7ebf06b3 1538
99870f4d
KW
1539 if ($message) {
1540 $message = main::join_lines($message);
1541 $message =~ s/^$0: *//; # Remove initial program name
1542 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1543 $message = "\n$0: $message;";
12ac2576 1544
99870f4d
KW
1545 # Fold the message with program name, semi-colon end punctuation
1546 # (which looks good with the message that carp appends to it), and a
1547 # hanging indent for continuation lines.
1548 $message = main::simple_fold($message, "", 4) unless $nofold;
1549 $message =~ s/\n$//; # Remove the trailing nl so what carp
1550 # appends is to the same line
1551 }
12ac2576 1552
99870f4d 1553 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1554
99870f4d
KW
1555 carp $message;
1556 return;
1557}
7ebf06b3 1558
99870f4d
KW
1559sub my_carp_bug {
1560 # This is called when it is clear that the problem is caused by a bug in
1561 # this program.
7ebf06b3 1562
99870f4d
KW
1563 my $message = shift;
1564 $message =~ s/^$0: *//;
1565 $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");
1566 carp $message;
1567 return;
1568}
7ebf06b3 1569
99870f4d
KW
1570sub carp_too_few_args {
1571 if (@_ != 2) {
1572 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1573 return;
12ac2576 1574 }
7ebf06b3 1575
99870f4d
KW
1576 my $args_ref = shift;
1577 my $count = shift;
7ebf06b3 1578
99870f4d
KW
1579 my_carp_bug("Need at least $count arguments to "
1580 . (caller 1)[3]
1581 . ". Instead got: '"
1582 . join ', ', @$args_ref
1583 . "'. No action taken.");
1584 return;
12ac2576
JP
1585}
1586
99870f4d
KW
1587sub carp_extra_args {
1588 my $args_ref = shift;
1589 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1590
99870f4d
KW
1591 unless (ref $args_ref) {
1592 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1593 return;
1594 }
1595 my ($package, $file, $line) = caller;
1596 my $subroutine = (caller 1)[3];
cf25bb62 1597
99870f4d
KW
1598 my $list;
1599 if (ref $args_ref eq 'HASH') {
1600 foreach my $key (keys %$args_ref) {
1601 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1602 }
99870f4d 1603 $list = join ', ', each %{$args_ref};
cf25bb62 1604 }
99870f4d
KW
1605 elsif (ref $args_ref eq 'ARRAY') {
1606 foreach my $arg (@$args_ref) {
1607 $arg = $UNDEF unless defined $arg;
1608 }
1609 $list = join ', ', @$args_ref;
1610 }
1611 else {
1612 my_carp_bug("Can't cope with ref "
1613 . ref($args_ref)
1614 . " . argument to 'carp_extra_args'. Not checking arguments.");
1615 return;
1616 }
1617
1618 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1619 return;
d73e5302
JH
1620}
1621
99870f4d
KW
1622package main;
1623
1624{ # Closure
1625
1626 # This program uses the inside-out method for objects, as recommended in
1627 # "Perl Best Practices". This closure aids in generating those. There
1628 # are two routines. setup_package() is called once per package to set
1629 # things up, and then set_access() is called for each hash representing a
1630 # field in the object. These routines arrange for the object to be
1631 # properly destroyed when no longer used, and for standard accessor
1632 # functions to be generated. If you need more complex accessors, just
1633 # write your own and leave those accesses out of the call to set_access().
1634 # More details below.
1635
1636 my %constructor_fields; # fields that are to be used in constructors; see
1637 # below
1638
1639 # The values of this hash will be the package names as keys to other
1640 # hashes containing the name of each field in the package as keys, and
1641 # references to their respective hashes as values.
1642 my %package_fields;
1643
1644 sub setup_package {
1645 # Sets up the package, creating standard DESTROY and dump methods
1646 # (unless already defined). The dump method is used in debugging by
1647 # simple_dumper().
1648 # The optional parameters are:
1649 # a) a reference to a hash, that gets populated by later
1650 # set_access() calls with one of the accesses being
1651 # 'constructor'. The caller can then refer to this, but it is
1652 # not otherwise used by these two routines.
1653 # b) a reference to a callback routine to call during destruction
1654 # of the object, before any fields are actually destroyed
1655
1656 my %args = @_;
1657 my $constructor_ref = delete $args{'Constructor_Fields'};
1658 my $destroy_callback = delete $args{'Destroy_Callback'};
1659 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1660
1661 my %fields;
1662 my $package = (caller)[0];
1663
1664 $package_fields{$package} = \%fields;
1665 $constructor_fields{$package} = $constructor_ref;
1666
1667 unless ($package->can('DESTROY')) {
1668 my $destroy_name = "${package}::DESTROY";
1669 no strict "refs";
1670
1671 # Use typeglob to give the anonymous subroutine the name we want
1672 *$destroy_name = sub {
1673 my $self = shift;
ffe43484 1674 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1675
1676 $self->$destroy_callback if $destroy_callback;
1677 foreach my $field (keys %{$package_fields{$package}}) {
1678 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1679 delete $package_fields{$package}{$field}{$addr};
1680 }
1681 return;
1682 }
1683 }
1684
1685 unless ($package->can('dump')) {
1686 my $dump_name = "${package}::dump";
1687 no strict "refs";
1688 *$dump_name = sub {
1689 my $self = shift;
1690 return dump_inside_out($self, $package_fields{$package}, @_);
1691 }
1692 }
1693 return;
1694 }
1695
1696 sub set_access {
1697 # Arrange for the input field to be garbage collected when no longer
1698 # needed. Also, creates standard accessor functions for the field
1699 # based on the optional parameters-- none if none of these parameters:
1700 # 'addable' creates an 'add_NAME()' accessor function.
1701 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1702 # function.
1703 # 'settable' creates a 'set_NAME()' accessor function.
1704 # 'constructor' doesn't create an accessor function, but adds the
1705 # field to the hash that was previously passed to
1706 # setup_package();
1707 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1708 # 'add' etc. all mean 'addable'.
1709 # The read accessor function will work on both array and scalar
1710 # values. If another accessor in the parameter list is 'a', the read
1711 # access assumes an array. You can also force it to be array access
1712 # by specifying 'readable_array' instead of 'readable'
1713 #
1714 # A sort-of 'protected' access can be set-up by preceding the addable,
1715 # readable or settable with some initial portion of 'protected_' (but,
1716 # the underscore is required), like 'p_a', 'pro_set', etc. The
1717 # "protection" is only by convention. All that happens is that the
1718 # accessor functions' names begin with an underscore. So instead of
1719 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1720 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1721 # end of each package, and then storing the __LINE__ ranges and
1722 # checking them on every accessor. But that is way overkill.)
1723
1724 # We create anonymous subroutines as the accessors and then use
1725 # typeglobs to assign them to the proper package and name
1726
1727 my $name = shift; # Name of the field
1728 my $field = shift; # Reference to the inside-out hash containing the
1729 # field
1730
1731 my $package = (caller)[0];
1732
1733 if (! exists $package_fields{$package}) {
1734 croak "$0: Must call 'setup_package' before 'set_access'";
1735 }
d73e5302 1736
99870f4d
KW
1737 # Stash the field so DESTROY can get it.
1738 $package_fields{$package}{$name} = $field;
cf25bb62 1739
99870f4d
KW
1740 # Remaining arguments are the accessors. For each...
1741 foreach my $access (@_) {
1742 my $access = lc $access;
cf25bb62 1743
99870f4d 1744 my $protected = "";
cf25bb62 1745
99870f4d
KW
1746 # Match the input as far as it goes.
1747 if ($access =~ /^(p[^_]*)_/) {
1748 $protected = $1;
1749 if (substr('protected_', 0, length $protected)
1750 eq $protected)
1751 {
1752
1753 # Add 1 for the underscore not included in $protected
1754 $access = substr($access, length($protected) + 1);
1755 $protected = '_';
1756 }
1757 else {
1758 $protected = "";
1759 }
1760 }
1761
1762 if (substr('addable', 0, length $access) eq $access) {
1763 my $subname = "${package}::${protected}add_$name";
1764 no strict "refs";
1765
1766 # add_ accessor. Don't add if already there, which we
1767 # determine using 'eq' for scalars and '==' otherwise.
1768 *$subname = sub {
1769 use strict "refs";
1770 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1771 my $self = shift;
1772 my $value = shift;
ffe43484 1773 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1774 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1775 if (ref $value) {
f998e60c 1776 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1777 }
1778 else {
f998e60c 1779 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1780 }
f998e60c 1781 push @{$field->{$addr}}, $value;
99870f4d
KW
1782 return;
1783 }
1784 }
1785 elsif (substr('constructor', 0, length $access) eq $access) {
1786 if ($protected) {
1787 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1788 }
1789 else {
1790 $constructor_fields{$package}{$name} = $field;
1791 }
1792 }
1793 elsif (substr('readable_array', 0, length $access) eq $access) {
1794
1795 # Here has read access. If one of the other parameters for
1796 # access is array, or this one specifies array (by being more
1797 # than just 'readable_'), then create a subroutine that
1798 # assumes the data is an array. Otherwise just a scalar
1799 my $subname = "${package}::${protected}$name";
1800 if (grep { /^a/i } @_
1801 or length($access) > length('readable_'))
1802 {
1803 no strict "refs";
1804 *$subname = sub {
1805 use strict "refs";
23e33b60 1806 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 1807 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
1808 if (ref $field->{$addr} ne 'ARRAY') {
1809 my $type = ref $field->{$addr};
1810 $type = 'scalar' unless $type;
1811 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
1812 return;
1813 }
1814 return scalar @{$field->{$addr}} unless wantarray;
1815
1816 # Make a copy; had problems with caller modifying the
1817 # original otherwise
1818 my @return = @{$field->{$addr}};
1819 return @return;
1820 }
1821 }
1822 else {
1823
1824 # Here not an array value, a simpler function.
1825 no strict "refs";
1826 *$subname = sub {
1827 use strict "refs";
23e33b60 1828 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 1829 no overloading;
051df77b 1830 return $field->{pack 'J', $_[0]};
99870f4d
KW
1831 }
1832 }
1833 }
1834 elsif (substr('settable', 0, length $access) eq $access) {
1835 my $subname = "${package}::${protected}set_$name";
1836 no strict "refs";
1837 *$subname = sub {
1838 use strict "refs";
23e33b60
KW
1839 if (main::DEBUG) {
1840 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
1841 Carp::carp_extra_args(\@_) if @_ > 2;
1842 }
1843 # $self is $_[0]; $value is $_[1]
f998e60c 1844 no overloading;
051df77b 1845 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
1846 return;
1847 }
1848 }
1849 else {
1850 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
1851 }
cf25bb62 1852 }
99870f4d 1853 return;
cf25bb62 1854 }
99870f4d
KW
1855}
1856
1857package Input_file;
1858
1859# All input files use this object, which stores various attributes about them,
1860# and provides for convenient, uniform handling. The run method wraps the
1861# processing. It handles all the bookkeeping of opening, reading, and closing
1862# the file, returning only significant input lines.
1863#
1864# Each object gets a handler which processes the body of the file, and is
1865# called by run(). Most should use the generic, default handler, which has
1866# code scrubbed to handle things you might not expect. A handler should
1867# basically be a while(next_line()) {...} loop.
1868#
1869# You can also set up handlers to
1870# 1) call before the first line is read for pre processing
1871# 2) call to adjust each line of the input before the main handler gets them
1872# 3) call upon EOF before the main handler exits its loop
1873# 4) call at the end for post processing
1874#
1875# $_ is used to store the input line, and is to be filtered by the
1876# each_line_handler()s. So, if the format of the line is not in the desired
1877# format for the main handler, these are used to do that adjusting. They can
1878# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
1879# so the $_ output of one is used as the input to the next. None of the other
1880# handlers are stackable, but could easily be changed to be so.
1881#
1882# Most of the handlers can call insert_lines() or insert_adjusted_lines()
1883# which insert the parameters as lines to be processed before the next input
1884# file line is read. This allows the EOF handler to flush buffers, for
1885# example. The difference between the two routines is that the lines inserted
1886# by insert_lines() are subjected to the each_line_handler()s. (So if you
1887# called it from such a handler, you would get infinite recursion.) Lines
1888# inserted by insert_adjusted_lines() go directly to the main handler without
1889# any adjustments. If the post-processing handler calls any of these, there
1890# will be no effect. Some error checking for these conditions could be added,
1891# but it hasn't been done.
1892#
1893# carp_bad_line() should be called to warn of bad input lines, which clears $_
1894# to prevent further processing of the line. This routine will output the
1895# message as a warning once, and then keep a count of the lines that have the
1896# same message, and output that count at the end of the file's processing.
1897# This keeps the number of messages down to a manageable amount.
1898#
1899# get_missings() should be called to retrieve any @missing input lines.
1900# Messages will be raised if this isn't done if the options aren't to ignore
1901# missings.
1902
1903sub trace { return main::trace(@_); }
1904
99870f4d
KW
1905{ # Closure
1906 # Keep track of fields that are to be put into the constructor.
1907 my %constructor_fields;
1908
1909 main::setup_package(Constructor_Fields => \%constructor_fields);
1910
1911 my %file; # Input file name, required
1912 main::set_access('file', \%file, qw{ c r });
1913
1914 my %first_released; # Unicode version file was first released in, required
1915 main::set_access('first_released', \%first_released, qw{ c r });
1916
1917 my %handler; # Subroutine to process the input file, defaults to
1918 # 'process_generic_property_file'
1919 main::set_access('handler', \%handler, qw{ c });
1920
1921 my %property;
1922 # name of property this file is for. defaults to none, meaning not
1923 # applicable, or is otherwise determinable, for example, from each line.
1924 main::set_access('property', \%property, qw{ c });
1925
1926 my %optional;
1927 # If this is true, the file is optional. If not present, no warning is
1928 # output. If it is present, the string given by this parameter is
1929 # evaluated, and if false the file is not processed.
1930 main::set_access('optional', \%optional, 'c', 'r');
1931
1932 my %non_skip;
1933 # This is used for debugging, to skip processing of all but a few input
1934 # files. Add 'non_skip => 1' to the constructor for those files you want
1935 # processed when you set the $debug_skip global.
1936 main::set_access('non_skip', \%non_skip, 'c');
1937
37e2e78e
KW
1938 my %skip;
1939 # This is used to skip processing of this input file semi-permanently.
1940 # It is used for files that we aren't planning to process anytime soon,
1941 # but want to allow to be in the directory and not raise a message that we
1942 # are not handling. Mostly for test files. This is in contrast to the
1943 # non_skip element, which is supposed to be used very temporarily for
1944 # debugging. Sets 'optional' to 1
1945 main::set_access('skip', \%skip, 'c');
1946
99870f4d
KW
1947 my %each_line_handler;
1948 # list of subroutines to look at and filter each non-comment line in the
1949 # file. defaults to none. The subroutines are called in order, each is
1950 # to adjust $_ for the next one, and the final one adjusts it for
1951 # 'handler'
1952 main::set_access('each_line_handler', \%each_line_handler, 'c');
1953
1954 my %has_missings_defaults;
1955 # ? Are there lines in the file giving default values for code points
1956 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
1957 # the norm, but IGNORED means it has such lines, but the handler doesn't
1958 # use them. Having these three states allows us to catch changes to the
1959 # UCD that this program should track
1960 main::set_access('has_missings_defaults',
1961 \%has_missings_defaults, qw{ c r });
1962
1963 my %pre_handler;
1964 # Subroutine to call before doing anything else in the file. If undef, no
1965 # such handler is called.
1966 main::set_access('pre_handler', \%pre_handler, qw{ c });
1967
1968 my %eof_handler;
1969 # Subroutine to call upon getting an EOF on the input file, but before
1970 # that is returned to the main handler. This is to allow buffers to be
1971 # flushed. The handler is expected to call insert_lines() or
1972 # insert_adjusted() with the buffered material
1973 main::set_access('eof_handler', \%eof_handler, qw{ c r });
1974
1975 my %post_handler;
1976 # Subroutine to call after all the lines of the file are read in and
1977 # processed. If undef, no such handler is called.
1978 main::set_access('post_handler', \%post_handler, qw{ c });
1979
1980 my %progress_message;
1981 # Message to print to display progress in lieu of the standard one
1982 main::set_access('progress_message', \%progress_message, qw{ c });
1983
1984 my %handle;
1985 # cache open file handle, internal. Is undef if file hasn't been
1986 # processed at all, empty if has;
1987 main::set_access('handle', \%handle);
1988
1989 my %added_lines;
1990 # cache of lines added virtually to the file, internal
1991 main::set_access('added_lines', \%added_lines);
1992
1993 my %errors;
1994 # cache of errors found, internal
1995 main::set_access('errors', \%errors);
1996
1997 my %missings;
1998 # storage of '@missing' defaults lines
1999 main::set_access('missings', \%missings);
2000
2001 sub new {
2002 my $class = shift;
2003
2004 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2005 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2006
2007 # Set defaults
2008 $handler{$addr} = \&main::process_generic_property_file;
2009 $non_skip{$addr} = 0;
37e2e78e 2010 $skip{$addr} = 0;
99870f4d
KW
2011 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2012 $handle{$addr} = undef;
2013 $added_lines{$addr} = [ ];
2014 $each_line_handler{$addr} = [ ];
2015 $errors{$addr} = { };
2016 $missings{$addr} = [ ];
2017
2018 # Two positional parameters.
99f78760 2019 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2020 $file{$addr} = main::internal_file_to_platform(shift);
2021 $first_released{$addr} = shift;
2022
2023 # The rest of the arguments are key => value pairs
2024 # %constructor_fields has been set up earlier to list all possible
2025 # ones. Either set or push, depending on how the default has been set
2026 # up just above.
2027 my %args = @_;
2028 foreach my $key (keys %args) {
2029 my $argument = $args{$key};
2030
2031 # Note that the fields are the lower case of the constructor keys
2032 my $hash = $constructor_fields{lc $key};
2033 if (! defined $hash) {
2034 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2035 next;
2036 }
2037 if (ref $hash->{$addr} eq 'ARRAY') {
2038 if (ref $argument eq 'ARRAY') {
2039 foreach my $argument (@{$argument}) {
2040 next if ! defined $argument;
2041 push @{$hash->{$addr}}, $argument;
2042 }
2043 }
2044 else {
2045 push @{$hash->{$addr}}, $argument if defined $argument;
2046 }
2047 }
2048 else {
2049 $hash->{$addr} = $argument;
2050 }
2051 delete $args{$key};
2052 };
2053
2054 # If the file has a property for it, it means that the property is not
2055 # listed in the file's entries. So add a handler to the list of line
2056 # handlers to insert the property name into the lines, to provide a
2057 # uniform interface to the final processing subroutine.
2058 # the final code doesn't have to worry about that.
2059 if ($property{$addr}) {
2060 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
2061 }
2062
2063 if ($non_skip{$addr} && ! $debug_skip && $verbosity) {
2064 print "Warning: " . __PACKAGE__ . " constructor for $file{$addr} has useless 'non_skip' in it\n";
a3a8c5f0 2065 }
99870f4d 2066
37e2e78e
KW
2067 $optional{$addr} = 1 if $skip{$addr};
2068
99870f4d 2069 return $self;
d73e5302
JH
2070 }
2071
cf25bb62 2072
99870f4d
KW
2073 use overload
2074 fallback => 0,
2075 qw("") => "_operator_stringify",
2076 "." => \&main::_operator_dot,
2077 ;
cf25bb62 2078
99870f4d
KW
2079 sub _operator_stringify {
2080 my $self = shift;
cf25bb62 2081
99870f4d 2082 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2083 }
d73e5302 2084
99870f4d
KW
2085 # flag to make sure extracted files are processed early
2086 my $seen_non_extracted_non_age = 0;
d73e5302 2087
99870f4d
KW
2088 sub run {
2089 # Process the input object $self. This opens and closes the file and
2090 # calls all the handlers for it. Currently, this can only be called
2091 # once per file, as it destroy's the EOF handler
d73e5302 2092
99870f4d
KW
2093 my $self = shift;
2094 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2095
ffe43484 2096 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2097
99870f4d 2098 my $file = $file{$addr};
d73e5302 2099
99870f4d
KW
2100 # Don't process if not expecting this file (because released later
2101 # than this Unicode version), and isn't there. This means if someone
2102 # copies it into an earlier version's directory, we will go ahead and
2103 # process it.
2104 return if $first_released{$addr} gt $v_version && ! -e $file;
2105
2106 # If in debugging mode and this file doesn't have the non-skip
2107 # flag set, and isn't one of the critical files, skip it.
2108 if ($debug_skip
2109 && $first_released{$addr} ne v0
2110 && ! $non_skip{$addr})
2111 {
2112 print "Skipping $file in debugging\n" if $verbosity;
2113 return;
2114 }
2115
2116 # File could be optional
37e2e78e 2117 if ($optional{$addr}) {
99870f4d
KW
2118 return unless -e $file;
2119 my $result = eval $optional{$addr};
2120 if (! defined $result) {
2121 Carp::my_carp_bug("Got '$@' when tried to eval $optional{$addr}. $file Skipped.");
2122 return;
2123 }
2124 if (! $result) {
2125 if ($verbosity) {
2126 print STDERR "Skipping processing input file '$file' because '$optional{$addr}' is not true\n";
2127 }
2128 return;
2129 }
2130 }
2131
2132 if (! defined $file || ! -e $file) {
2133
2134 # If the file doesn't exist, see if have internal data for it
2135 # (based on first_released being 0).
2136 if ($first_released{$addr} eq v0) {
2137 $handle{$addr} = 'pretend_is_open';
2138 }
2139 else {
2140 if (! $optional{$addr} # File could be optional
2141 && $v_version ge $first_released{$addr})
2142 {
2143 print STDERR "Skipping processing input file '$file' because not found\n" if $v_version ge $first_released{$addr};
2144 }
2145 return;
2146 }
2147 }
2148 else {
2149
37e2e78e
KW
2150 # Here, the file exists. Some platforms may change the case of
2151 # its name
99870f4d 2152 if ($seen_non_extracted_non_age) {
517956bf 2153 if ($file =~ /$EXTRACTED/i) {
99870f4d 2154 Carp::my_carp_bug(join_lines(<<END
99f78760 2155$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2156anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2157have subtle problems
2158END
2159 ));
2160 }
2161 }
2162 elsif ($EXTRACTED_DIR
2163 && $first_released{$addr} ne v0
517956bf
CB
2164 && $file !~ /$EXTRACTED/i
2165 && lc($file) ne 'dage.txt')
99870f4d
KW
2166 {
2167 # We don't set this (by the 'if' above) if we have no
2168 # extracted directory, so if running on an early version,
2169 # this test won't work. Not worth worrying about.
2170 $seen_non_extracted_non_age = 1;
2171 }
2172
2173 # And mark the file as having being processed, and warn if it
2174 # isn't a file we are expecting. As we process the files,
2175 # they are deleted from the hash, so any that remain at the
2176 # end of the program are files that we didn't process.
517956bf
CB
2177 my $fkey = File::Spec->rel2abs($file);
2178 my $expecting = delete $potential_files{$fkey};
2179 $expecting = delete $potential_files{lc($fkey)} unless defined $expecting;
678f13d5
KW
2180 Carp::my_carp("Was not expecting '$file'.") if
2181 ! $expecting
99870f4d
KW
2182 && ! defined $handle{$addr};
2183
37e2e78e
KW
2184 # Having deleted from expected files, we can quit if not to do
2185 # anything. Don't print progress unless really want verbosity
2186 if ($skip{$addr}) {
2187 print "Skipping $file.\n" if $verbosity >= $VERBOSE;
2188 return;
2189 }
2190
99870f4d
KW
2191 # Open the file, converting the slashes used in this program
2192 # into the proper form for the OS
2193 my $file_handle;
2194 if (not open $file_handle, "<", $file) {
2195 Carp::my_carp("Can't open $file. Skipping: $!");
2196 return 0;
2197 }
2198 $handle{$addr} = $file_handle; # Cache the open file handle
2199 }
2200
2201 if ($verbosity >= $PROGRESS) {
2202 if ($progress_message{$addr}) {
2203 print "$progress_message{$addr}\n";
2204 }
2205 else {
2206 # If using a virtual file, say so.
2207 print "Processing ", (-e $file)
2208 ? $file
2209 : "substitute $file",
2210 "\n";
2211 }
2212 }
2213
2214
2215 # Call any special handler for before the file.
2216 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2217
2218 # Then the main handler
2219 &{$handler{$addr}}($self);
2220
2221 # Then any special post-file handler.
2222 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2223
2224 # If any errors have been accumulated, output the counts (as the first
2225 # error message in each class was output when it was encountered).
2226 if ($errors{$addr}) {
2227 my $total = 0;
2228 my $types = 0;
2229 foreach my $error (keys %{$errors{$addr}}) {
2230 $total += $errors{$addr}->{$error};
2231 delete $errors{$addr}->{$error};
2232 $types++;
2233 }
2234 if ($total > 1) {
2235 my $message
2236 = "A total of $total lines had errors in $file. ";
2237
2238 $message .= ($types == 1)
2239 ? '(Only the first one was displayed.)'
2240 : '(Only the first of each type was displayed.)';
2241 Carp::my_carp($message);
2242 }
2243 }
2244
2245 if (@{$missings{$addr}}) {
2246 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2247 }
2248
2249 # If a real file handle, close it.
2250 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2251 ref $handle{$addr};
2252 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2253 # the file, as opposed to undef
2254 return;
2255 }
2256
2257 sub next_line {
2258 # Sets $_ to be the next logical input line, if any. Returns non-zero
2259 # if such a line exists. 'logical' means that any lines that have
2260 # been added via insert_lines() will be returned in $_ before the file
2261 # is read again.
2262
2263 my $self = shift;
2264 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2265
ffe43484 2266 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2267
2268 # Here the file is open (or if the handle is not a ref, is an open
2269 # 'virtual' file). Get the next line; any inserted lines get priority
2270 # over the file itself.
2271 my $adjusted;
2272
2273 LINE:
2274 while (1) { # Loop until find non-comment, non-empty line
2275 #local $to_trace = 1 if main::DEBUG;
2276 my $inserted_ref = shift @{$added_lines{$addr}};
2277 if (defined $inserted_ref) {
2278 ($adjusted, $_) = @{$inserted_ref};
2279 trace $adjusted, $_ if main::DEBUG && $to_trace;
2280 return 1 if $adjusted;
2281 }
2282 else {
2283 last if ! ref $handle{$addr}; # Don't read unless is real file
2284 last if ! defined ($_ = readline $handle{$addr});
2285 }
2286 chomp;
2287 trace $_ if main::DEBUG && $to_trace;
2288
2289 # See if this line is the comment line that defines what property
2290 # value that code points that are not listed in the file should
2291 # have. The format or existence of these lines is not guaranteed
2292 # by Unicode since they are comments, but the documentation says
2293 # that this was added for machine-readability, so probably won't
2294 # change. This works starting in Unicode Version 5.0. They look
2295 # like:
2296 #
2297 # @missing: 0000..10FFFF; Not_Reordered
2298 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2299 # @missing: 0000..10FFFF; ; NaN
2300 #
2301 # Save the line for a later get_missings() call.
2302 if (/$missing_defaults_prefix/) {
2303 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2304 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2305 }
2306 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2307 my @defaults = split /\s* ; \s*/x, $_;
2308
2309 # The first field is the @missing, which ends in a
2310 # semi-colon, so can safely shift.
2311 shift @defaults;
2312
2313 # Some of these lines may have empty field placeholders
2314 # which get in the way. An example is:
2315 # @missing: 0000..10FFFF; ; NaN
2316 # Remove them. Process starting from the top so the
2317 # splice doesn't affect things still to be looked at.
2318 for (my $i = @defaults - 1; $i >= 0; $i--) {
2319 next if $defaults[$i] ne "";
2320 splice @defaults, $i, 1;
2321 }
2322
2323 # What's left should be just the property (maybe) and the
2324 # default. Having only one element means it doesn't have
2325 # the property.
2326 my $default;
2327 my $property;
2328 if (@defaults >= 1) {
2329 if (@defaults == 1) {
2330 $default = $defaults[0];
2331 }
2332 else {
2333 $property = $defaults[0];
2334 $default = $defaults[1];
2335 }
2336 }
2337
2338 if (@defaults < 1
2339 || @defaults > 2
2340 || ($default =~ /^</
2341 && $default !~ /^<code *point>$/i
2342 && $default !~ /^<none>$/i))
2343 {
2344 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2345 }
2346 else {
2347
2348 # If the property is missing from the line, it should
2349 # be the one for the whole file
2350 $property = $property{$addr} if ! defined $property;
2351
2352 # Change <none> to the null string, which is what it
2353 # really means. If the default is the code point
2354 # itself, set it to <code point>, which is what
2355 # Unicode uses (but sometimes they've forgotten the
2356 # space)
2357 if ($default =~ /^<none>$/i) {
2358 $default = "";
2359 }
2360 elsif ($default =~ /^<code *point>$/i) {
2361 $default = $CODE_POINT;
2362 }
2363
2364 # Store them as a sub-arrays with both components.
2365 push @{$missings{$addr}}, [ $default, $property ];
2366 }
2367 }
2368
2369 # There is nothing for the caller to process on this comment
2370 # line.
2371 next;
2372 }
2373
2374 # Remove comments and trailing space, and skip this line if the
2375 # result is empty
2376 s/#.*//;
2377 s/\s+$//;
2378 next if /^$/;
2379
2380 # Call any handlers for this line, and skip further processing of
2381 # the line if the handler sets the line to null.
2382 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2383 &{$sub_ref}($self);
2384 next LINE if /^$/;
2385 }
2386
2387 # Here the line is ok. return success.
2388 return 1;
2389 } # End of looping through lines.
2390
2391 # If there is an EOF handler, call it (only once) and if it generates
2392 # more lines to process go back in the loop to handle them.
2393 if ($eof_handler{$addr}) {
2394 &{$eof_handler{$addr}}($self);
2395 $eof_handler{$addr} = ""; # Currently only get one shot at it.
2396 goto LINE if $added_lines{$addr};
2397 }
2398
2399 # Return failure -- no more lines.
2400 return 0;
2401
2402 }
2403
2404# Not currently used, not fully tested.
2405# sub peek {
2406# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
2407# # record. Not callable from an each_line_handler(), nor does it call
2408# # an each_line_handler() on the line.
2409#
2410# my $self = shift;
ffe43484 2411# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2412#
2413# foreach my $inserted_ref (@{$added_lines{$addr}}) {
2414# my ($adjusted, $line) = @{$inserted_ref};
2415# next if $adjusted;
2416#
2417# # Remove comments and trailing space, and return a non-empty
2418# # resulting line
2419# $line =~ s/#.*//;
2420# $line =~ s/\s+$//;
2421# return $line if $line ne "";
2422# }
2423#
2424# return if ! ref $handle{$addr}; # Don't read unless is real file
2425# while (1) { # Loop until find non-comment, non-empty line
2426# local $to_trace = 1 if main::DEBUG;
2427# trace $_ if main::DEBUG && $to_trace;
2428# return if ! defined (my $line = readline $handle{$addr});
2429# chomp $line;
2430# push @{$added_lines{$addr}}, [ 0, $line ];
2431#
2432# $line =~ s/#.*//;
2433# $line =~ s/\s+$//;
2434# return $line if $line ne "";
2435# }
2436#
2437# return;
2438# }
2439
2440
2441 sub insert_lines {
2442 # Lines can be inserted so that it looks like they were in the input
2443 # file at the place it was when this routine is called. See also
2444 # insert_adjusted_lines(). Lines inserted via this routine go through
2445 # any each_line_handler()
2446
2447 my $self = shift;
2448
2449 # Each inserted line is an array, with the first element being 0 to
2450 # indicate that this line hasn't been adjusted, and needs to be
2451 # processed.
f998e60c 2452 no overloading;
051df77b 2453 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
2454 return;
2455 }
2456
2457 sub insert_adjusted_lines {
2458 # Lines can be inserted so that it looks like they were in the input
2459 # file at the place it was when this routine is called. See also
2460 # insert_lines(). Lines inserted via this routine are already fully
2461 # adjusted, ready to be processed; each_line_handler()s handlers will
2462 # not be called. This means this is not a completely general
2463 # facility, as only the last each_line_handler on the stack should
2464 # call this. It could be made more general, by passing to each of the
2465 # line_handlers their position on the stack, which they would pass on
2466 # to this routine, and that would replace the boolean first element in
2467 # the anonymous array pushed here, so that the next_line routine could
2468 # use that to call only those handlers whose index is after it on the
2469 # stack. But this is overkill for what is needed now.
2470
2471 my $self = shift;
2472 trace $_[0] if main::DEBUG && $to_trace;
2473
2474 # Each inserted line is an array, with the first element being 1 to
2475 # indicate that this line has been adjusted
f998e60c 2476 no overloading;
051df77b 2477 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
2478 return;
2479 }
2480
2481 sub get_missings {
2482 # Returns the stored up @missings lines' values, and clears the list.
2483 # The values are in an array, consisting of the default in the first
2484 # element, and the property in the 2nd. However, since these lines
2485 # can be stacked up, the return is an array of all these arrays.
2486
2487 my $self = shift;
2488 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2489
ffe43484 2490 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2491
2492 # If not accepting a list return, just return the first one.
2493 return shift @{$missings{$addr}} unless wantarray;
2494
2495 my @return = @{$missings{$addr}};
2496 undef @{$missings{$addr}};
2497 return @return;
2498 }
2499
2500 sub _insert_property_into_line {
2501 # Add a property field to $_, if this file requires it.
2502
f998e60c 2503 my $self = shift;
ffe43484 2504 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 2505 my $property = $property{$addr};
99870f4d
KW
2506 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2507
2508 $_ =~ s/(;|$)/; $property$1/;
2509 return;
2510 }
2511
2512 sub carp_bad_line {
2513 # Output consistent error messages, using either a generic one, or the
2514 # one given by the optional parameter. To avoid gazillions of the
2515 # same message in case the syntax of a file is way off, this routine
2516 # only outputs the first instance of each message, incrementing a
2517 # count so the totals can be output at the end of the file.
2518
2519 my $self = shift;
2520 my $message = shift;
2521 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2522
ffe43484 2523 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2524
2525 $message = 'Unexpected line' unless $message;
2526
2527 # No trailing punctuation so as to fit with our addenda.
2528 $message =~ s/[.:;,]$//;
2529
2530 # If haven't seen this exact message before, output it now. Otherwise
2531 # increment the count of how many times it has occurred
2532 unless ($errors{$addr}->{$message}) {
2533 Carp::my_carp("$message in '$_' in "
f998e60c 2534 . $file{$addr}
99870f4d
KW
2535 . " at line $.. Skipping this line;");
2536 $errors{$addr}->{$message} = 1;
2537 }
2538 else {
2539 $errors{$addr}->{$message}++;
2540 }
2541
2542 # Clear the line to prevent any further (meaningful) processing of it.
2543 $_ = "";
2544
2545 return;
2546 }
2547} # End closure
2548
2549package Multi_Default;
2550
2551# Certain properties in early versions of Unicode had more than one possible
2552# default for code points missing from the files. In these cases, one
2553# default applies to everything left over after all the others are applied,
2554# and for each of the others, there is a description of which class of code
2555# points applies to it. This object helps implement this by storing the
2556# defaults, and for all but that final default, an eval string that generates
2557# the class that it applies to.
2558
2559
2560{ # Closure
2561
2562 main::setup_package();
2563
2564 my %class_defaults;
2565 # The defaults structure for the classes
2566 main::set_access('class_defaults', \%class_defaults);
2567
2568 my %other_default;
2569 # The default that applies to everything left over.
2570 main::set_access('other_default', \%other_default, 'r');
2571
2572
2573 sub new {
2574 # The constructor is called with default => eval pairs, terminated by
2575 # the left-over default. e.g.
2576 # Multi_Default->new(
2577 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
2578 # - 0x200D',
2579 # 'R' => 'some other expression that evaluates to code points',
2580 # .
2581 # .
2582 # .
2583 # 'U'));
2584
2585 my $class = shift;
2586
2587 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 2588 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2589
2590 while (@_ > 1) {
2591 my $default = shift;
2592 my $eval = shift;
2593 $class_defaults{$addr}->{$default} = $eval;
2594 }
2595
2596 $other_default{$addr} = shift;
2597
2598 return $self;
2599 }
2600
2601 sub get_next_defaults {
2602 # Iterates and returns the next class of defaults.
2603 my $self = shift;
2604 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2605
ffe43484 2606 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2607
2608 return each %{$class_defaults{$addr}};
2609 }
2610}
2611
2612package Alias;
2613
2614# An alias is one of the names that a table goes by. This class defines them
2615# including some attributes. Everything is currently setup in the
2616# constructor.
2617
2618
2619{ # Closure
2620
2621 main::setup_package();
2622
2623 my %name;
2624 main::set_access('name', \%name, 'r');
2625
2626 my %loose_match;
2627 # Determined by the constructor code if this name should match loosely or
2628 # not. The constructor parameters can override this, but it isn't fully
2629 # implemented, as should have ability to override Unicode one's via
2630 # something like a set_loose_match()
2631 main::set_access('loose_match', \%loose_match, 'r');
2632
2633 my %make_pod_entry;
2634 # Some aliases should not get their own entries because they are covered
2635 # by a wild-card, and some we want to discourage use of. Binary
2636 main::set_access('make_pod_entry', \%make_pod_entry, 'r');
2637
2638 my %status;
2639 # Aliases have a status, like deprecated, or even suppressed (which means
2640 # they don't appear in documentation). Enum
2641 main::set_access('status', \%status, 'r');
2642
2643 my %externally_ok;
2644 # Similarly, some aliases should not be considered as usable ones for
2645 # external use, such as file names, or we don't want documentation to
2646 # recommend them. Boolean
2647 main::set_access('externally_ok', \%externally_ok, 'r');
2648
2649 sub new {
2650 my $class = shift;
2651
2652 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2653 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2654
2655 $name{$addr} = shift;
2656 $loose_match{$addr} = shift;
2657 $make_pod_entry{$addr} = shift;
2658 $externally_ok{$addr} = shift;
2659 $status{$addr} = shift;
2660
2661 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2662
2663 # Null names are never ok externally
2664 $externally_ok{$addr} = 0 if $name{$addr} eq "";
2665
2666 return $self;
2667 }
2668}
2669
2670package Range;
2671
2672# A range is the basic unit for storing code points, and is described in the
2673# comments at the beginning of the program. Each range has a starting code
2674# point; an ending code point (not less than the starting one); a value
2675# that applies to every code point in between the two end-points, inclusive;
2676# and an enum type that applies to the value. The type is for the user's
2677# convenience, and has no meaning here, except that a non-zero type is
2678# considered to not obey the normal Unicode rules for having standard forms.
2679#
2680# The same structure is used for both map and match tables, even though in the
2681# latter, the value (and hence type) is irrelevant and could be used as a
2682# comment. In map tables, the value is what all the code points in the range
2683# map to. Type 0 values have the standardized version of the value stored as
2684# well, so as to not have to recalculate it a lot.
2685
2686sub trace { return main::trace(@_); }
2687
2688{ # Closure
2689
2690 main::setup_package();
2691
2692 my %start;
2693 main::set_access('start', \%start, 'r', 's');
2694
2695 my %end;
2696 main::set_access('end', \%end, 'r', 's');
2697
2698 my %value;
2699 main::set_access('value', \%value, 'r');
2700
2701 my %type;
2702 main::set_access('type', \%type, 'r');
2703
2704 my %standard_form;
2705 # The value in internal standard form. Defined only if the type is 0.
2706 main::set_access('standard_form', \%standard_form);
2707
2708 # Note that if these fields change, the dump() method should as well
2709
2710 sub new {
2711 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
2712 my $class = shift;
2713
2714 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2715 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2716
2717 $start{$addr} = shift;
2718 $end{$addr} = shift;
2719
2720 my %args = @_;
2721
2722 my $value = delete $args{'Value'}; # Can be 0
2723 $value = "" unless defined $value;
2724 $value{$addr} = $value;
2725
2726 $type{$addr} = delete $args{'Type'} || 0;
2727
2728 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2729
2730 if (! $type{$addr}) {
2731 $standard_form{$addr} = main::standardize($value);
2732 }
2733
2734 return $self;
2735 }
2736
2737 use overload
2738 fallback => 0,
2739 qw("") => "_operator_stringify",
2740 "." => \&main::_operator_dot,
2741 ;
2742
2743 sub _operator_stringify {
2744 my $self = shift;
ffe43484 2745 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2746
2747 # Output it like '0041..0065 (value)'
2748 my $return = sprintf("%04X", $start{$addr})
2749 . '..'
2750 . sprintf("%04X", $end{$addr});
2751 my $value = $value{$addr};
2752 my $type = $type{$addr};
2753 $return .= ' (';
2754 $return .= "$value";
2755 $return .= ", Type=$type" if $type != 0;
2756 $return .= ')';
2757
2758 return $return;
2759 }
2760
2761 sub standard_form {
2762 # The standard form is the value itself if the standard form is
2763 # undefined (that is if the value is special)
2764
2765 my $self = shift;
2766 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2767
ffe43484 2768 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2769
2770 return $standard_form{$addr} if defined $standard_form{$addr};
2771 return $value{$addr};
2772 }
2773
2774 sub dump {
2775 # Human, not machine readable. For machine readable, comment out this
2776 # entire routine and let the standard one take effect.
2777 my $self = shift;
2778 my $indent = shift;
2779 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2780
ffe43484 2781 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2782
2783 my $return = $indent
2784 . sprintf("%04X", $start{$addr})
2785 . '..'
2786 . sprintf("%04X", $end{$addr})
2787 . " '$value{$addr}';";
2788 if (! defined $standard_form{$addr}) {
2789 $return .= "(type=$type{$addr})";
2790 }
2791 elsif ($standard_form{$addr} ne $value{$addr}) {
2792 $return .= "(standard '$standard_form{$addr}')";
2793 }
2794 return $return;
2795 }
2796} # End closure
2797
2798package _Range_List_Base;
2799
2800# Base class for range lists. A range list is simply an ordered list of
2801# ranges, so that the ranges with the lowest starting numbers are first in it.
2802#
2803# When a new range is added that is adjacent to an existing range that has the
2804# same value and type, it merges with it to form a larger range.
2805#
2806# Ranges generally do not overlap, except that there can be multiple entries
2807# of single code point ranges. This is because of NameAliases.txt.
2808#
2809# In this program, there is a standard value such that if two different
2810# values, have the same standard value, they are considered equivalent. This
2811# value was chosen so that it gives correct results on Unicode data
2812
2813# There are a number of methods to manipulate range lists, and some operators
2814# are overloaded to handle them.
2815
99870f4d
KW
2816sub trace { return main::trace(@_); }
2817
2818{ # Closure
2819
2820 our $addr;
2821
2822 main::setup_package();
2823
2824 my %ranges;
2825 # The list of ranges
2826 main::set_access('ranges', \%ranges, 'readable_array');
2827
2828 my %max;
2829 # The highest code point in the list. This was originally a method, but
2830 # actual measurements said it was used a lot.
2831 main::set_access('max', \%max, 'r');
2832
2833 my %each_range_iterator;
2834 # Iterator position for each_range()
2835 main::set_access('each_range_iterator', \%each_range_iterator);
2836
2837 my %owner_name_of;
2838 # Name of parent this is attached to, if any. Solely for better error
2839 # messages.
2840 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
2841
2842 my %_search_ranges_cache;
2843 # A cache of the previous result from _search_ranges(), for better
2844 # performance
2845 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
2846
2847 sub new {
2848 my $class = shift;
2849 my %args = @_;
2850
2851 # Optional initialization data for the range list.
2852 my $initialize = delete $args{'Initialize'};
2853
2854 my $self;
2855
2856 # Use _union() to initialize. _union() returns an object of this
2857 # class, which means that it will call this constructor recursively.
2858 # But it won't have this $initialize parameter so that it won't
2859 # infinitely loop on this.
2860 return _union($class, $initialize, %args) if defined $initialize;
2861
2862 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2863 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2864
2865 # Optional parent object, only for debug info.
2866 $owner_name_of{$addr} = delete $args{'Owner'};
2867 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
2868
2869 # Stringify, in case it is an object.
2870 $owner_name_of{$addr} = "$owner_name_of{$addr}";
2871
2872 # This is used only for error messages, and so a colon is added
2873 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
2874
2875 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2876
2877 # Max is initialized to a negative value that isn't adjacent to 0,
2878 # for simpler tests
2879 $max{$addr} = -2;
2880
2881 $_search_ranges_cache{$addr} = 0;
2882 $ranges{$addr} = [];
2883
2884 return $self;
2885 }
2886
2887 use overload
2888 fallback => 0,
2889 qw("") => "_operator_stringify",
2890 "." => \&main::_operator_dot,
2891 ;
2892
2893 sub _operator_stringify {
2894 my $self = shift;
ffe43484 2895 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2896
2897 return "Range_List attached to '$owner_name_of{$addr}'"
2898 if $owner_name_of{$addr};
2899 return "anonymous Range_List " . \$self;
2900 }
2901
2902 sub _union {
2903 # Returns the union of the input code points. It can be called as
2904 # either a constructor or a method. If called as a method, the result
2905 # will be a new() instance of the calling object, containing the union
2906 # of that object with the other parameter's code points; if called as
2907 # a constructor, the first parameter gives the class the new object
2908 # should be, and the second parameter gives the code points to go into
2909 # it.
2910 # In either case, there are two parameters looked at by this routine;
2911 # any additional parameters are passed to the new() constructor.
2912 #
2913 # The code points can come in the form of some object that contains
2914 # ranges, and has a conventionally named method to access them; or
2915 # they can be an array of individual code points (as integers); or
2916 # just a single code point.
2917 #
2918 # If they are ranges, this routine doesn't make any effort to preserve
2919 # the range values of one input over the other. Therefore this base
2920 # class should not allow _union to be called from other than
2921 # initialization code, so as to prevent two tables from being added
2922 # together where the range values matter. The general form of this
2923 # routine therefore belongs in a derived class, but it was moved here
2924 # to avoid duplication of code. The failure to overload this in this
2925 # class keeps it safe.
2926 #
2927
2928 my $self;
2929 my @args; # Arguments to pass to the constructor
2930
2931 my $class = shift;
2932
2933 # If a method call, will start the union with the object itself, and
2934 # the class of the new object will be the same as self.
2935 if (ref $class) {
2936 $self = $class;
2937 $class = ref $self;
2938 push @args, $self;
2939 }
2940
2941 # Add the other required parameter.
2942 push @args, shift;
2943 # Rest of parameters are passed on to the constructor
2944
2945 # Accumulate all records from both lists.
2946 my @records;
2947 for my $arg (@args) {
2948 #local $to_trace = 0 if main::DEBUG;
2949 trace "argument = $arg" if main::DEBUG && $to_trace;
2950 if (! defined $arg) {
2951 my $message = "";
2952 if (defined $self) {
f998e60c 2953 no overloading;
051df77b 2954 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2955 }
2956 Carp::my_carp_bug($message .= "Undefined argument to _union. No union done.");
2957 return;
2958 }
2959 $arg = [ $arg ] if ! ref $arg;
2960 my $type = ref $arg;
2961 if ($type eq 'ARRAY') {
2962 foreach my $element (@$arg) {
2963 push @records, Range->new($element, $element);
2964 }
2965 }
2966 elsif ($arg->isa('Range')) {
2967 push @records, $arg;
2968 }
2969 elsif ($arg->can('ranges')) {
2970 push @records, $arg->ranges;
2971 }
2972 else {
2973 my $message = "";
2974 if (defined $self) {
f998e60c 2975 no overloading;
051df77b 2976 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2977 }
2978 Carp::my_carp_bug($message . "Cannot take the union of a $type. No union done.");
2979 return;
2980 }
2981 }
2982
2983 # Sort with the range containing the lowest ordinal first, but if
2984 # two ranges start at the same code point, sort with the bigger range
2985 # of the two first, because it takes fewer cycles.
2986 @records = sort { ($a->start <=> $b->start)
2987 or
2988 # if b is shorter than a, b->end will be
2989 # less than a->end, and we want to select
2990 # a, so want to return -1
2991 ($b->end <=> $a->end)
2992 } @records;
2993
2994 my $new = $class->new(@_);
2995
2996 # Fold in records so long as they add new information.
2997 for my $set (@records) {
2998 my $start = $set->start;
2999 my $end = $set->end;
3000 my $value = $set->value;
3001 if ($start > $new->max) {
3002 $new->_add_delete('+', $start, $end, $value);
3003 }
3004 elsif ($end > $new->max) {
3005 $new->_add_delete('+', $new->max +1, $end, $value);
3006 }
3007 }
3008
3009 return $new;
3010 }
3011
3012 sub range_count { # Return the number of ranges in the range list
3013 my $self = shift;
3014 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3015
f998e60c 3016 no overloading;
051df77b 3017 return scalar @{$ranges{pack 'J', $self}};
99870f4d
KW
3018 }
3019
3020 sub min {
3021 # Returns the minimum code point currently in the range list, or if
3022 # the range list is empty, 2 beyond the max possible. This is a
3023 # method because used so rarely, that not worth saving between calls,
3024 # and having to worry about changing it as ranges are added and
3025 # deleted.
3026
3027 my $self = shift;
3028 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3029
ffe43484 3030 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3031
3032 # If the range list is empty, return a large value that isn't adjacent
3033 # to any that could be in the range list, for simpler tests
3034 return $LAST_UNICODE_CODEPOINT + 2 unless scalar @{$ranges{$addr}};
3035 return $ranges{$addr}->[0]->start;
3036 }
3037
3038 sub contains {
3039 # Boolean: Is argument in the range list? If so returns $i such that:
3040 # range[$i]->end < $codepoint <= range[$i+1]->end
3041 # which is one beyond what you want; this is so that the 0th range
3042 # doesn't return false
3043 my $self = shift;
3044 my $codepoint = shift;
3045 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3046
99870f4d
KW
3047 my $i = $self->_search_ranges($codepoint);
3048 return 0 unless defined $i;
3049
3050 # The search returns $i, such that
3051 # range[$i-1]->end < $codepoint <= range[$i]->end
3052 # So is in the table if and only iff it is at least the start position
3053 # of range $i.
f998e60c 3054 no overloading;
051df77b 3055 return 0 if $ranges{pack 'J', $self}->[$i]->start > $codepoint;
99870f4d
KW
3056 return $i + 1;
3057 }
3058
2f7a8815
KW
3059 sub containing_range {
3060 # Returns the range object that contains the code point, undef if none
3061
3062 my $self = shift;
3063 my $codepoint = shift;
3064 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3065
3066 my $i = $self->contains($codepoint);
3067 return unless $i;
3068
3069 # contains() returns 1 beyond where we should look
3070 no overloading;
3071 return $ranges{pack 'J', $self}->[$i-1];
3072 }
3073
99870f4d
KW
3074 sub value_of {
3075 # Returns the value associated with the code point, undef if none
3076
3077 my $self = shift;
3078 my $codepoint = shift;
3079 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3080
d69c231b
KW
3081 my $range = $self->containing_range($codepoint);
3082 return unless defined $range;
99870f4d 3083
d69c231b 3084 return $range->value;
99870f4d
KW
3085 }
3086
0a9dbafc
KW
3087 sub type_of {
3088 # Returns the type of the range containing the code point, undef if
3089 # the code point is not in the table
3090
3091 my $self = shift;
3092 my $codepoint = shift;
3093 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3094
3095 my $range = $self->containing_range($codepoint);
3096 return unless defined $range;
3097
3098 return $range->type;
3099 }
3100
99870f4d
KW
3101 sub _search_ranges {
3102 # Find the range in the list which contains a code point, or where it
3103 # should go if were to add it. That is, it returns $i, such that:
3104 # range[$i-1]->end < $codepoint <= range[$i]->end
3105 # Returns undef if no such $i is possible (e.g. at end of table), or
3106 # if there is an error.
3107
3108 my $self = shift;
3109 my $code_point = shift;
3110 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3111
ffe43484 3112 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3113
3114 return if $code_point > $max{$addr};
3115 my $r = $ranges{$addr}; # The current list of ranges
3116 my $range_list_size = scalar @$r;
3117 my $i;
3118
3119 use integer; # want integer division
3120
3121 # Use the cached result as the starting guess for this one, because,
3122 # an experiment on 5.1 showed that 90% of the time the cache was the
3123 # same as the result on the next call (and 7% it was one less).
3124 $i = $_search_ranges_cache{$addr};
3125 $i = 0 if $i >= $range_list_size; # Reset if no longer valid (prob.
3126 # from an intervening deletion
3127 #local $to_trace = 1 if main::DEBUG;
3128 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);
3129 return $i if $code_point <= $r->[$i]->end
3130 && ($i == 0 || $r->[$i-1]->end < $code_point);
3131
3132 # Here the cache doesn't yield the correct $i. Try adding 1.
3133 if ($i < $range_list_size - 1
3134 && $r->[$i]->end < $code_point &&
3135 $code_point <= $r->[$i+1]->end)
3136 {
3137 $i++;
3138 trace "next \$i is correct: $i" if main::DEBUG && $to_trace;
3139 $_search_ranges_cache{$addr} = $i;
3140 return $i;
3141 }
3142
3143 # Here, adding 1 also didn't work. We do a binary search to
3144 # find the correct position, starting with current $i
3145 my $lower = 0;
3146 my $upper = $range_list_size - 1;
3147 while (1) {
3148 trace "top of loop i=$i:", sprintf("%04X", $r->[$lower]->start), "[$lower] .. ", sprintf("%04X", $r->[$i]->start), "[$i] .. ", sprintf("%04X", $r->[$upper]->start), "[$upper]" if main::DEBUG && $to_trace;
3149
3150 if ($code_point <= $r->[$i]->end) {
3151
3152 # Here we have met the upper constraint. We can quit if we
3153 # also meet the lower one.
3154 last if $i == 0 || $r->[$i-1]->end < $code_point;
3155
3156 $upper = $i; # Still too high.
3157
3158 }
3159 else {
3160
3161 # Here, $r[$i]->end < $code_point, so look higher up.
3162 $lower = $i;
3163 }
3164
3165 # Split search domain in half to try again.
3166 my $temp = ($upper + $lower) / 2;
3167
3168 # No point in continuing unless $i changes for next time
3169 # in the loop.
3170 if ($temp == $i) {
3171
3172 # We can't reach the highest element because of the averaging.
3173 # So if one below the upper edge, force it there and try one
3174 # more time.
3175 if ($i == $range_list_size - 2) {
3176
3177 trace "Forcing to upper edge" if main::DEBUG && $to_trace;
3178 $i = $range_list_size - 1;
3179
3180 # Change $lower as well so if fails next time through,
3181 # taking the average will yield the same $i, and we will
3182 # quit with the error message just below.
3183 $lower = $i;
3184 next;
3185 }
3186 Carp::my_carp_bug("$owner_name_of{$addr}Can't find where the range ought to go. No action taken.");
3187 return;
3188 }
3189 $i = $temp;
3190 } # End of while loop
3191
3192 if (main::DEBUG && $to_trace) {
3193 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i;
3194 trace "i= [ $i ]", $r->[$i];
3195 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < $range_list_size - 1;
3196 }
3197
3198 # Here we have found the offset. Cache it as a starting point for the
3199 # next call.
3200 $_search_ranges_cache{$addr} = $i;
3201 return $i;
3202 }
3203
3204 sub _add_delete {
3205 # Add, replace or delete ranges to or from a list. The $type
3206 # parameter gives which:
3207 # '+' => insert or replace a range, returning a list of any changed
3208 # ranges.
3209 # '-' => delete a range, returning a list of any deleted ranges.
3210 #
3211 # The next three parameters give respectively the start, end, and
3212 # value associated with the range. 'value' should be null unless the
3213 # operation is '+';
3214 #
3215 # The range list is kept sorted so that the range with the lowest
3216 # starting position is first in the list, and generally, adjacent
c1739a4a 3217 # ranges with the same values are merged into a single larger one (see
99870f4d
KW
3218 # exceptions below).
3219 #
c1739a4a 3220 # There are more parameters; all are key => value pairs:
99870f4d
KW
3221 # Type gives the type of the value. It is only valid for '+'.
3222 # All ranges have types; if this parameter is omitted, 0 is
3223 # assumed. Ranges with type 0 are assumed to obey the
3224 # Unicode rules for casing, etc; ranges with other types are
3225 # not. Otherwise, the type is arbitrary, for the caller's
3226 # convenience, and looked at only by this routine to keep
3227 # adjacent ranges of different types from being merged into
3228 # a single larger range, and when Replace =>
3229 # $IF_NOT_EQUIVALENT is specified (see just below).
3230 # Replace determines what to do if the range list already contains
3231 # ranges which coincide with all or portions of the input
3232 # range. It is only valid for '+':
3233 # => $NO means that the new value is not to replace
3234 # any existing ones, but any empty gaps of the
3235 # range list coinciding with the input range
3236 # will be filled in with the new value.
3237 # => $UNCONDITIONALLY means to replace the existing values with
3238 # this one unconditionally. However, if the
3239 # new and old values are identical, the
3240 # replacement is skipped to save cycles
3241 # => $IF_NOT_EQUIVALENT means to replace the existing values
3242 # with this one if they are not equivalent.
3243 # Ranges are equivalent if their types are the
c1739a4a 3244 # same, and they are the same string; or if
99870f4d
KW
3245 # both are type 0 ranges, if their Unicode
3246 # standard forms are identical. In this last
3247 # case, the routine chooses the more "modern"
3248 # one to use. This is because some of the
3249 # older files are formatted with values that
3250 # are, for example, ALL CAPs, whereas the
3251 # derived files have a more modern style,
3252 # which looks better. By looking for this
3253 # style when the pre-existing and replacement
3254 # standard forms are the same, we can move to
3255 # the modern style
3256 # => $MULTIPLE means that if this range duplicates an
3257 # existing one, but has a different value,
3258 # don't replace the existing one, but insert
3259 # this, one so that the same range can occur
53d84487
KW
3260 # multiple times. They are stored LIFO, so
3261 # that the final one inserted is the first one
3262 # returned in an ordered search of the table.
99870f4d
KW
3263 # => anything else is the same as => $IF_NOT_EQUIVALENT
3264 #
c1739a4a
KW
3265 # "same value" means identical for non-type-0 ranges, and it means
3266 # having the same standard forms for type-0 ranges.
99870f4d
KW
3267
3268 return Carp::carp_too_few_args(\@_, 5) if main::DEBUG && @_ < 5;
3269
3270 my $self = shift;
3271 my $operation = shift; # '+' for add/replace; '-' for delete;
3272 my $start = shift;
3273 my $end = shift;
3274 my $value = shift;
3275
3276 my %args = @_;
3277
3278 $value = "" if not defined $value; # warning: $value can be "0"
3279
3280 my $replace = delete $args{'Replace'};
3281 $replace = $IF_NOT_EQUIVALENT unless defined $replace;
3282
3283 my $type = delete $args{'Type'};
3284 $type = 0 unless defined $type;
3285
3286 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3287
ffe43484 3288 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3289
3290 if ($operation ne '+' && $operation ne '-') {
3291 Carp::my_carp_bug("$owner_name_of{$addr}First parameter to _add_delete must be '+' or '-'. No action taken.");
3292 return;
3293 }
3294 unless (defined $start && defined $end) {
3295 Carp::my_carp_bug("$owner_name_of{$addr}Undefined start and/or end to _add_delete. No action taken.");
3296 return;
3297 }
3298 unless ($end >= $start) {
3299 Carp::my_carp_bug("$owner_name_of{$addr}End of range (" . sprintf("%04X", $end) . ") must not be before start (" . sprintf("%04X", $start) . "). No action taken.");
3300 return;
3301 }
3302 #local $to_trace = 1 if main::DEBUG;
3303
3304 if ($operation eq '-') {
3305 if ($replace != $IF_NOT_EQUIVALENT) {
3306 Carp::my_carp_bug("$owner_name_of{$addr}Replace => \$IF_NOT_EQUIVALENT is required when deleting a range from a range list. Assuming Replace => \$IF_NOT_EQUIVALENT.");
3307 $replace = $IF_NOT_EQUIVALENT;
3308 }
3309 if ($type) {
3310 Carp::my_carp_bug("$owner_name_of{$addr}Type => 0 is required when deleting a range from a range list. Assuming Type => 0.");
3311 $type = 0;
3312 }
3313 if ($value ne "") {
3314 Carp::my_carp_bug("$owner_name_of{$addr}Value => \"\" is required when deleting a range from a range list. Assuming Value => \"\".");
3315 $value = "";
3316 }
3317 }
3318
3319 my $r = $ranges{$addr}; # The current list of ranges
3320 my $range_list_size = scalar @$r; # And its size
3321 my $max = $max{$addr}; # The current high code point in
3322 # the list of ranges
3323
3324 # Do a special case requiring fewer machine cycles when the new range
3325 # starts after the current highest point. The Unicode input data is
3326 # structured so this is common.
3327 if ($start > $max) {
3328
3329 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) type=$type" if main::DEBUG && $to_trace;
3330 return if $operation eq '-'; # Deleting a non-existing range is a
3331 # no-op
3332
3333 # If the new range doesn't logically extend the current final one
3334 # in the range list, create a new range at the end of the range
3335 # list. (max cleverly is initialized to a negative number not
3336 # adjacent to 0 if the range list is empty, so even adding a range
3337 # to an empty range list starting at 0 will have this 'if'
3338 # succeed.)
3339 if ($start > $max + 1 # non-adjacent means can't extend.
3340 || @{$r}[-1]->value ne $value # values differ, can't extend.
3341 || @{$r}[-1]->type != $type # types differ, can't extend.
3342 ) {
3343 push @$r, Range->new($start, $end,
3344 Value => $value,
3345 Type => $type);
3346 }
3347 else {
3348
3349 # Here, the new range starts just after the current highest in
3350 # the range list, and they have the same type and value.
3351 # Extend the current range to incorporate the new one.
3352 @{$r}[-1]->set_end($end);
3353 }
3354
3355 # This becomes the new maximum.
3356 $max{$addr} = $end;
3357
3358 return;
3359 }
3360 #local $to_trace = 0 if main::DEBUG;
3361
3362 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) replace=$replace" if main::DEBUG && $to_trace;
3363
3364 # Here, the input range isn't after the whole rest of the range list.
3365 # Most likely 'splice' will be needed. The rest of the routine finds
3366 # the needed splice parameters, and if necessary, does the splice.
3367 # First, find the offset parameter needed by the splice function for
3368 # the input range. Note that the input range may span multiple
3369 # existing ones, but we'll worry about that later. For now, just find
3370 # the beginning. If the input range is to be inserted starting in a
3371 # position not currently in the range list, it must (obviously) come
3372 # just after the range below it, and just before the range above it.
3373 # Slightly less obviously, it will occupy the position currently
3374 # occupied by the range that is to come after it. More formally, we
3375 # are looking for the position, $i, in the array of ranges, such that:
3376 #
3377 # r[$i-1]->start <= r[$i-1]->end < $start < r[$i]->start <= r[$i]->end
3378 #
3379 # (The ordered relationships within existing ranges are also shown in
3380 # the equation above). However, if the start of the input range is
3381 # within an existing range, the splice offset should point to that
3382 # existing range's position in the list; that is $i satisfies a
3383 # somewhat different equation, namely:
3384 #
3385 #r[$i-1]->start <= r[$i-1]->end < r[$i]->start <= $start <= r[$i]->end
3386 #
3387 # More briefly, $start can come before or after r[$i]->start, and at
3388 # this point, we don't know which it will be. However, these
3389 # two equations share these constraints:
3390 #
3391 # r[$i-1]->end < $start <= r[$i]->end
3392 #
3393 # And that is good enough to find $i.
3394
3395 my $i = $self->_search_ranges($start);
3396 if (! defined $i) {
3397 Carp::my_carp_bug("Searching $self for range beginning with $start unexpectedly returned undefined. Operation '$operation' not performed");
3398 return;
3399 }
3400
3401 # The search function returns $i such that:
3402 #
3403 # r[$i-1]->end < $start <= r[$i]->end
3404 #
3405 # That means that $i points to the first range in the range list
3406 # that could possibly be affected by this operation. We still don't
3407 # know if the start of the input range is within r[$i], or if it
3408 # points to empty space between r[$i-1] and r[$i].
3409 trace "[$i] is the beginning splice point. Existing range there is ", $r->[$i] if main::DEBUG && $to_trace;
3410
3411 # Special case the insertion of data that is not to replace any
3412 # existing data.
3413 if ($replace == $NO) { # If $NO, has to be operation '+'
3414 #local $to_trace = 1 if main::DEBUG;
3415 trace "Doesn't replace" if main::DEBUG && $to_trace;
3416
3417 # Here, the new range is to take effect only on those code points
3418 # that aren't already in an existing range. This can be done by
3419 # looking through the existing range list and finding the gaps in
3420 # the ranges that this new range affects, and then calling this
3421 # function recursively on each of those gaps, leaving untouched
3422 # anything already in the list. Gather up a list of the changed
3423 # gaps first so that changes to the internal state as new ranges
3424 # are added won't be a problem.
3425 my @gap_list;
3426
3427 # First, if the starting point of the input range is outside an
3428 # existing one, there is a gap from there to the beginning of the
3429 # existing range -- add a span to fill the part that this new
3430 # range occupies
3431 if ($start < $r->[$i]->start) {
3432 push @gap_list, Range->new($start,
3433 main::min($end,
3434 $r->[$i]->start - 1),
3435 Type => $type);
3436 trace "gap before $r->[$i] [$i], will add", $gap_list[-1] if main::DEBUG && $to_trace;
3437 }
3438
3439 # Then look through the range list for other gaps until we reach
3440 # the highest range affected by the input one.
3441 my $j;
3442 for ($j = $i+1; $j < $range_list_size; $j++) {
3443 trace "j=[$j]", $r->[$j] if main::DEBUG && $to_trace;
3444 last if $end < $r->[$j]->start;
3445
3446 # If there is a gap between when this range starts and the
3447 # previous one ends, add a span to fill it. Note that just
3448 # because there are two ranges doesn't mean there is a
3449 # non-zero gap between them. It could be that they have
3450 # different values or types
3451 if ($r->[$j-1]->end + 1 != $r->[$j]->start) {
3452 push @gap_list,
3453 Range->new($r->[$j-1]->end + 1,
3454 $r->[$j]->start - 1,
3455 Type => $type);
3456 trace "gap between $r->[$j-1] and $r->[$j] [$j], will add: $gap_list[-1]" if main::DEBUG && $to_trace;
3457 }
3458 }
3459
3460 # Here, we have either found an existing range in the range list,
3461 # beyond the area affected by the input one, or we fell off the
3462 # end of the loop because the input range affects the whole rest
3463 # of the range list. In either case, $j is 1 higher than the
3464 # highest affected range. If $j == $i, it means that there are no
3465 # affected ranges, that the entire insertion is in the gap between
3466 # r[$i-1], and r[$i], which we already have taken care of before
3467 # the loop.
3468 # On the other hand, if there are affected ranges, it might be
3469 # that there is a gap that needs filling after the final such
3470 # range to the end of the input range
3471 if ($r->[$j-1]->end < $end) {
3472 push @gap_list, Range->new(main::max($start,
3473 $r->[$j-1]->end + 1),
3474 $end,
3475 Type => $type);
3476 trace "gap after $r->[$j-1], will add $gap_list[-1]" if main::DEBUG && $to_trace;
3477 }
3478
3479 # Call recursively to fill in all the gaps.
3480 foreach my $gap (@gap_list) {
3481 $self->_add_delete($operation,
3482 $gap->start,
3483 $gap->end,
3484 $value,
3485 Type => $type);
3486 }
3487
3488 return;
3489 }
3490
53d84487
KW
3491 # Here, we have taken care of the case where $replace is $NO.
3492 # Remember that here, r[$i-1]->end < $start <= r[$i]->end
3493 # If inserting a multiple record, this is where it goes, before the
3494 # first (if any) existing one. This implies an insertion, and no
3495 # change to any existing ranges. Note that $i can be -1 if this new
3496 # range doesn't actually duplicate any existing, and comes at the
3497 # beginning of the list.
3498 if ($replace == $MULTIPLE) {
3499
3500 if ($start != $end) {
3501 Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple record when the range ($start..$end) contains more than one code point. No action taken.");
3502 return;
3503 }
3504
3505 # Don't add an exact duplicate, as it isn't really a multiple
3506 if ($end >= $r->[$i]->start) {
3507 if ($r->[$i]->start != $r->[$i]->end) {
3508 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.");
3509 return;
3510 }
3511 return if $value eq $r->[$i]->value && $type eq $r->[$i]->type;
3512 }
3513
3514 trace "Adding multiple record at $i with $start..$end, $value" if main::DEBUG && $to_trace;
3515 my @return = splice @$r,
3516 $i,
3517 0,
3518 Range->new($start,
3519 $end,
3520 Value => $value,
3521 Type => $type);
3522 if (main::DEBUG && $to_trace) {
3523 trace "After splice:";
3524 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3525 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3526 trace "i =[", $i, "]", $r->[$i] if $i >= 0;
3527 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3528 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3529 trace 'i+3=[', $i+3, ']', $r->[$i+3] if $i < @$r - 3;
3530 }
3531 return @return;
3532 }
3533
3534 # Here, we have taken care of $NO and $MULTIPLE replaces. This leaves
3535 # delete, insert, and replace either unconditionally or if not
3536 # equivalent. $i still points to the first potential affected range.
3537 # Now find the highest range affected, which will determine the length
3538 # parameter to splice. (The input range can span multiple existing
3539 # ones.) If this isn't a deletion, while we are looking through the
3540 # range list, see also if this is a replacement rather than a clean
3541 # insertion; that is if it will change the values of at least one
3542 # existing range. Start off assuming it is an insert, until find it
3543 # isn't.
3544 my $clean_insert = $operation eq '+';
99870f4d
KW
3545 my $j; # This will point to the highest affected range
3546
3547 # For non-zero types, the standard form is the value itself;
3548 my $standard_form = ($type) ? $value : main::standardize($value);
3549
3550 for ($j = $i; $j < $range_list_size; $j++) {
3551 trace "Looking for highest affected range; the one at $j is ", $r->[$j] if main::DEBUG && $to_trace;
3552
3553 # If find a range that it doesn't overlap into, we can stop
3554 # searching
3555 last if $end < $r->[$j]->start;
3556
969a34cc
KW
3557 # Here, overlaps the range at $j. If the values don't match,
3558 # and so far we think this is a clean insertion, it becomes a
3559 # non-clean insertion, i.e., a 'change' or 'replace' instead.
3560 if ($clean_insert) {
99870f4d 3561 if ($r->[$j]->standard_form ne $standard_form) {
969a34cc 3562 $clean_insert = 0;
56343c78
KW
3563 if ($replace == $CROAK) {
3564 main::croak("The range to add "
3565 . sprintf("%04X", $start)
3566 . '-'
3567 . sprintf("%04X", $end)
3568 . " with value '$value' overlaps an existing range $r->[$j]");
3569 }
99870f4d
KW
3570 }
3571 else {
3572
3573 # Here, the two values are essentially the same. If the
3574 # two are actually identical, replacing wouldn't change
3575 # anything so skip it.
3576 my $pre_existing = $r->[$j]->value;
3577 if ($pre_existing ne $value) {
3578
3579 # Here the new and old standardized values are the
3580 # same, but the non-standardized values aren't. If
3581 # replacing unconditionally, then replace
3582 if( $replace == $UNCONDITIONALLY) {
969a34cc 3583 $clean_insert = 0;
99870f4d
KW
3584 }
3585 else {
3586
3587 # Here, are replacing conditionally. Decide to
3588 # replace or not based on which appears to look
3589 # the "nicest". If one is mixed case and the
3590 # other isn't, choose the mixed case one.
3591 my $new_mixed = $value =~ /[A-Z]/
3592 && $value =~ /[a-z]/;
3593 my $old_mixed = $pre_existing =~ /[A-Z]/
3594 && $pre_existing =~ /[a-z]/;
3595
3596 if ($old_mixed != $new_mixed) {
969a34cc 3597 $clean_insert = 0 if $new_mixed;
99870f4d 3598 if (main::DEBUG && $to_trace) {
969a34cc
KW
3599 if ($clean_insert) {
3600 trace "Retaining $pre_existing over $value";
99870f4d
KW
3601 }
3602 else {
969a34cc 3603 trace "Replacing $pre_existing with $value";
99870f4d
KW
3604 }
3605 }
3606 }
3607 else {
3608
3609 # Here casing wasn't different between the two.
3610 # If one has hyphens or underscores and the
3611 # other doesn't, choose the one with the
3612 # punctuation.
3613 my $new_punct = $value =~ /[-_]/;
3614 my $old_punct = $pre_existing =~ /[-_]/;
3615
3616 if ($old_punct != $new_punct) {
969a34cc 3617 $clean_insert = 0 if $new_punct;
99870f4d 3618 if (main::DEBUG && $to_trace) {
969a34cc
KW
3619 if ($clean_insert) {
3620 trace "Retaining $pre_existing over $value";
99870f4d
KW
3621 }
3622 else {
969a34cc 3623 trace "Replacing $pre_existing with $value";
99870f4d
KW
3624 }
3625 }
3626 } # else existing one is just as "good";
3627 # retain it to save cycles.
3628 }
3629 }
3630 }
3631 }
3632 }
3633 } # End of loop looking for highest affected range.
3634
3635 # Here, $j points to one beyond the highest range that this insertion
3636 # affects (hence to beyond the range list if that range is the final
3637 # one in the range list).
3638
3639 # The splice length is all the affected ranges. Get it before
3640 # subtracting, for efficiency, so we don't have to later add 1.
3641 my $length = $j - $i;
3642
3643 $j--; # $j now points to the highest affected range.
3644 trace "Final affected range is $j: $r->[$j]" if main::DEBUG && $to_trace;
3645
99870f4d
KW
3646 # Here, have taken care of $NO and $MULTIPLE replaces.
3647 # $j points to the highest affected range. But it can be < $i or even
3648 # -1. These happen only if the insertion is entirely in the gap
3649 # between r[$i-1] and r[$i]. Here's why: j < i means that the j loop
3650 # above exited first time through with $end < $r->[$i]->start. (And
3651 # then we subtracted one from j) This implies also that $start <
3652 # $r->[$i]->start, but we know from above that $r->[$i-1]->end <
3653 # $start, so the entire input range is in the gap.
3654 if ($j < $i) {
3655
3656 # Here the entire input range is in the gap before $i.
3657
3658 if (main::DEBUG && $to_trace) {
3659 if ($i) {
3660 trace "Entire range is between $r->[$i-1] and $r->[$i]";
3661 }
3662 else {
3663 trace "Entire range is before $r->[$i]";
3664 }
3665 }
3666 return if $operation ne '+'; # Deletion of a non-existent range is
3667 # a no-op
3668 }
3669 else {
3670
969a34cc
KW
3671 # Here part of the input range is not in the gap before $i. Thus,
3672 # there is at least one affected one, and $j points to the highest
3673 # such one.
99870f4d
KW
3674
3675 # At this point, here is the situation:
3676 # This is not an insertion of a multiple, nor of tentative ($NO)
3677 # data.
3678 # $i points to the first element in the current range list that
3679 # may be affected by this operation. In fact, we know
3680 # that the range at $i is affected because we are in
3681 # the else branch of this 'if'
3682 # $j points to the highest affected range.
3683 # In other words,
3684 # r[$i-1]->end < $start <= r[$i]->end
3685 # And:
3686 # r[$i-1]->end < $start <= $end <= r[$j]->end
3687 #
3688 # Also:
969a34cc
KW
3689 # $clean_insert is a boolean which is set true if and only if
3690 # this is a "clean insertion", i.e., not a change nor a
3691 # deletion (multiple was handled above).
99870f4d
KW
3692
3693 # We now have enough information to decide if this call is a no-op
969a34cc
KW
3694 # or not. It is a no-op if this is an insertion of already
3695 # existing data.
99870f4d 3696
969a34cc 3697 if (main::DEBUG && $to_trace && $clean_insert
99870f4d
KW
3698 && $i == $j
3699 && $start >= $r->[$i]->start)
3700 {
3701 trace "no-op";
3702 }
969a34cc 3703 return if $clean_insert
99870f4d
KW
3704 && $i == $j # more than one affected range => not no-op
3705
3706 # Here, r[$i-1]->end < $start <= $end <= r[$i]->end
3707 # Further, $start and/or $end is >= r[$i]->start
3708 # The test below hence guarantees that
3709 # r[$i]->start < $start <= $end <= r[$i]->end
3710 # This means the input range is contained entirely in
3711 # the one at $i, so is a no-op
3712 && $start >= $r->[$i]->start;
3713 }
3714
3715 # Here, we know that some action will have to be taken. We have
3716 # calculated the offset and length (though adjustments may be needed)
3717 # for the splice. Now start constructing the replacement list.
3718 my @replacement;
3719 my $splice_start = $i;
3720
3721 my $extends_below;
3722 my $extends_above;
3723
3724 # See if should extend any adjacent ranges.
3725 if ($operation eq '-') { # Don't extend deletions
3726 $extends_below = $extends_above = 0;
3727 }
3728 else { # Here, should extend any adjacent ranges. See if there are
3729 # any.
3730 $extends_below = ($i > 0
3731 # can't extend unless adjacent
3732 && $r->[$i-1]->end == $start -1
3733 # can't extend unless are same standard value
3734 && $r->[$i-1]->standard_form eq $standard_form
3735 # can't extend unless share type
3736 && $r->[$i-1]->type == $type);
3737 $extends_above = ($j+1 < $range_list_size
3738 && $r->[$j+1]->start == $end +1
3739 && $r->[$j+1]->standard_form eq $standard_form
23822bda 3740 && $r->[$j+1]->type == $type);
99870f4d
KW
3741 }
3742 if ($extends_below && $extends_above) { # Adds to both
3743 $splice_start--; # start replace at element below
3744 $length += 2; # will replace on both sides
3745 trace "Extends both below and above ranges" if main::DEBUG && $to_trace;
3746
3747 # The result will fill in any gap, replacing both sides, and
3748 # create one large range.
3749 @replacement = Range->new($r->[$i-1]->start,
3750 $r->[$j+1]->end,
3751 Value => $value,
3752 Type => $type);
3753 }
3754 else {
3755
3756 # Here we know that the result won't just be the conglomeration of
3757 # a new range with both its adjacent neighbors. But it could
3758 # extend one of them.
3759
3760 if ($extends_below) {
3761
3762 # Here the new element adds to the one below, but not to the
3763 # one above. If inserting, and only to that one range, can
3764 # just change its ending to include the new one.
969a34cc 3765 if ($length == 0 && $clean_insert) {
99870f4d
KW
3766 $r->[$i-1]->set_end($end);
3767 trace "inserted range extends range to below so it is now $r->[$i-1]" if main::DEBUG && $to_trace;
3768 return;
3769 }
3770 else {
3771 trace "Changing inserted range to start at ", sprintf("%04X", $r->[$i-1]->start), " instead of ", sprintf("%04X", $start) if main::DEBUG && $to_trace;
3772 $splice_start--; # start replace at element below
3773 $length++; # will replace the element below
3774 $start = $r->[$i-1]->start;
3775 }
3776 }
3777 elsif ($extends_above) {
3778
3779 # Here the new element adds to the one above, but not below.
3780 # Mirror the code above
969a34cc 3781 if ($length == 0 && $clean_insert) {
99870f4d
KW
3782 $r->[$j+1]->set_start($start);
3783 trace "inserted range extends range to above so it is now $r->[$j+1]" if main::DEBUG && $to_trace;
3784 return;
3785 }
3786 else {
3787 trace "Changing inserted range to end at ", sprintf("%04X", $r->[$j+1]->end), " instead of ", sprintf("%04X", $end) if main::DEBUG && $to_trace;
3788 $length++; # will replace the element above
3789 $end = $r->[$j+1]->end;
3790 }
3791 }
3792
3793 trace "Range at $i is $r->[$i]" if main::DEBUG && $to_trace;
3794
3795 # Finally, here we know there will have to be a splice.
3796 # If the change or delete affects only the highest portion of the
3797 # first affected range, the range will have to be split. The
3798 # splice will remove the whole range, but will replace it by a new
3799 # range containing just the unaffected part. So, in this case,
3800 # add to the replacement list just this unaffected portion.
3801 if (! $extends_below
3802 && $start > $r->[$i]->start && $start <= $r->[$i]->end)
3803 {
3804 push @replacement,
3805 Range->new($r->[$i]->start,
3806 $start - 1,
3807 Value => $r->[$i]->value,
3808 Type => $r->[$i]->type);
3809 }
3810
3811 # In the case of an insert or change, but not a delete, we have to
3812 # put in the new stuff; this comes next.
3813 if ($operation eq '+') {
3814 push @replacement, Range->new($start,
3815 $end,
3816 Value => $value,
3817 Type => $type);
3818 }
3819
3820 trace "Range at $j is $r->[$j]" if main::DEBUG && $to_trace && $j != $i;
3821 #trace "$end >=", $r->[$j]->start, " && $end <", $r->[$j]->end if main::DEBUG && $to_trace;
3822
3823 # And finally, if we're changing or deleting only a portion of the
3824 # highest affected range, it must be split, as the lowest one was.
3825 if (! $extends_above
3826 && $j >= 0 # Remember that j can be -1 if before first
3827 # current element
3828 && $end >= $r->[$j]->start
3829 && $end < $r->[$j]->end)
3830 {
3831 push @replacement,
3832 Range->new($end + 1,
3833 $r->[$j]->end,
3834 Value => $r->[$j]->value,
3835 Type => $r->[$j]->type);
3836 }
3837 }
3838
3839 # And do the splice, as calculated above
3840 if (main::DEBUG && $to_trace) {
3841 trace "replacing $length element(s) at $i with ";
3842 foreach my $replacement (@replacement) {
3843 trace " $replacement";
3844 }
3845 trace "Before splice:";
3846 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3847 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3848 trace "i =[", $i, "]", $r->[$i];
3849 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3850 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3851 }
3852
3853 my @return = splice @$r, $splice_start, $length, @replacement;
3854
3855 if (main::DEBUG && $to_trace) {
3856 trace "After splice:";
3857 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3858 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3859 trace "i =[", $i, "]", $r->[$i];
3860 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3861 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
e6451557 3862 trace "removed ", @return if @return;
99870f4d
KW
3863 }
3864
3865 # An actual deletion could have changed the maximum in the list.
3866 # There was no deletion if the splice didn't return something, but
3867 # otherwise recalculate it. This is done too rarely to worry about
3868 # performance.
3869 if ($operation eq '-' && @return) {
3870 $max{$addr} = $r->[-1]->end;
3871 }
3872 return @return;
3873 }
3874
3875 sub reset_each_range { # reset the iterator for each_range();
3876 my $self = shift;
3877 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3878
f998e60c 3879 no overloading;
051df77b 3880 undef $each_range_iterator{pack 'J', $self};
99870f4d
KW
3881 return;
3882 }
3883
3884 sub each_range {
3885 # Iterate over each range in a range list. Results are undefined if
3886 # the range list is changed during the iteration.
3887
3888 my $self = shift;
3889 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3890
ffe43484 3891 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3892
3893 return if $self->is_empty;
3894
3895 $each_range_iterator{$addr} = -1
3896 if ! defined $each_range_iterator{$addr};
3897 $each_range_iterator{$addr}++;
3898 return $ranges{$addr}->[$each_range_iterator{$addr}]
3899 if $each_range_iterator{$addr} < @{$ranges{$addr}};
3900 undef $each_range_iterator{$addr};
3901 return;
3902 }
3903
3904 sub count { # Returns count of code points in range list
3905 my $self = shift;
3906 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3907
ffe43484 3908 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3909
3910 my $count = 0;
3911 foreach my $range (@{$ranges{$addr}}) {
3912 $count += $range->end - $range->start + 1;
3913 }
3914 return $count;
3915 }
3916
3917 sub delete_range { # Delete a range
3918 my $self = shift;
3919 my $start = shift;
3920 my $end = shift;
3921
3922 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3923
3924 return $self->_add_delete('-', $start, $end, "");
3925 }
3926
3927 sub is_empty { # Returns boolean as to if a range list is empty
3928 my $self = shift;
3929 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3930
f998e60c