This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
RT now imports new tags automatically
[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
9be22ca5
KW
7# There was an attempt when this was first rewritten to make it 5.8
8# compatible, but that has now been abandoned, and newer constructs are used
9# as convenient.
6c68572b 10
80f41172
JH
11# NOTE: this script can run quite slowly in older/slower systems.
12# It can also consume a lot of memory (128 MB or more), you may need
13# to raise your process resource limits (e.g. in bash, "ulimit -a"
14# to inspect, and "ulimit -d ..." or "ulimit -m ..." to set)
15
cdcef19a 16my $start_time;
98dc9551 17BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
18 # get it as close as possible
19 $start_time= time;
20}
21
23e33b60 22require 5.010_001;
d73e5302 23use strict;
99870f4d 24use warnings;
cf25bb62 25use Carp;
bd9ebcfd 26use Config;
99870f4d
KW
27use File::Find;
28use File::Path;
d07a55ed 29use File::Spec;
99870f4d 30use Text::Tabs;
6b64c11c 31use re "/aa";
0dee21d6 32use feature 'state';
99870f4d
KW
33
34sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 35my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d 36
74cd47d0
KW
37sub NON_ASCII_PLATFORM { ord("A") != 65 }
38
99870f4d
KW
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
232ed87f 43# a pod file and .t files, depending on option parameters.
99870f4d
KW
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#
232ed87f
KW
55# This program works on all releases of Unicode so far. The outputs have been
56# scrutinized most intently for release 5.1. The others have been checked for
57# somewhat more than just sanity. It can handle all non-provisional Unicode
58# character properties in those releases.
99870f4d 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
232ed87f
KW
68# value". (Some more recently defined properties, map a code point to a set
69# of values.)
99870f4d
KW
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
232ed87f 122# that is algorithmically determinable from its code point (and the reverse).
99870f4d
KW
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.
232ed87f
KW
134# Again, this is so that methods can be defined on one and not the others so
135# as to prevent operating on them in incorrect ways.
99870f4d
KW
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
232ed87f 157# code points that map to N. (For each binary property, a third table is also
99870f4d 158# generated for the pseudo Perl property. It contains the identical code
232ed87f
KW
159# points as the Y table, but can be written in regular expressions, not in the
160# compound form, but in a "single" form like \p{IsUppercase}.) Many
161# properties are binary, but some properties have several possible values,
162# some have many, and properties like Name have a different value for every
163# named code point. Those will not, unless the controlling lists are changed,
164# have their match tables written out. But all the ones which can be used in
165# regular expression \p{} and \P{} constructs will. Prior to 5.14, generally
166# a property would have either its map table or its match tables written but
167# not both. Again, what gets written is controlled by lists which can easily
168# be changed. Starting in 5.14, advantage was taken of this, and all the map
169# tables needed to reconstruct the Unicode db are now written out, while
170# suppressing the Unicode .txt files that contain the data. Our tables are
171# much more compact than the .txt files, so a significant space savings was
172# achieved. Also, tables are not written out that are trivially derivable
173# from tables that do get written. So, there typically is no file containing
174# the code points not matched by a binary property (the table for \P{} versus
175# lowercase \p{}), since you just need to invert the True table to get the
176# False table.
177
178# Properties have a 'Type', like 'binary', or 'string', or 'enum' depending on
179# how many match tables there are and the content of the maps. This 'Type' is
c12f2655
KW
180# different than a range 'Type', so don't get confused by the two concepts
181# having the same name.
678f13d5 182#
99870f4d
KW
183# For information about the Unicode properties, see Unicode's UAX44 document:
184
185my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
186
187# As stated earlier, this program will work on any release of Unicode so far.
188# Most obvious problems in earlier data have NOT been corrected except when
be864b6c 189# necessary to make Perl or this program work reasonably, and to keep out
232ed87f
KW
190# potential security issues. For example, no folding information was given in
191# early releases, so this program substitutes lower case instead, just so that
192# a regular expression with the /i option will do something that actually
193# gives the right results in many cases. There are also a couple other
194# corrections for version 1.1.5, commented at the point they are made. As an
195# example of corrections that weren't made (but could be) is this statement
196# from DerivedAge.txt: "The supplementary private use code points and the
197# non-character code points were assigned in version 2.0, but not specifically
198# listed in the UCD until versions 3.0 and 3.1 respectively." (To be precise
199# it was 3.0.1 not 3.0.0) More information on Unicode version glitches is
200# further down in these introductory comments.
99870f4d 201#
232ed87f
KW
202# This program works on all non-provisional properties as of the current
203# Unicode release, though the files for some are suppressed for various
204# reasons. You can change which are output by changing lists in this program.
678f13d5 205#
dc85bd38 206# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
207# loose matchings rules (from Unicode TR18):
208#
209# The recommended names for UCD properties and property values are in
210# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
211# [PropValue]. There are both abbreviated names and longer, more
212# descriptive names. It is strongly recommended that both names be
213# recognized, and that loose matching of property names be used,
214# whereby the case distinctions, whitespace, hyphens, and underbar
215# are ignored.
232ed87f 216#
99870f4d
KW
217# The program still allows Fuzzy to override its determination of if loose
218# matching should be used, but it isn't currently used, as it is no longer
219# needed; the calculations it makes are good enough.
678f13d5 220#
99870f4d
KW
221# SUMMARY OF HOW IT WORKS:
222#
223# Process arguments
224#
225# A list is constructed containing each input file that is to be processed
226#
227# Each file on the list is processed in a loop, using the associated handler
228# code for each:
229# The PropertyAliases.txt and PropValueAliases.txt files are processed
230# first. These files name the properties and property values.
231# Objects are created of all the property and property value names
232# that the rest of the input should expect, including all synonyms.
233# The other input files give mappings from properties to property
234# values. That is, they list code points and say what the mapping
235# is under the given property. Some files give the mappings for
236# just one property; and some for many. This program goes through
232ed87f
KW
237# each file and populates the properties and their map tables from
238# them. Some properties are listed in more than one file, and
239# Unicode has set up a precedence as to which has priority if there
240# is a conflict. Thus the order of processing matters, and this
241# program handles the conflict possibility by processing the
242# overriding input files last, so that if necessary they replace
243# earlier values.
99870f4d
KW
244# After this is all done, the program creates the property mappings not
245# furnished by Unicode, but derivable from what it does give.
246# The tables of code points that match each property value in each
247# property that is accessible by regular expressions are created.
248# The Perl-defined properties are created and populated. Many of these
249# require data determined from the earlier steps
250# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 251# and Unicode are reconciled and warned about.
99870f4d
KW
252# All the properties are written to files
253# Any other files are written, and final warnings issued.
678f13d5 254#
99870f4d
KW
255# For clarity, a number of operators have been overloaded to work on tables:
256# ~ means invert (take all characters not in the set). The more
257# conventional '!' is not used because of the possibility of confusing
258# it with the actual boolean operation.
259# + means union
260# - means subtraction
261# & means intersection
262# The precedence of these is the order listed. Parentheses should be
263# copiously used. These are not a general scheme. The operations aren't
264# defined for a number of things, deliberately, to avoid getting into trouble.
265# Operations are done on references and affect the underlying structures, so
266# that the copy constructors for them have been overloaded to not return a new
267# clone, but the input object itself.
678f13d5 268#
99870f4d
KW
269# The bool operator is deliberately not overloaded to avoid confusion with
270# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
271#
272# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
273#
274# This program needs to be able to run under miniperl. Therefore, it uses a
275# minimum of other modules, and hence implements some things itself that could
276# be gotten from CPAN
277#
278# This program uses inputs published by the Unicode Consortium. These can
279# change incompatibly between releases without the Perl maintainers realizing
280# it. Therefore this program is now designed to try to flag these. It looks
281# at the directories where the inputs are, and flags any unrecognized files.
282# It keeps track of all the properties in the files it handles, and flags any
283# that it doesn't know how to handle. It also flags any input lines that
284# don't match the expected syntax, among other checks.
285#
286# It is also designed so if a new input file matches one of the known
287# templates, one hopefully just needs to add it to a list to have it
288# processed.
289#
290# As mentioned earlier, some properties are given in more than one file. In
291# particular, the files in the extracted directory are supposedly just
292# reformattings of the others. But they contain information not easily
9be22ca5
KW
293# derivable from the other files, including results for Unihan (which isn't
294# usually available to this program) and for unassigned code points. They
678f13d5
KW
295# also have historically had errors or been incomplete. In an attempt to
296# create the best possible data, this program thus processes them first to
297# glean information missing from the other files; then processes those other
298# files to override any errors in the extracted ones. Much of the design was
299# driven by this need to store things and then possibly override them.
300#
301# It tries to keep fatal errors to a minimum, to generate something usable for
302# testing purposes. It always looks for files that could be inputs, and will
303# warn about any that it doesn't know how to handle (the -q option suppresses
304# the warning).
99870f4d 305#
678f13d5
KW
306# Why is there more than one type of range?
307# This simplified things. There are some very specialized code points that
308# have to be handled specially for output, such as Hangul syllable names.
309# By creating a range type (done late in the development process), it
310# allowed this to be stored with the range, and overridden by other input.
311# Originally these were stored in another data structure, and it became a
312# mess trying to decide if a second file that was for the same property was
313# overriding the earlier one or not.
314#
315# Why are there two kinds of tables, match and map?
316# (And there is a base class shared by the two as well.) As stated above,
317# they actually are for different things. Development proceeded much more
318# smoothly when I (khw) realized the distinction. Map tables are used to
319# give the property value for every code point (actually every code point
320# that doesn't map to a default value). Match tables are used for regular
321# expression matches, and are essentially the inverse mapping. Separating
322# the two allows more specialized methods, and error checks so that one
323# can't just take the intersection of two map tables, for example, as that
324# is nonsensical.
99870f4d 325#
232ed87f
KW
326# What about 'fate' and 'status'. The concept of a table's fate was created
327# late when it became clear that something more was needed. The difference
328# between this and 'status' is unclean, and could be improved if someone
329# wanted to spend the effort.
330#
23e33b60
KW
331# DEBUGGING
332#
678f13d5
KW
333# This program is written so it will run under miniperl. Occasionally changes
334# will cause an error where the backtrace doesn't work well under miniperl.
335# To diagnose the problem, you can instead run it under regular perl, if you
336# have one compiled.
337#
338# There is a good trace facility. To enable it, first sub DEBUG must be set
339# to return true. Then a line like
340#
341# local $to_trace = 1 if main::DEBUG;
342#
232ed87f
KW
343# can be added to enable tracing in its lexical scope (plus dynamic) or until
344# you insert another line:
678f13d5
KW
345#
346# local $to_trace = 0 if main::DEBUG;
347#
232ed87f 348# To actually trace, use a line like "trace $a, @b, %c, ...;
678f13d5
KW
349#
350# Some of the more complex subroutines already have trace statements in them.
351# Permanent trace statements should be like:
352#
353# trace ... if main::DEBUG && $to_trace;
354#
fd2f7781
KW
355# main::stack_trace() will display what its name implies
356#
678f13d5
KW
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
232ed87f
KW
363# a, 'non_skip => 1,' in its constructor will be skipped. However, skipping
364# Jamo.txt or UnicodeData.txt will likely cause fatal errors.
678f13d5 365#
b4a0206c 366# To compare the output tables, it may be useful to specify the -annotate
97a8a595
KW
367# flag. (As of this writing, this can't be done on a clean workspace, due to
368# requirements in Text::Tabs used in this option; so first run mktables
369# without this option.) This option adds comment lines to each table, one for
370# each non-algorithmically named character giving, currently its code point,
371# name, and graphic representation if printable (and you have a font that
372# knows about it). This makes it easier to see what the particular code
373# points are in each output table. Non-named code points are annotated with a
374# description of their status, and contiguous ones with the same description
375# will be output as a range rather than individually. Algorithmically named
376# characters are also output as ranges, except when there are just a few
377# contiguous ones.
c4019d52 378#
99870f4d
KW
379# FUTURE ISSUES
380#
381# The program would break if Unicode were to change its names so that
382# interior white space, underscores, or dashes differences were significant
383# within property and property value names.
384#
385# It might be easier to use the xml versions of the UCD if this program ever
386# would need heavy revision, and the ability to handle old versions was not
387# required.
388#
389# There is the potential for name collisions, in that Perl has chosen names
390# that Unicode could decide it also likes. There have been such collisions in
391# the past, with mostly Perl deciding to adopt the Unicode definition of the
392# name. However in the 5.2 Unicode beta testing, there were a number of such
393# collisions, which were withdrawn before the final release, because of Perl's
394# and other's protests. These all involved new properties which began with
395# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
396# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
397# Unicode document, so they are unlikely to be used by Unicode for another
398# purpose. However, they might try something beginning with 'In', or use any
399# of the other Perl-defined properties. This program will warn you of name
400# collisions, and refuse to generate tables with them, but manual intervention
401# will be required in this event. One scheme that could be implemented, if
402# necessary, would be to have this program generate another file, or add a
403# field to mktables.lst that gives the date of first definition of a property.
404# Each new release of Unicode would use that file as a basis for the next
405# iteration. And the Perl synonym addition code could sort based on the age
406# of the property, so older properties get priority, and newer ones that clash
407# would be refused; hence existing code would not be impacted, and some other
408# synonym would have to be used for the new property. This is ugly, and
409# manual intervention would certainly be easier to do in the short run; lets
410# hope it never comes to this.
678f13d5 411#
99870f4d
KW
412# A NOTE ON UNIHAN
413#
b9e57660
KW
414# This program can generate tables from the Unihan database. But that db
415# isn't normally available, so it is marked as optional. Prior to version
416# 5.2, this database was in a single file, Unihan.txt. In 5.2 the database
417# was split into 8 different files, all beginning with the letters 'Unihan'.
418# If you plunk those files down into the directory mktables ($0) is in, this
419# program will read them and automatically create tables for the properties
420# from it that are listed in PropertyAliases.txt and PropValueAliases.txt,
421# plus any you add to the @cjk_properties array and the @cjk_property_values
422# array, being sure to add necessary '# @missings' lines to the latter. For
423# Unicode versions earlier than 5.2, most of the Unihan properties are not
424# listed at all in PropertyAliases nor PropValueAliases. This program assumes
425# for these early releases that you want the properties that are specified in
426# the 5.2 release.
99870f4d
KW
427#
428# You may need to adjust the entries to suit your purposes. setup_unihan(),
429# and filter_unihan_line() are the functions where this is done. This program
430# already does some adjusting to make the lines look more like the rest of the
431# Unicode DB; You can see what that is in filter_unihan_line()
432#
433# There is a bug in the 3.2 data file in which some values for the
434# kPrimaryNumeric property have commas and an unexpected comment. A filter
9be22ca5
KW
435# could be added to correct these; or for a particular installation, the
436# Unihan.txt file could be edited to fix them.
99870f4d 437#
678f13d5
KW
438# HOW TO ADD A FILE TO BE PROCESSED
439#
440# A new file from Unicode needs to have an object constructed for it in
441# @input_file_objects, probably at the end or at the end of the extracted
442# ones. The program should warn you if its name will clash with others on
443# restrictive file systems, like DOS. If so, figure out a better name, and
444# add lines to the README.perl file giving that. If the file is a character
232ed87f 445# property, it should be in the format that Unicode has implicitly
678f13d5
KW
446# standardized for such files for the more recently introduced ones.
447# If so, the Input_file constructor for @input_file_objects can just be the
448# file name and release it first appeared in. If not, then it should be
449# possible to construct an each_line_handler() to massage the line into the
450# standardized form.
451#
452# For non-character properties, more code will be needed. You can look at
453# the existing entries for clues.
454#
455# UNICODE VERSIONS NOTES
456#
457# The Unicode UCD has had a number of errors in it over the versions. And
458# these remain, by policy, in the standard for that version. Therefore it is
459# risky to correct them, because code may be expecting the error. So this
460# program doesn't generally make changes, unless the error breaks the Perl
461# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
462# for U+1105, which causes real problems for the algorithms for Jamo
463# calculations, so it is changed here.
464#
465# But it isn't so clear cut as to what to do about concepts that are
466# introduced in a later release; should they extend back to earlier releases
467# where the concept just didn't exist? It was easier to do this than to not,
468# so that's what was done. For example, the default value for code points not
469# in the files for various properties was probably undefined until changed by
470# some version. No_Block for blocks is such an example. This program will
471# assign No_Block even in Unicode versions that didn't have it. This has the
472# benefit that code being written doesn't have to special case earlier
473# versions; and the detriment that it doesn't match the Standard precisely for
474# the affected versions.
475#
476# Here are some observations about some of the issues in early versions:
477#
232ed87f
KW
478# Prior to version 3.0, there were 3 character decompositions. These are not
479# handled by Unicode::Normalize, nor will it compile when presented a version
480# that has them. However, you can trivially get it to compile by simply
481# ignoring those decompositions, by changing the croak to a carp. At the time
9be22ca5
KW
482# of this writing, the line (in cpan/Unicode-Normalize/Normalize.pm or
483# cpan/Unicode-Normalize/mkheader) reads
232ed87f
KW
484#
485# croak("Weird Canonical Decomposition of U+$h");
486#
28807e1d 487# Simply comment it out. It will compile, but will not know about any three
9be22ca5 488# character decompositions.
232ed87f
KW
489
490# The number of code points in \p{alpha=True} halved in 2.1.9. It turns out
491# that the reason is that the CJK block starting at 4E00 was removed from
492# PropList, and was not put back in until 3.1.0. The Perl extension (the
493# single property name \p{alpha}) has the correct values. But the compound
494# form is simply not generated until 3.1, as it can be argued that prior to
495# this release, this was not an official property. The comments for
496# filter_old_style_proplist() give more details.
678f13d5
KW
497#
498# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
499# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
500# reason is that 3.2 introduced U+205F=medium math space, which was not
501# classed as white space, but Perl figured out that it should have been. 4.0
502# reclassified it correctly.
503#
504# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
232ed87f
KW
505# this was erroneously a synonym for 202 (it should be 200). In 4.0, ATB
506# became 202, and ATBL was left with no code points, as all the ones that
507# mapped to 202 stayed mapped to 202. Thus if your program used the numeric
508# name for the class, it would not have been affected, but if it used the
509# mnemonic, it would have been.
678f13d5 510#
9be22ca5 511# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that, code
678f13d5
KW
512# points which eventually came to have this script property value, instead
513# mapped to "Unknown". But in the next release all these code points were
514# moved to \p{sc=common} instead.
9be22ca5
KW
515
516# The tests furnished by Unicode for testing WordBreak and SentenceBreak
517# generate errors in 5.0 and earlier.
99870f4d
KW
518#
519# The default for missing code points for BidiClass is complicated. Starting
520# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
521# tries to do the best it can for earlier releases. It is done in
522# process_PropertyAliases()
523#
232ed87f
KW
524# In version 2.1.2, the entry in UnicodeData.txt:
525# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;;019F;
526# should instead be
527# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;019F;;019F
528# Without this change, there are casing problems for this character.
529#
7803ad2d
KW
530# Search for $string_compare_versions to see how to compare changes to
531# properties between Unicode versions
532#
99870f4d
KW
533##############################################################################
534
535my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
536 # and errors
537my $MAX_LINE_WIDTH = 78;
538
539# Debugging aid to skip most files so as to not be distracted by them when
540# concentrating on the ones being debugged. Add
541# non_skip => 1,
542# to the constructor for those files you want processed when you set this.
543# Files with a first version number of 0 are special: they are always
c12f2655
KW
544# processed regardless of the state of this flag. Generally, Jamo.txt and
545# UnicodeData.txt must not be skipped if you want this program to not die
546# before normal completion.
99870f4d
KW
547my $debug_skip = 0;
548
e9c4b4f8
KW
549
550# Normally these are suppressed.
551my $write_Unicode_deprecated_tables = 0;
552
99870f4d
KW
553# Set to 1 to enable tracing.
554our $to_trace = 0;
555
556{ # Closure for trace: debugging aid
557 my $print_caller = 1; # ? Include calling subroutine name
558 my $main_with_colon = 'main::';
559 my $main_colon_length = length($main_with_colon);
560
561 sub trace {
562 return unless $to_trace; # Do nothing if global flag not set
563
564 my @input = @_;
565
566 local $DB::trace = 0;
567 $DB::trace = 0; # Quiet 'used only once' message
568
569 my $line_number;
570
571 # Loop looking up the stack to get the first non-trace caller
572 my $caller_line;
573 my $caller_name;
574 my $i = 0;
575 do {
576 $line_number = $caller_line;
577 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
578 $caller = $main_with_colon unless defined $caller;
579
580 $caller_name = $caller;
581
582 # get rid of pkg
583 $caller_name =~ s/.*:://;
584 if (substr($caller_name, 0, $main_colon_length)
585 eq $main_with_colon)
586 {
587 $caller_name = substr($caller_name, $main_colon_length);
588 }
589
590 } until ($caller_name ne 'trace');
591
592 # If the stack was empty, we were called from the top level
593 $caller_name = 'main' if ($caller_name eq ""
594 || $caller_name eq 'trace');
595
596 my $output = "";
9346f59a 597 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
99870f4d 598 foreach my $string (@input) {
99870f4d
KW
599 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
600 $output .= simple_dumper($string);
601 }
602 else {
603 $string = "$string" if ref $string;
604 $string = $UNDEF unless defined $string;
605 chomp $string;
606 $string = '""' if $string eq "";
607 $output .= " " if $output ne ""
608 && $string ne ""
609 && substr($output, -1, 1) ne " "
610 && substr($string, 0, 1) ne " ";
611 $output .= $string;
612 }
613 }
614
99f78760
KW
615 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
616 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
617 print STDERR $output, "\n";
618 return;
619 }
620}
621
fd2f7781
KW
622sub stack_trace() {
623 local $to_trace = 1 if main::DEBUG;
624 my $line = (caller(0))[2];
625 my $i = 1;
626
627 # Accumulate the stack trace
628 while (1) {
629 my ($pkg, $file, $caller_line, $caller) = caller $i++;
630
631 last unless defined $caller;
632
633 trace "called from $caller() at line $line";
634 $line = $caller_line;
635 }
636}
637
99870f4d
KW
638# This is for a rarely used development feature that allows you to compare two
639# versions of the Unicode standard without having to deal with changes caused
9be22ca5 640# by the code points introduced in the later version. You probably also want
8f21889c
KW
641# to use the -annotate option when using this. Run this program on a unicore
642# containing the starting release you want to compare. Save that output
643# structrue. Then, switching to a unicore with the ending release, change the
644# 0 in the $string_compare_versions definition just below to a string
645# containing a SINGLE dotted Unicode release number (e.g. "2.1") corresponding
646# to the starting release. This program will then compile, but throw away all
647# code points introduced after the starting release. Finally use a diff tool
648# to compare the two directory structures. They include only the code points
649# common to both releases, and you can see the changes caused just by the
650# underlying release semantic changes. For versions earlier than 3.2, you
651# must copy a version of DAge.txt into the directory.
c12f2655 652my $string_compare_versions = DEBUG && 0; # e.g., "2.1";
99870f4d
KW
653my $compare_versions = DEBUG
654 && $string_compare_versions
655 && pack "C*", split /\./, $string_compare_versions;
656
657sub uniques {
658 # Returns non-duplicated input values. From "Perl Best Practices:
659 # Encapsulated Cleverness". p. 455 in first edition.
660
661 my %seen;
0e407844
NC
662 # Arguably this breaks encapsulation, if the goal is to permit multiple
663 # distinct objects to stringify to the same value, and be interchangeable.
664 # However, for this program, no two objects stringify identically, and all
665 # lists passed to this function are either objects or strings. So this
666 # doesn't affect correctness, but it does give a couple of percent speedup.
667 no overloading;
99870f4d
KW
668 return grep { ! $seen{$_}++ } @_;
669}
670
671$0 = File::Spec->canonpath($0);
672
673my $make_test_script = 0; # ? Should we output a test script
6b5ab373 674my $make_norm_test_script = 0; # ? Should we output a normalization test script
99870f4d
KW
675my $write_unchanged_files = 0; # ? Should we update the output files even if
676 # we don't think they have changed
677my $use_directory = ""; # ? Should we chdir somewhere.
678my $pod_directory; # input directory to store the pod file.
679my $pod_file = 'perluniprops';
680my $t_path; # Path to the .t test file
681my $file_list = 'mktables.lst'; # File to store input and output file names.
682 # This is used to speed up the build, by not
683 # executing the main body of the program if
684 # nothing on the list has changed since the
685 # previous build
686my $make_list = 1; # ? Should we write $file_list. Set to always
687 # make a list so that when the pumpking is
688 # preparing a release, s/he won't have to do
689 # special things
690my $glob_list = 0; # ? Should we try to include unknown .txt files
691 # in the input.
bd9ebcfd
KW
692my $output_range_counts = $debugging_build; # ? Should we include the number
693 # of code points in ranges in
694 # the output
558712cf 695my $annotate = 0; # ? Should character names be in the output
9ef2b94f 696
99870f4d
KW
697# Verbosity levels; 0 is quiet
698my $NORMAL_VERBOSITY = 1;
699my $PROGRESS = 2;
700my $VERBOSE = 3;
701
702my $verbosity = $NORMAL_VERBOSITY;
703
0458fbc1
KW
704# Stored in mktables.lst so that if this program is called with different
705# options, will regenerate even if the files otherwise look like they're
706# up-to-date.
707my $command_line_arguments = join " ", @ARGV;
708
99870f4d
KW
709# Process arguments
710while (@ARGV) {
cf25bb62
JH
711 my $arg = shift @ARGV;
712 if ($arg eq '-v') {
99870f4d
KW
713 $verbosity = $VERBOSE;
714 }
715 elsif ($arg eq '-p') {
716 $verbosity = $PROGRESS;
717 $| = 1; # Flush buffers as we go.
718 }
719 elsif ($arg eq '-q') {
720 $verbosity = 0;
721 }
722 elsif ($arg eq '-w') {
723 $write_unchanged_files = 1; # update the files even if havent changed
724 }
725 elsif ($arg eq '-check') {
6ae7e459
YO
726 my $this = shift @ARGV;
727 my $ok = shift @ARGV;
728 if ($this ne $ok) {
729 print "Skipping as check params are not the same.\n";
730 exit(0);
731 }
00a8df5c 732 }
99870f4d
KW
733 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
734 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
735 }
3df51b85
KW
736 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
737 {
99870f4d 738 $make_test_script = 1;
99870f4d 739 }
6b5ab373
KW
740 elsif ($arg eq '-makenormtest')
741 {
742 $make_norm_test_script = 1;
743 }
99870f4d
KW
744 elsif ($arg eq '-makelist') {
745 $make_list = 1;
746 }
747 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
748 -d $use_directory or croak "Unknown directory '$use_directory'";
749 }
750 elsif ($arg eq '-L') {
751
752 # Existence not tested until have chdir'd
753 $file_list = shift;
754 }
755 elsif ($arg eq '-globlist') {
756 $glob_list = 1;
757 }
758 elsif ($arg eq '-c') {
759 $output_range_counts = ! $output_range_counts
760 }
b4a0206c 761 elsif ($arg eq '-annotate') {
558712cf 762 $annotate = 1;
bd9ebcfd
KW
763 $debugging_build = 1;
764 $output_range_counts = 1;
9ef2b94f 765 }
99870f4d
KW
766 else {
767 my $with_c = 'with';
768 $with_c .= 'out' if $output_range_counts; # Complements the state
769 croak <<END;
770usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
771 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
772 [-check A B ]
773 -c : Output comments $with_c number of code points in ranges
774 -q : Quiet Mode: Only output serious warnings.
775 -p : Set verbosity level to normal plus show progress.
776 -v : Set Verbosity level high: Show progress and non-serious
777 warnings
778 -w : Write files regardless
779 -C dir : Change to this directory before proceeding. All relative paths
780 except those specified by the -P and -T options will be done
781 with respect to this directory.
782 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 783 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
784 -L filelist : Use alternate 'filelist' instead of standard one
785 -globlist : Take as input all non-Test *.txt files in current and sub
786 directories
3df51b85
KW
787 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
788 overrides -T
99870f4d 789 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 790 -annotate : Output an annotation for each character in the table files;
97a8a595
KW
791 useful for debugging mktables, looking at diffs; but is slow
792 and memory intensive
99870f4d
KW
793 -check A B : Executes $0 only if A and B are the same
794END
795 }
796}
797
798# Stores the most-recently changed file. If none have changed, can skip the
799# build
aeab6150 800my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
801
802# Change directories now, because need to read 'version' early.
803if ($use_directory) {
3df51b85 804 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
805 $pod_directory = File::Spec->rel2abs($pod_directory);
806 }
3df51b85 807 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 808 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 809 }
99870f4d 810 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 811 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 812 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 813 }
3df51b85 814 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 815 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 816 }
00a8df5c
YO
817}
818
99870f4d
KW
819# Get Unicode version into regular and v-string. This is done now because
820# various tables below get populated based on it. These tables are populated
821# here to be near the top of the file, and so easily seeable by those needing
822# to modify things.
823open my $VERSION, "<", "version"
824 or croak "$0: can't open required file 'version': $!\n";
825my $string_version = <$VERSION>;
826close $VERSION;
827chomp $string_version;
828my $v_version = pack "C*", split /\./, $string_version; # v string
829
8f21889c
KW
830my $unicode_version = ($compare_versions)
831 ? ( "$string_compare_versions (using "
832 . "$string_version rules)")
833 : $string_version;
834
99870f4d
KW
835# The following are the complete names of properties with property values that
836# are known to not match any code points in some versions of Unicode, but that
837# may change in the future so they should be matchable, hence an empty file is
838# generated for them.
a9c9e371
KW
839my @tables_that_may_be_empty;
840push @tables_that_may_be_empty, 'Joining_Type=Left_Joining'
841 if $v_version lt v6.3.0;
99870f4d
KW
842push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
843push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
844push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
845 if $v_version ge v4.1.0;
82aed44a
KW
846push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
847 if $v_version ge v6.0.0;
f583b44c
KW
848push @tables_that_may_be_empty, 'Grapheme_Cluster_Break=Prepend'
849 if $v_version ge v6.1.0;
1e958ea9
KW
850push @tables_that_may_be_empty, 'Canonical_Combining_Class=CCC133'
851 if $v_version ge v6.2.0;
99870f4d
KW
852
853# The lists below are hashes, so the key is the item in the list, and the
854# value is the reason why it is in the list. This makes generation of
855# documentation easier.
856
857my %why_suppressed; # No file generated for these.
858
859# Files aren't generated for empty extraneous properties. This is arguable.
860# Extraneous properties generally come about because a property is no longer
861# used in a newer version of Unicode. If we generated a file without code
862# points, programs that used to work on that property will still execute
863# without errors. It just won't ever match (or will always match, with \P{}).
864# This means that the logic is now likely wrong. I (khw) think its better to
865# find this out by getting an error message. Just move them to the table
866# above to change this behavior
867my %why_suppress_if_empty_warn_if_not = (
868
869 # It is the only property that has ever officially been removed from the
870 # Standard. The database never contained any code points for it.
871 'Special_Case_Condition' => 'Obsolete',
872
873 # Apparently never official, but there were code points in some versions of
874 # old-style PropList.txt
875 'Non_Break' => 'Obsolete',
876);
877
878# These would normally go in the warn table just above, but they were changed
879# a long time before this program was written, so warnings about them are
880# moot.
881if ($v_version gt v3.2.0) {
882 push @tables_that_may_be_empty,
883 'Canonical_Combining_Class=Attached_Below_Left'
884}
885
9be22ca5
KW
886# Enum values for to_output_map() method in the Map_Table package. (0 is don't
887# output)
272501f6
KW
888my $EXTERNAL_MAP = 1;
889my $INTERNAL_MAP = 2;
ce712c88 890my $OUTPUT_ADJUSTED = 3;
272501f6 891
fcf1973c
KW
892# To override computed values for writing the map tables for these properties.
893# The default for enum map tables is to write them out, so that the Unicode
894# .txt files can be removed, but all the data to compute any property value
895# for any code point is available in a more compact form.
896my %global_to_output_map = (
897 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
898 # get stuck supporting it if things change. Since it is a STRING
899 # property, it normally would be listed in the pod, but INTERNAL_MAP
900 # suppresses that.
fcf1973c
KW
901 Unicode_1_Name => $INTERNAL_MAP,
902
903 Present_In => 0, # Suppress, as easily computed from Age
74cd47d0
KW
904 Block => (NON_ASCII_PLATFORM) ? 1 : 0, # Suppress, as Blocks.txt is
905 # retained, but needed for
906 # non-ASCII
53d34b6c
KW
907
908 # Suppress, as mapping can be found instead from the
909 # Perl_Decomposition_Mapping file
910 Decomposition_Type => 0,
fcf1973c
KW
911);
912
99870f4d
KW
913# There are several types of obsolete properties defined by Unicode. These
914# must be hand-edited for every new Unicode release.
915my %why_deprecated; # Generates a deprecated warning message if used.
916my %why_stabilized; # Documentation only
917my %why_obsolete; # Documentation only
918
919{ # Closure
8364f9eb 920 my $simple = 'Perl uses the more complete version';
99870f4d
KW
921 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
922
923 my $other_properties = 'other properties';
924 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 925 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
926
927 %why_deprecated = (
5f7264c7 928 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
929 'Jamo_Short_Name' => $contributory,
930 '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',
931 'Other_Alphabetic' => $contributory,
932 'Other_Default_Ignorable_Code_Point' => $contributory,
933 'Other_Grapheme_Extend' => $contributory,
934 'Other_ID_Continue' => $contributory,
935 'Other_ID_Start' => $contributory,
936 'Other_Lowercase' => $contributory,
937 'Other_Math' => $contributory,
938 'Other_Uppercase' => $contributory,
e22aaf5c
KW
939 'Expands_On_NFC' => $why_no_expand,
940 'Expands_On_NFD' => $why_no_expand,
941 'Expands_On_NFKC' => $why_no_expand,
942 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
943 );
944
945 %why_suppressed = (
5f7264c7 946 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
947 # contains the same information, but without the algorithmically
948 # determinable Hangul syllables'. This file is not published, so it's
949 # existence is not noted in the comment.
12fee290 950 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize or prop_invmap() or charprop() in Unicode::UCD::',
99870f4d 951
5f8d1a89
KW
952 # Don't suppress ISO_Comment, as otherwise special handling is needed
953 # to differentiate between it and gc=c, which can be written as 'isc',
954 # which is the same characters as ISO_Comment's short name.
99870f4d 955
12fee290 956 'Name' => "Accessible via \\N{...} or 'use charnames;' or charprop() or prop_invmap() in Unicode::UCD::",
e0b29447 957
12fee290
KW
958 'Simple_Case_Folding' => "$simple. Can access this through casefold(), charprop(), or prop_invmap() in Unicode::UCD",
959 'Simple_Lowercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
960 'Simple_Titlecase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
961 'Simple_Uppercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
99870f4d 962
dac6f618 963 FC_NFKC_Closure => 'Deprecated by Unicode, and supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
964 );
965
1704a0ea
KW
966 foreach my $property (
967
968 # The following are suppressed because they were made contributory
969 # or deprecated by Unicode before Perl ever thought about
970 # supporting them.
971 'Jamo_Short_Name',
972 'Grapheme_Link',
973 'Expands_On_NFC',
974 'Expands_On_NFD',
975 'Expands_On_NFKC',
976 'Expands_On_NFKD',
977
978 # The following are suppressed because they have been marked
979 # as deprecated for a sufficient amount of time
980 'Other_Alphabetic',
981 'Other_Default_Ignorable_Code_Point',
982 'Other_Grapheme_Extend',
983 'Other_ID_Continue',
984 'Other_ID_Start',
985 'Other_Lowercase',
986 'Other_Math',
987 'Other_Uppercase',
e22aaf5c 988 ) {
99870f4d
KW
989 $why_suppressed{$property} = $why_deprecated{$property};
990 }
cf25bb62 991
99870f4d
KW
992 # Customize the message for all the 'Other_' properties
993 foreach my $property (keys %why_deprecated) {
994 next if (my $main_property = $property) !~ s/^Other_//;
995 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
996 }
997}
998
e9c4b4f8
KW
999if ($write_Unicode_deprecated_tables) {
1000 foreach my $property (keys %why_suppressed) {
1001 delete $why_suppressed{$property} if $property =~
1002 / ^ Other | Grapheme /x;
1003 }
1004}
1005
99870f4d
KW
1006if ($v_version ge 4.0.0) {
1007 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
1008 if ($v_version ge 6.0.0) {
1009 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
1010 }
99870f4d 1011}
5f7264c7 1012if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 1013 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 1014 if ($v_version ge 6.0.0) {
63f74647 1015 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 1016 }
99870f4d
KW
1017}
1018
1019# Probably obsolete forever
1020if ($v_version ge v4.1.0) {
82aed44a
KW
1021 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
1022}
1023if ($v_version ge v6.0.0) {
caa75395 1024 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana" (or both)';
2b352efd 1025 $why_suppressed{'Script_Extensions=Katakana_Or_Hiragana'} = 'All code points that would be matched by this are matched by either "Script_Extensions=Katakana" or "Script_Extensions=Hiragana"';
99870f4d
KW
1026}
1027
1028# This program can create files for enumerated-like properties, such as
1029# 'Numeric_Type'. This file would be the same format as for a string
1030# property, with a mapping from code point to its value, so you could look up,
1031# for example, the script a code point is in. But no one so far wants this
1032# mapping, or they have found another way to get it since this is a new
1033# feature. So no file is generated except if it is in this list.
1034my @output_mapped_properties = split "\n", <<END;
1035END
1036
b9e57660
KW
1037# If you want more Unihan properties than the default, you need to add them to
1038# these arrays. Depending on the property type, @missing lines might have to
1039# be added to the second array. A sample entry would be (including the '#'):
1040# @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
99870f4d 1041my @cjk_properties = split "\n", <<'END';
99870f4d 1042END
99870f4d 1043my @cjk_property_values = split "\n", <<'END';
99870f4d
KW
1044END
1045
1046# The input files don't list every code point. Those not listed are to be
1047# defaulted to some value. Below are hard-coded what those values are for
1048# non-binary properties as of 5.1. Starting in 5.0, there are
caa75395 1049# machine-parsable comment lines in the files that give the defaults; so this
99870f4d
KW
1050# list shouldn't have to be extended. The claim is that all missing entries
1051# for binary properties will default to 'N'. Unicode tried to change that in
1052# 5.2, but the beta period produced enough protest that they backed off.
1053#
1054# The defaults for the fields that appear in UnicodeData.txt in this hash must
1055# be in the form that it expects. The others may be synonyms.
1056my $CODE_POINT = '<code point>';
1057my %default_mapping = (
1058 Age => "Unassigned",
1059 # Bidi_Class => Complicated; set in code
1060 Bidi_Mirroring_Glyph => "",
1061 Block => 'No_Block',
1062 Canonical_Combining_Class => 0,
1063 Case_Folding => $CODE_POINT,
1064 Decomposition_Mapping => $CODE_POINT,
1065 Decomposition_Type => 'None',
1066 East_Asian_Width => "Neutral",
1067 FC_NFKC_Closure => $CODE_POINT,
ec2cd995 1068 General_Category => ($v_version le 6.3.0) ? 'Cn' : 'Unassigned',
99870f4d
KW
1069 Grapheme_Cluster_Break => 'Other',
1070 Hangul_Syllable_Type => 'NA',
1071 ISO_Comment => "",
1072 Jamo_Short_Name => "",
1073 Joining_Group => "No_Joining_Group",
1074 # Joining_Type => Complicated; set in code
1075 kIICore => 'N', # Is converted to binary
1076 #Line_Break => Complicated; set in code
1077 Lowercase_Mapping => $CODE_POINT,
1078 Name => "",
1079 Name_Alias => "",
1080 NFC_QC => 'Yes',
1081 NFD_QC => 'Yes',
1082 NFKC_QC => 'Yes',
1083 NFKD_QC => 'Yes',
1084 Numeric_Type => 'None',
1085 Numeric_Value => 'NaN',
1086 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1087 Sentence_Break => 'Other',
1088 Simple_Case_Folding => $CODE_POINT,
1089 Simple_Lowercase_Mapping => $CODE_POINT,
1090 Simple_Titlecase_Mapping => $CODE_POINT,
1091 Simple_Uppercase_Mapping => $CODE_POINT,
1092 Titlecase_Mapping => $CODE_POINT,
1093 Unicode_1_Name => "",
1094 Unicode_Radical_Stroke => "",
1095 Uppercase_Mapping => $CODE_POINT,
1096 Word_Break => 'Other',
1097);
1098
678f13d5 1099### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1100
1101my $HEADER=<<"EOF";
1102# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85 1103# This file is machine-generated by $0 from the Unicode
8f21889c 1104# database, Version $unicode_version. Any changes made here will be lost!
cf25bb62
JH
1105EOF
1106
126c3d4e 1107my $INTERNAL_ONLY_HEADER = <<"EOF";
99870f4d
KW
1108
1109# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1110# This file is for internal use by core Perl only. The format and even the
1111# name or existence of this file are subject to change without notice. Don't
d9ae3878
KW
1112# use it directly. Use Unicode::UCD to access the Unicode character data
1113# base.
99870f4d
KW
1114EOF
1115
1116my $DEVELOPMENT_ONLY=<<"EOF";
1117# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1118# This file contains information artificially constrained to code points
1119# present in Unicode release $string_compare_versions.
1120# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1121# not be used for production.
b6922eda
KW
1122
1123EOF
1124
346a20cf
KW
1125my $MAX_UNICODE_CODEPOINT_STRING = ($v_version ge v2.0.0)
1126 ? "10FFFF"
1127 : "FFFF";
6189eadc
KW
1128my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1129my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d 1130
2d88a86a
KW
1131# We work with above-Unicode code points, up to UV_MAX. But when you get
1132# that high, above IV_MAX, some operations don't work, and you can easily get
1133# overflow. Therefore for internal use, we use a much smaller number,
1134# translating it to UV_MAX only for output. The exact number is immaterial
1135# (all Unicode code points are treated exactly the same), but the algorithm
1136# requires it to be at least 2 * $MAX_UNICODE_CODEPOINTS + 1;
1137my $MAX_WORKING_CODEPOINTS= $MAX_UNICODE_CODEPOINT * 8;
1138my $MAX_WORKING_CODEPOINT = $MAX_WORKING_CODEPOINTS - 1;
1139my $MAX_WORKING_CODEPOINT_STRING = sprintf("%X", $MAX_WORKING_CODEPOINT);
1140
1141my $MAX_PLATFORM_CODEPOINT = ~0;
1142
99870f4d
KW
1143# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1144# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1145# to decrease backtracking. The first regex allows the code point to be at
1146# the end of a word, but to work properly, the word shouldn't end with a valid
1147# hex character. The second one won't match a code point at the end of a
1148# word, and doesn't have the run-on issue
8c32d378
KW
1149my $run_on_code_point_re =
1150 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1151my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d
KW
1152
1153# This matches the beginning of the line in the Unicode db files that give the
1154# defaults for code points not listed (i.e., missing) in the file. The code
1155# depends on this ending with a semi-colon, so it can assume it is a valid
1156# field when the line is split() by semi-colons
346a20cf 1157my $missing_defaults_prefix = qr/^#\s+\@missing:\s+0000\.\.10FFFF\s*;/;
99870f4d
KW
1158
1159# Property types. Unicode has more types, but these are sufficient for our
1160# purposes.
1161my $UNKNOWN = -1; # initialized to illegal value
1162my $NON_STRING = 1; # Either binary or enum
1163my $BINARY = 2;
06f26c45
KW
1164my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1165 # tables, additional true and false tables are
1166 # generated so that false is anything matching the
1167 # default value, and true is everything else.
1168my $ENUM = 4; # Include catalog
1169my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1170
1171# Some input files have lines that give default values for code points not
1172# contained in the file. Sometimes these should be ignored.
1173my $NO_DEFAULTS = 0; # Must evaluate to false
1174my $NOT_IGNORED = 1;
1175my $IGNORED = 2;
1176
1177# Range types. Each range has a type. Most ranges are type 0, for normal,
1178# and will appear in the main body of the tables in the output files, but
1179# there are other types of ranges as well, listed below, that are specially
1180# handled. There are pseudo-types as well that will never be stored as a
1181# type, but will affect the calculation of the type.
1182
1183# 0 is for normal, non-specials
1184my $MULTI_CP = 1; # Sequence of more than code point
1185my $HANGUL_SYLLABLE = 2;
1186my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1187my $NULL = 4; # The map is to the null string; utf8.c can't
1188 # handle these, nor is there an accepted syntax
1189 # for them in \p{} constructs
f86864ac 1190my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1191 # otherwise be $MULTI_CP type are instead type 0
1192
1193# process_generic_property_file() can accept certain overrides in its input.
1194# Each of these must begin AND end with $CMD_DELIM.
1195my $CMD_DELIM = "\a";
1196my $REPLACE_CMD = 'replace'; # Override the Replace
1197my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1198
1199my $NO = 0;
1200my $YES = 1;
1201
1202# Values for the Replace argument to add_range.
1203# $NO # Don't replace; add only the code points not
1204 # already present.
1205my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1206 # the comments at the subroutine definition.
1207my $UNCONDITIONALLY = 2; # Replace without conditions.
9470941f 1208my $MULTIPLE_BEFORE = 4; # Don't replace, but add a duplicate record if
99870f4d 1209 # already there
7f4b1e25
KW
1210my $MULTIPLE_AFTER = 5; # Don't replace, but add a duplicate record if
1211 # already there
1212my $CROAK = 6; # Die with an error if is already there
99870f4d
KW
1213
1214# Flags to give property statuses. The phrases are to remind maintainers that
1215# if the flag is changed, the indefinite article referring to it in the
1216# documentation may need to be as well.
1217my $NORMAL = "";
99870f4d
KW
1218my $DEPRECATED = 'D';
1219my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1220my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1221my $DISCOURAGED = 'X';
1222my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1223my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1224my $STRICTER = 'T';
1225my $a_bold_stricter = "a 'B<$STRICTER>'";
1226my $A_bold_stricter = "A 'B<$STRICTER>'";
1227my $STABILIZED = 'S';
1228my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1229my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1230my $OBSOLETE = 'O';
1231my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1232my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1233
5c6d9947
KW
1234# Aliases can also have an extra status:
1235my $INTERNAL_ALIAS = 'P';
1236
99870f4d
KW
1237my %status_past_participles = (
1238 $DISCOURAGED => 'discouraged',
99870f4d
KW
1239 $STABILIZED => 'stabilized',
1240 $OBSOLETE => 'obsolete',
37e2e78e 1241 $DEPRECATED => 'deprecated',
5c6d9947 1242 $INTERNAL_ALIAS => 'reserved for Perl core internal use only',
99870f4d
KW
1243);
1244
395dfc19
KW
1245# Table fates. These are somewhat ordered, so that fates < $MAP_PROXIED should be
1246# externally documented.
301ba948 1247my $ORDINARY = 0; # The normal fate.
395dfc19
KW
1248my $MAP_PROXIED = 1; # The map table for the property isn't written out,
1249 # but there is a file written that can be used to
1250 # reconstruct this table
3cdaf629 1251my $INTERNAL_ONLY = 2; # The file for this table is written out, but it is
301ba948 1252 # for Perl's internal use only
277b7b16
KW
1253my $LEGACY_ONLY = 3; # Like $INTERNAL_ONLY, but not actually used by Perl.
1254 # Is for backwards compatibility for applications that
1255 # read the file directly, so it's format is
1256 # unchangeable.
1257my $SUPPRESSED = 4; # The file for this table is not written out, and as a
3cdaf629
KW
1258 # result, we don't bother to do many computations on
1259 # it.
277b7b16 1260my $PLACEHOLDER = 5; # Like $SUPPRESSED, but we go through all the
3cdaf629
KW
1261 # computations anyway, as the values are needed for
1262 # things to work. This happens when we have Perl
1263 # extensions that depend on Unicode tables that
1264 # wouldn't normally be in a given Unicode version.
301ba948 1265
f5817e0a
KW
1266# The format of the values of the tables:
1267my $EMPTY_FORMAT = "";
99870f4d
KW
1268my $BINARY_FORMAT = 'b';
1269my $DECIMAL_FORMAT = 'd';
1270my $FLOAT_FORMAT = 'f';
1271my $INTEGER_FORMAT = 'i';
1272my $HEX_FORMAT = 'x';
1273my $RATIONAL_FORMAT = 'r';
1274my $STRING_FORMAT = 's';
d11155ec 1275my $ADJUST_FORMAT = 'a';
24303724 1276my $HEX_ADJUST_FORMAT = 'ax';
a14f3cb1 1277my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1278my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1279
1280my %map_table_formats = (
1281 $BINARY_FORMAT => 'binary',
1282 $DECIMAL_FORMAT => 'single decimal digit',
1283 $FLOAT_FORMAT => 'floating point number',
1284 $INTEGER_FORMAT => 'integer',
add63c13 1285 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1286 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1287 $STRING_FORMAT => 'string',
d11155ec 1288 $ADJUST_FORMAT => 'some entries need adjustment',
24303724 1289 $HEX_ADJUST_FORMAT => 'mapped value in hex; some entries need adjustment',
92f9d56c 1290 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1291 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1292);
1293
1294# Unicode didn't put such derived files in a separate directory at first.
1295my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1296my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1297my $AUXILIARY = 'auxiliary';
1298
3854b4b8
KW
1299# Hashes and arrays that will eventually go into Heavy.pl for the use of
1300# utf8_heavy.pl and into UCD.pl for the use of UCD.pm
99870f4d
KW
1301my %loose_to_file_of; # loosely maps table names to their respective
1302 # files
1303my %stricter_to_file_of; # same; but for stricter mapping.
315bfd4e 1304my %loose_property_to_file_of; # Maps a loose property name to its map file
e1dc048d 1305my %strict_property_to_file_of; # Same, but strict
3854b4b8
KW
1306my @inline_definitions = "V0"; # Each element gives a definition of a unique
1307 # inversion list. When a definition is inlined,
1308 # its value in the hash it's in (one of the two
1309 # defined just above) will include an index into
1310 # this array. The 0th element is initialized to
9be22ca5 1311 # the definition for a zero length inversion list
89cf10cc
KW
1312my %file_to_swash_name; # Maps the file name to its corresponding key name
1313 # in the hash %utf8::SwashInfo
99870f4d
KW
1314my %nv_floating_to_rational; # maps numeric values floating point numbers to
1315 # their rational equivalent
c12f2655
KW
1316my %loose_property_name_of; # Loosely maps (non_string) property names to
1317 # standard form
e1dc048d
KW
1318my %strict_property_name_of; # Strictly maps (non_string) property names to
1319 # standard form
86a52d1e 1320my %string_property_loose_to_name; # Same, for string properties.
c15fda25
KW
1321my %loose_defaults; # keys are of form "prop=value", where 'prop' is
1322 # the property name in standard loose form, and
1323 # 'value' is the default value for that property,
1324 # also in standard loose form.
9e4a1e86
KW
1325my %loose_to_standard_value; # loosely maps table names to the canonical
1326 # alias for them
2df7880f
KW
1327my %ambiguous_names; # keys are alias names (in standard form) that
1328 # have more than one possible meaning.
d963b40d
KW
1329my %combination_property; # keys are alias names (in standard form) that
1330 # have both a map table, and a binary one that
1331 # yields true for all non-null maps.
5d1df013
KW
1332my %prop_aliases; # Keys are standard property name; values are each
1333 # one's aliases
1e863613
KW
1334my %prop_value_aliases; # Keys of top level are standard property name;
1335 # values are keys to another hash, Each one is
1336 # one of the property's values, in standard form.
1337 # The values are that prop-val's aliases.
b9e57660 1338my %skipped_files; # List of files that we skip
2df7880f 1339my %ucd_pod; # Holds entries that will go into the UCD section of the pod
99870f4d 1340
d867ccfb
KW
1341# Most properties are immune to caseless matching, otherwise you would get
1342# nonsensical results, as properties are a function of a code point, not
1343# everything that is caselessly equivalent to that code point. For example,
1344# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1345# be true because 's' and 'S' are equivalent caselessly. However,
1346# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1347# extend that concept to those very few properties that are like this. Each
1348# such property will match the full range caselessly. They are hard-coded in
1349# the program; it's not worth trying to make it general as it's extremely
1350# unlikely that they will ever change.
1351my %caseless_equivalent_to;
1352
53adf6a2
KW
1353# This is the range of characters that were in Release 1 of Unicode, and
1354# removed in Release 2 (replaced with the current Hangul syllables starting at
1355# U+AC00). The range was reused starting in Release 3 for other purposes.
1356my $FIRST_REMOVED_HANGUL_SYLLABLE = 0x3400;
1357my $FINAL_REMOVED_HANGUL_SYLLABLE = 0x4DFF;
1358
99870f4d
KW
1359# These constants names and values were taken from the Unicode standard,
1360# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1361# syllables. The '_string' versions are so generated tables can retain the
1362# hex format, which is the more familiar value
1363my $SBase_string = "0xAC00";
1364my $SBase = CORE::hex $SBase_string;
1365my $LBase_string = "0x1100";
1366my $LBase = CORE::hex $LBase_string;
1367my $VBase_string = "0x1161";
1368my $VBase = CORE::hex $VBase_string;
1369my $TBase_string = "0x11A7";
1370my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1371my $SCount = 11172;
1372my $LCount = 19;
1373my $VCount = 21;
1374my $TCount = 28;
1375my $NCount = $VCount * $TCount;
1376
1377# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1378# with the above published constants.
1379my %Jamo;
1380my %Jamo_L; # Leading consonants
1381my %Jamo_V; # Vowels
1382my %Jamo_T; # Trailing consonants
1383
bb1dd3da
KW
1384# For code points whose name contains its ordinal as a '-ABCD' suffix.
1385# The key is the base name of the code point, and the value is an
1386# array giving all the ranges that use this base name. Each range
1387# is actually a hash giving the 'low' and 'high' values of it.
1388my %names_ending_in_code_point;
1389my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1390 # removed from the names
1391# Inverse mapping. The list of ranges that have these kinds of
1392# names. Each element contains the low, high, and base names in an
1393# anonymous hash.
1394my @code_points_ending_in_code_point;
1395
6b5ab373
KW
1396# To hold Unicode's normalization test suite
1397my @normalization_tests;
1398
bb1dd3da
KW
1399# Boolean: does this Unicode version have the hangul syllables, and are we
1400# writing out a table for them?
1401my $has_hangul_syllables = 0;
1402
1403# Does this Unicode version have code points whose names end in their
1404# respective code points, and are we writing out a table for them? 0 for no;
1405# otherwise points to first property that a table is needed for them, so that
1406# if multiple tables are needed, we don't create duplicates
1407my $needing_code_points_ending_in_code_point = 0;
1408
37e2e78e 1409my @backslash_X_tests; # List of tests read in for testing \X
6b659339 1410my @LB_tests; # List of tests read in for testing \b{lb}
06ae2722 1411my @SB_tests; # List of tests read in for testing \b{sb}
ae3bb8ea 1412my @WB_tests; # List of tests read in for testing \b{wb}
99870f4d
KW
1413my @unhandled_properties; # Will contain a list of properties found in
1414 # the input that we didn't process.
f86864ac 1415my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1416 # listed in the pod
1417my @map_properties; # Properties that get map files written
1418my @named_sequences; # NamedSequences.txt contents.
1419my %potential_files; # Generated list of all .txt files in the directory
1420 # structure so we can warn if something is being
1421 # ignored.
a0204bb1
KW
1422my @missing_early_files; # Generated list of absent files that we need to
1423 # proceed in compiling this early Unicode version
99870f4d
KW
1424my @files_actually_output; # List of files we generated.
1425my @more_Names; # Some code point names are compound; this is used
1426 # to store the extra components of them.
1427my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1428 # the minimum before we consider it equivalent to a
1429 # candidate rational
1430my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1431
1432# These store references to certain commonly used property objects
1254636b 1433my $age;
3c88a801 1434my $ccc;
99870f4d
KW
1435my $gc;
1436my $perl;
1437my $block;
3e20195b
KW
1438my $perl_charname;
1439my $print;
ac7dbdb6 1440my $All;
bc0c431d 1441my $Assigned; # All assigned characters in this Unicode release
e47e66b9
KW
1442my $DI; # Default_Ignorable_Code_Point property
1443my $NChar; # Noncharacter_Code_Point property
359523e2 1444my $script;
99870f4d
KW
1445
1446# Are there conflicting names because of beginning with 'In_', or 'Is_'
1447my $has_In_conflicts = 0;
1448my $has_Is_conflicts = 0;
1449
1450sub internal_file_to_platform ($) {
1451 # Convert our file paths which have '/' separators to those of the
1452 # platform.
1453
1454 my $file = shift;
1455 return undef unless defined $file;
1456
1457 return File::Spec->join(split '/', $file);
d07a55ed 1458}
5beb625e 1459
99870f4d
KW
1460sub file_exists ($) { # platform independent '-e'. This program internally
1461 # uses slash as a path separator.
1462 my $file = shift;
1463 return 0 if ! defined $file;
1464 return -e internal_file_to_platform($file);
1465}
5beb625e 1466
99870f4d 1467sub objaddr($) {
23e33b60
KW
1468 # Returns the address of the blessed input object.
1469 # It doesn't check for blessedness because that would do a string eval
1470 # every call, and the program is structured so that this is never called
1471 # for a non-blessed object.
99870f4d 1472
23e33b60 1473 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1474
1475 # Numifying a ref gives its address.
051df77b 1476 return pack 'J', $_[0];
99870f4d
KW
1477}
1478
558712cf 1479# These are used only if $annotate is true.
c4019d52
KW
1480# The entire range of Unicode characters is examined to populate these
1481# after all the input has been processed. But most can be skipped, as they
1482# have the same descriptive phrases, such as being unassigned
1483my @viacode; # Contains the 1 million character names
4b06b8de 1484my @age; # And their ages ("" if none)
c4019d52
KW
1485my @printable; # boolean: And are those characters printable?
1486my @annotate_char_type; # Contains a type of those characters, specifically
1487 # for the purposes of annotation.
1488my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1489 # name for the purposes of annotation. They map to the
c4019d52
KW
1490 # upper edge of the range, so that the end point can
1491 # be immediately found. This is used to skip ahead to
1492 # the end of a range, and avoid processing each
1493 # individual code point in it.
1494my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1495 # characters, but excluding those which are
1496 # also noncharacter code points
1497
1498# The annotation types are an extension of the regular range types, though
1499# some of the latter are folded into one. Make the new types negative to
1500# avoid conflicting with the regular types
1501my $SURROGATE_TYPE = -1;
1502my $UNASSIGNED_TYPE = -2;
1503my $PRIVATE_USE_TYPE = -3;
1504my $NONCHARACTER_TYPE = -4;
1505my $CONTROL_TYPE = -5;
2d88a86a
KW
1506my $ABOVE_UNICODE_TYPE = -6;
1507my $UNKNOWN_TYPE = -7; # Used only if there is a bug in this program
c4019d52
KW
1508
1509sub populate_char_info ($) {
558712cf 1510 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1511 # input code point's info that are needed for outputting more detailed
1512 # comments. If calling context wants a return, it is the end point of
1513 # any contiguous range of characters that share essentially the same info
1514
1515 my $i = shift;
1516 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1517
1518 $viacode[$i] = $perl_charname->value_of($i) || "";
1254636b
KW
1519 $age[$i] = (defined $age)
1520 ? (($age->value_of($i) =~ / ^ \d \. \d $ /x)
1521 ? $age->value_of($i)
1522 : "")
1523 : "";
c4019d52
KW
1524
1525 # A character is generally printable if Unicode says it is,
1526 # but below we make sure that most Unicode general category 'C' types
1527 # aren't.
1528 $printable[$i] = $print->contains($i);
1529
1254636b
KW
1530 # But the characters in this range were removed in v2.0 and replaced by
1531 # different ones later. Modern fonts will be for the replacement
1532 # characters, so suppress printing them.
1533 if (($v_version lt v2.0
1534 || ($compare_versions && $compare_versions lt v2.0))
1535 && ( $i >= $FIRST_REMOVED_HANGUL_SYLLABLE
1536 && $i <= $FINAL_REMOVED_HANGUL_SYLLABLE))
1537 {
1538 $printable[$i] = 0;
1539 }
1540
c4019d52
KW
1541 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1542
1543 # Only these two regular types are treated specially for annotations
1544 # purposes
1545 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1546 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1547
1548 # Give a generic name to all code points that don't have a real name.
1549 # We output ranges, if applicable, for these. Also calculate the end
1550 # point of the range.
1551 my $end;
1552 if (! $viacode[$i]) {
2d88a86a
KW
1553 if ($i > $MAX_UNICODE_CODEPOINT) {
1554 $viacode[$i] = 'Above-Unicode';
1555 $annotate_char_type[$i] = $ABOVE_UNICODE_TYPE;
1556 $printable[$i] = 0;
1557 $end = $MAX_WORKING_CODEPOINT;
1558 }
1559 elsif ($gc-> table('Private_use')->contains($i)) {
c4019d52
KW
1560 $viacode[$i] = 'Private Use';
1561 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1562 $printable[$i] = 0;
1563 $end = $gc->table('Private_Use')->containing_range($i)->end;
1564 }
e47e66b9 1565 elsif ($NChar->contains($i)) {
c4019d52
KW
1566 $viacode[$i] = 'Noncharacter';
1567 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1568 $printable[$i] = 0;
e47e66b9 1569 $end = $NChar->containing_range($i)->end;
c4019d52
KW
1570 }
1571 elsif ($gc-> table('Control')->contains($i)) {
1254636b
KW
1572 my $name_ref = property_ref('Name_Alias');
1573 $name_ref = property_ref('Unicode_1_Name') if ! defined $name_ref;
1574 $viacode[$i] = (defined $name_ref)
1575 ? $name_ref->value_of($i)
1576 : 'Control';
c4019d52
KW
1577 $annotate_char_type[$i] = $CONTROL_TYPE;
1578 $printable[$i] = 0;
c4019d52
KW
1579 }
1580 elsif ($gc-> table('Unassigned')->contains($i)) {
c4019d52
KW
1581 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1582 $printable[$i] = 0;
1254636b
KW
1583 $viacode[$i] = 'Unassigned';
1584
349edad4 1585 if (defined $block) { # No blocks in earliest releases
1254636b 1586 $viacode[$i] .= ', block=' . $block-> value_of($i);
1d025d66 1587 $end = $gc-> table('Unassigned')->containing_range($i)->end;
c4019d52 1588
bf06c733
KW
1589 # Because we name the unassigned by the blocks they are in, it
1590 # can't go past the end of that block, and it also can't go
1591 # past the unassigned range it is in. The special table makes
1592 # sure that the non-characters, which are unassigned, are
1593 # separated out.
1594 $end = min($block->containing_range($i)->end,
1595 $unassigned_sans_noncharacters->
1596 containing_range($i)->end);
1d025d66 1597 }
1254636b
KW
1598 else {
1599 $end = $i + 1;
1600 while ($unassigned_sans_noncharacters->contains($end)) {
1601 $end++;
1602 }
1603 $end--;
1604 }
1d025d66 1605 }
35e2b416 1606 elsif ($perl->table('_Perl_Surrogate')->contains($i)) {
1d025d66
KW
1607 $viacode[$i] = 'Surrogate';
1608 $annotate_char_type[$i] = $SURROGATE_TYPE;
1609 $printable[$i] = 0;
1610 $end = $gc->table('Surrogate')->containing_range($i)->end;
13ca76ff
KW
1611 }
1612 else {
1613 Carp::my_carp_bug("Can't figure out how to annotate "
1614 . sprintf("U+%04X", $i)
1615 . ". Proceeding anyway.");
c4019d52
KW
1616 $viacode[$i] = 'UNKNOWN';
1617 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1618 $printable[$i] = 0;
1619 }
1620 }
1621
1622 # Here, has a name, but if it's one in which the code point number is
1623 # appended to the name, do that.
1624 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1625 $viacode[$i] .= sprintf("-%04X", $i);
4b06b8de 1626
4b06b8de 1627 my $limit = $perl_charname->containing_range($i)->end;
1254636b
KW
1628 if (defined $age) {
1629 # Do all these as groups of the same age, instead of individually,
1630 # because their names are so meaningless, and there are typically
1631 # large quantities of them.
1632 $end = $i + 1;
1633 while ($end <= $limit && $age->value_of($end) == $age[$i]) {
1634 $end++;
1635 }
1636 $end--;
1637 }
1638 else {
1639 $end = $limit;
4b06b8de 1640 }
c4019d52
KW
1641 }
1642
1643 # And here, has a name, but if it's a hangul syllable one, replace it with
1644 # the correct name from the Unicode algorithm
1645 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1646 use integer;
1647 my $SIndex = $i - $SBase;
1648 my $L = $LBase + $SIndex / $NCount;
1649 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1650 my $T = $TBase + $SIndex % $TCount;
1651 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1652 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1653 $end = $perl_charname->containing_range($i)->end;
1654 }
1655
1656 return if ! defined wantarray;
1657 return $i if ! defined $end; # If not a range, return the input
1658
1659 # Save this whole range so can find the end point quickly
1660 $annotate_ranges->add_map($i, $end, $end);
1661
1662 return $end;
1663}
1664
23e33b60
KW
1665# Commented code below should work on Perl 5.8.
1666## This 'require' doesn't necessarily work in miniperl, and even if it does,
1667## the native perl version of it (which is what would operate under miniperl)
1668## is extremely slow, as it does a string eval every call.
7e017d6d 1669#my $has_fast_scalar_util = $^X !~ /miniperl/
23e33b60
KW
1670# && defined eval "require Scalar::Util";
1671#
1672#sub objaddr($) {
1673# # Returns the address of the blessed input object. Uses the XS version if
1674# # available. It doesn't check for blessedness because that would do a
1675# # string eval every call, and the program is structured so that this is
1676# # never called for a non-blessed object.
1677#
1678# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1679#
1680# # Check at least that is a ref.
1681# my $pkg = ref($_[0]) or return undef;
1682#
1683# # Change to a fake package to defeat any overloaded stringify
1684# bless $_[0], 'main::Fake';
1685#
1686# # Numifying a ref gives its address.
051df77b 1687# my $addr = pack 'J', $_[0];
23e33b60
KW
1688#
1689# # Return to original class
1690# bless $_[0], $pkg;
1691# return $addr;
1692#}
1693
99870f4d
KW
1694sub max ($$) {
1695 my $a = shift;
1696 my $b = shift;
1697 return $a if $a >= $b;
1698 return $b;
1699}
1700
1701sub min ($$) {
1702 my $a = shift;
1703 my $b = shift;
1704 return $a if $a <= $b;
1705 return $b;
1706}
1707
1708sub clarify_number ($) {
1709 # This returns the input number with underscores inserted every 3 digits
1710 # in large (5 digits or more) numbers. Input must be entirely digits, not
1711 # checked.
1712
1713 my $number = shift;
1714 my $pos = length($number) - 3;
1715 return $number if $pos <= 1;
1716 while ($pos > 0) {
1717 substr($number, $pos, 0) = '_';
1718 $pos -= 3;
5beb625e 1719 }
99870f4d 1720 return $number;
99598c8c
JH
1721}
1722
731cb813
KW
1723sub clarify_code_point_count ($) {
1724 # This is like clarify_number(), but the input is assumed to be a count of
1725 # code points, rather than a generic number.
1726
2d88a86a
KW
1727 my $append = "";
1728
1729 my $number = shift;
1730 if ($number > $MAX_UNICODE_CODEPOINTS) {
1731 $number -= ($MAX_WORKING_CODEPOINTS - $MAX_UNICODE_CODEPOINTS);
1732 return "All above-Unicode code points" if $number == 0;
1733 $append = " + all above-Unicode code points";
1734 }
1735 return clarify_number($number) . $append;
731cb813 1736}
12ac2576 1737
99870f4d 1738package Carp;
7ebf06b3 1739
99870f4d
KW
1740# These routines give a uniform treatment of messages in this program. They
1741# are placed in the Carp package to cause the stack trace to not include them,
1742# although an alternative would be to use another package and set @CARP_NOT
1743# for it.
12ac2576 1744
99870f4d 1745our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1746
99f78760
KW
1747# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1748# and overload trying to load Scalar:Util under miniperl. See
1749# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1750undef $overload::VERSION;
1751
99870f4d
KW
1752sub my_carp {
1753 my $message = shift || "";
1754 my $nofold = shift || 0;
7ebf06b3 1755
99870f4d
KW
1756 if ($message) {
1757 $message = main::join_lines($message);
1758 $message =~ s/^$0: *//; # Remove initial program name
1759 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1760 $message = "\n$0: $message;";
12ac2576 1761
99870f4d
KW
1762 # Fold the message with program name, semi-colon end punctuation
1763 # (which looks good with the message that carp appends to it), and a
1764 # hanging indent for continuation lines.
1765 $message = main::simple_fold($message, "", 4) unless $nofold;
1766 $message =~ s/\n$//; # Remove the trailing nl so what carp
1767 # appends is to the same line
1768 }
12ac2576 1769
99870f4d 1770 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1771
99870f4d
KW
1772 carp $message;
1773 return;
1774}
7ebf06b3 1775
99870f4d
KW
1776sub my_carp_bug {
1777 # This is called when it is clear that the problem is caused by a bug in
1778 # this program.
7ebf06b3 1779
99870f4d
KW
1780 my $message = shift;
1781 $message =~ s/^$0: *//;
1782 $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");
1783 carp $message;
1784 return;
1785}
7ebf06b3 1786
99870f4d
KW
1787sub carp_too_few_args {
1788 if (@_ != 2) {
1789 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1790 return;
12ac2576 1791 }
7ebf06b3 1792
99870f4d
KW
1793 my $args_ref = shift;
1794 my $count = shift;
7ebf06b3 1795
99870f4d
KW
1796 my_carp_bug("Need at least $count arguments to "
1797 . (caller 1)[3]
1798 . ". Instead got: '"
1799 . join ', ', @$args_ref
1800 . "'. No action taken.");
1801 return;
12ac2576
JP
1802}
1803
99870f4d
KW
1804sub carp_extra_args {
1805 my $args_ref = shift;
1806 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1807
99870f4d
KW
1808 unless (ref $args_ref) {
1809 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1810 return;
1811 }
1812 my ($package, $file, $line) = caller;
1813 my $subroutine = (caller 1)[3];
cf25bb62 1814
99870f4d
KW
1815 my $list;
1816 if (ref $args_ref eq 'HASH') {
1817 foreach my $key (keys %$args_ref) {
1818 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1819 }
99870f4d 1820 $list = join ', ', each %{$args_ref};
cf25bb62 1821 }
99870f4d
KW
1822 elsif (ref $args_ref eq 'ARRAY') {
1823 foreach my $arg (@$args_ref) {
1824 $arg = $UNDEF unless defined $arg;
1825 }
1826 $list = join ', ', @$args_ref;
1827 }
1828 else {
1829 my_carp_bug("Can't cope with ref "
1830 . ref($args_ref)
1831 . " . argument to 'carp_extra_args'. Not checking arguments.");
1832 return;
1833 }
1834
1835 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1836 return;
d73e5302
JH
1837}
1838
99870f4d
KW
1839package main;
1840
1841{ # Closure
1842
1843 # This program uses the inside-out method for objects, as recommended in
1b6b3fa9
KW
1844 # "Perl Best Practices". (This is the best solution still, since this has
1845 # to run under miniperl.) This closure aids in generating those. There
99870f4d
KW
1846 # are two routines. setup_package() is called once per package to set
1847 # things up, and then set_access() is called for each hash representing a
1848 # field in the object. These routines arrange for the object to be
1849 # properly destroyed when no longer used, and for standard accessor
1850 # functions to be generated. If you need more complex accessors, just
1851 # write your own and leave those accesses out of the call to set_access().
1852 # More details below.
1853
1854 my %constructor_fields; # fields that are to be used in constructors; see
1855 # below
1856
1857 # The values of this hash will be the package names as keys to other
1858 # hashes containing the name of each field in the package as keys, and
1859 # references to their respective hashes as values.
1860 my %package_fields;
1861
1862 sub setup_package {
1863 # Sets up the package, creating standard DESTROY and dump methods
1864 # (unless already defined). The dump method is used in debugging by
1865 # simple_dumper().
1866 # The optional parameters are:
1867 # a) a reference to a hash, that gets populated by later
1868 # set_access() calls with one of the accesses being
1869 # 'constructor'. The caller can then refer to this, but it is
1870 # not otherwise used by these two routines.
1871 # b) a reference to a callback routine to call during destruction
1872 # of the object, before any fields are actually destroyed
1873
1874 my %args = @_;
1875 my $constructor_ref = delete $args{'Constructor_Fields'};
1876 my $destroy_callback = delete $args{'Destroy_Callback'};
1877 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1878
1879 my %fields;
1880 my $package = (caller)[0];
1881
1882 $package_fields{$package} = \%fields;
1883 $constructor_fields{$package} = $constructor_ref;
1884
1885 unless ($package->can('DESTROY')) {
1886 my $destroy_name = "${package}::DESTROY";
1887 no strict "refs";
1888
1889 # Use typeglob to give the anonymous subroutine the name we want
1890 *$destroy_name = sub {
1891 my $self = shift;
ffe43484 1892 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1893
1894 $self->$destroy_callback if $destroy_callback;
1895 foreach my $field (keys %{$package_fields{$package}}) {
1896 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1897 delete $package_fields{$package}{$field}{$addr};
1898 }
1899 return;
1900 }
1901 }
1902
1903 unless ($package->can('dump')) {
1904 my $dump_name = "${package}::dump";
1905 no strict "refs";
1906 *$dump_name = sub {
1907 my $self = shift;
1908 return dump_inside_out($self, $package_fields{$package}, @_);
1909 }
1910 }
1911 return;
1912 }
1913
1914 sub set_access {
1915 # Arrange for the input field to be garbage collected when no longer
1916 # needed. Also, creates standard accessor functions for the field
1917 # based on the optional parameters-- none if none of these parameters:
1918 # 'addable' creates an 'add_NAME()' accessor function.
1919 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1920 # function.
1921 # 'settable' creates a 'set_NAME()' accessor function.
1922 # 'constructor' doesn't create an accessor function, but adds the
1923 # field to the hash that was previously passed to
1924 # setup_package();
1925 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1926 # 'add' etc. all mean 'addable'.
1927 # The read accessor function will work on both array and scalar
1928 # values. If another accessor in the parameter list is 'a', the read
1929 # access assumes an array. You can also force it to be array access
1930 # by specifying 'readable_array' instead of 'readable'
1931 #
1932 # A sort-of 'protected' access can be set-up by preceding the addable,
1933 # readable or settable with some initial portion of 'protected_' (but,
1934 # the underscore is required), like 'p_a', 'pro_set', etc. The
1935 # "protection" is only by convention. All that happens is that the
1936 # accessor functions' names begin with an underscore. So instead of
1937 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1938 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1939 # end of each package, and then storing the __LINE__ ranges and
1940 # checking them on every accessor. But that is way overkill.)
1941
1942 # We create anonymous subroutines as the accessors and then use
1943 # typeglobs to assign them to the proper package and name
1944
1945 my $name = shift; # Name of the field
1946 my $field = shift; # Reference to the inside-out hash containing the
1947 # field
1948
1949 my $package = (caller)[0];
1950
1951 if (! exists $package_fields{$package}) {
1952 croak "$0: Must call 'setup_package' before 'set_access'";
1953 }
d73e5302 1954
99870f4d
KW
1955 # Stash the field so DESTROY can get it.
1956 $package_fields{$package}{$name} = $field;
cf25bb62 1957
99870f4d
KW
1958 # Remaining arguments are the accessors. For each...
1959 foreach my $access (@_) {
1960 my $access = lc $access;
cf25bb62 1961
99870f4d 1962 my $protected = "";
cf25bb62 1963
99870f4d
KW
1964 # Match the input as far as it goes.
1965 if ($access =~ /^(p[^_]*)_/) {
1966 $protected = $1;
1967 if (substr('protected_', 0, length $protected)
1968 eq $protected)
1969 {
1970
1971 # Add 1 for the underscore not included in $protected
1972 $access = substr($access, length($protected) + 1);
1973 $protected = '_';
1974 }
1975 else {
1976 $protected = "";
1977 }
1978 }
1979
1980 if (substr('addable', 0, length $access) eq $access) {
1981 my $subname = "${package}::${protected}add_$name";
1982 no strict "refs";
1983
1984 # add_ accessor. Don't add if already there, which we
1985 # determine using 'eq' for scalars and '==' otherwise.
1986 *$subname = sub {
1987 use strict "refs";
1988 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1989 my $self = shift;
1990 my $value = shift;
ffe43484 1991 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1992 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1993 if (ref $value) {
f998e60c 1994 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1995 }
1996 else {
f998e60c 1997 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1998 }
f998e60c 1999 push @{$field->{$addr}}, $value;
99870f4d
KW
2000 return;
2001 }
2002 }
2003 elsif (substr('constructor', 0, length $access) eq $access) {
2004 if ($protected) {
2005 Carp::my_carp_bug("Can't set-up 'protected' constructors")
2006 }
2007 else {
2008 $constructor_fields{$package}{$name} = $field;
2009 }
2010 }
2011 elsif (substr('readable_array', 0, length $access) eq $access) {
2012
2013 # Here has read access. If one of the other parameters for
2014 # access is array, or this one specifies array (by being more
2015 # than just 'readable_'), then create a subroutine that
2016 # assumes the data is an array. Otherwise just a scalar
2017 my $subname = "${package}::${protected}$name";
2018 if (grep { /^a/i } @_
2019 or length($access) > length('readable_'))
2020 {
2021 no strict "refs";
2022 *$subname = sub {
2023 use strict "refs";
23e33b60 2024 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 2025 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
2026 if (ref $field->{$addr} ne 'ARRAY') {
2027 my $type = ref $field->{$addr};
2028 $type = 'scalar' unless $type;
2029 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
2030 return;
2031 }
2032 return scalar @{$field->{$addr}} unless wantarray;
2033
2034 # Make a copy; had problems with caller modifying the
2035 # original otherwise
2036 my @return = @{$field->{$addr}};
2037 return @return;
2038 }
2039 }
2040 else {
2041
2042 # Here not an array value, a simpler function.
2043 no strict "refs";
2044 *$subname = sub {
2045 use strict "refs";
23e33b60 2046 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 2047 no overloading;
051df77b 2048 return $field->{pack 'J', $_[0]};
99870f4d
KW
2049 }
2050 }
2051 }
2052 elsif (substr('settable', 0, length $access) eq $access) {
2053 my $subname = "${package}::${protected}set_$name";
2054 no strict "refs";
2055 *$subname = sub {
2056 use strict "refs";
23e33b60
KW
2057 if (main::DEBUG) {
2058 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
2059 Carp::carp_extra_args(\@_) if @_ > 2;
2060 }
2061 # $self is $_[0]; $value is $_[1]
f998e60c 2062 no overloading;
051df77b 2063 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
2064 return;
2065 }
2066 }
2067 else {
2068 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
2069 }
cf25bb62 2070 }
99870f4d 2071 return;
cf25bb62 2072 }
99870f4d
KW
2073}
2074
2075package Input_file;
2076
2077# All input files use this object, which stores various attributes about them,
2078# and provides for convenient, uniform handling. The run method wraps the
2079# processing. It handles all the bookkeeping of opening, reading, and closing
2080# the file, returning only significant input lines.
2081#
2082# Each object gets a handler which processes the body of the file, and is
74cd47d0
KW
2083# called by run(). All character property files must use the generic,
2084# default handler, which has code scrubbed to handle things you might not
2085# expect, including automatic EBCDIC handling. For files that don't deal with
2086# mapping code points to a property value, such as test files,
2087# PropertyAliases, PropValueAliases, and named sequences, you can override the
2088# handler to be a custom one. Such a handler should basically be a
2089# while(next_line()) {...} loop.
99870f4d
KW
2090#
2091# You can also set up handlers to
40045c2d 2092# 0) call during object construction time, after everything else is done
537124e4 2093# 1) call before the first line is read, for pre processing
83b68635
KW
2094# 2) call to adjust each line of the input before the main handler gets
2095# them. This can be automatically generated, if appropriately simple
2096# enough, by specifiying a Properties parameter in the constructor.
99870f4d 2097# 3) call upon EOF before the main handler exits its loop
537124e4 2098# 4) call at the end, for post processing
99870f4d
KW
2099#
2100# $_ is used to store the input line, and is to be filtered by the
2101# each_line_handler()s. So, if the format of the line is not in the desired
2102# format for the main handler, these are used to do that adjusting. They can
2103# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
3848e5b7
KW
2104# so the $_ output of one is used as the input to the next. The eof handler
2105# is also stackable, but none of the others are, but could easily be changed
2106# to be so.
99870f4d 2107#
a0204bb1
KW
2108# Some properties are used by the Perl core but aren't defined until later
2109# Unicode releases. The perl interpreter would have problems working when
2110# compiled with an earlier Unicode version that doesn't have them, so we need
2111# to define them somehow for those releases. The 'Early' constructor
2112# parameter can be used to automatically handle this. It is essentially
2113# ignored if the Unicode version being compiled has a data file for this
2114# property. Either code to execute or a file to read can be specified.
2115# Details are at the %early definition.
2116#
99870f4d
KW
2117# Most of the handlers can call insert_lines() or insert_adjusted_lines()
2118# which insert the parameters as lines to be processed before the next input
3848e5b7 2119# file line is read. This allows the EOF handler(s) to flush buffers, for
99870f4d
KW
2120# example. The difference between the two routines is that the lines inserted
2121# by insert_lines() are subjected to the each_line_handler()s. (So if you
9be22ca5
KW
2122# called it from such a handler, you would get infinite recursion without some
2123# mechanism to prevent that.) Lines inserted by insert_adjusted_lines() go
2124# directly to the main handler without any adjustments. If the
2125# post-processing handler calls any of these, there will be no effect. Some
2126# error checking for these conditions could be added, but it hasn't been done.
99870f4d
KW
2127#
2128# carp_bad_line() should be called to warn of bad input lines, which clears $_
2129# to prevent further processing of the line. This routine will output the
2130# message as a warning once, and then keep a count of the lines that have the
2131# same message, and output that count at the end of the file's processing.
2132# This keeps the number of messages down to a manageable amount.
2133#
2134# get_missings() should be called to retrieve any @missing input lines.
2135# Messages will be raised if this isn't done if the options aren't to ignore
2136# missings.
2137
2138sub trace { return main::trace(@_); }
2139
99870f4d
KW
2140{ # Closure
2141 # Keep track of fields that are to be put into the constructor.
2142 my %constructor_fields;
2143
2144 main::setup_package(Constructor_Fields => \%constructor_fields);
2145
2146 my %file; # Input file name, required
2147 main::set_access('file', \%file, qw{ c r });
2148
2149 my %first_released; # Unicode version file was first released in, required
2150 main::set_access('first_released', \%first_released, qw{ c r });
2151
2152 my %handler; # Subroutine to process the input file, defaults to
2153 # 'process_generic_property_file'
2154 main::set_access('handler', \%handler, qw{ c });
2155
2156 my %property;
2157 # name of property this file is for. defaults to none, meaning not
2158 # applicable, or is otherwise determinable, for example, from each line.
696609bf 2159 main::set_access('property', \%property, qw{ c r });
99870f4d
KW
2160
2161 my %optional;
3274d896
KW
2162 # This is either an unsigned number, or a list of property names. In the
2163 # former case, if it is non-zero, it means the file is optional, so if the
2164 # file is absent, no warning about that is output. In the latter case, it
2165 # is a list of properties that the file (exclusively) defines. If the
2166 # file is present, tables for those properties will be produced; if
2167 # absent, none will, even if they are listed elsewhere (namely
2168 # PropertyAliases.txt and PropValueAliases.txt) as being in this release,
2169 # and no warnings will be raised about them not being available. (And no
2170 # warning about the file itself will be raised.)
2171 main::set_access('optional', \%optional, qw{ c readable_array } );
99870f4d
KW
2172
2173 my %non_skip;
2174 # This is used for debugging, to skip processing of all but a few input
2175 # files. Add 'non_skip => 1' to the constructor for those files you want
2176 # processed when you set the $debug_skip global.
2177 main::set_access('non_skip', \%non_skip, 'c');
2178
37e2e78e 2179 my %skip;
0dee21d6
KW
2180 # This is used to skip processing of this input file (semi-) permanently.
2181 # The value should be the reason the file is being skipped. It is used
2182 # for files that we aren't planning to process anytime soon, but want to
2183 # allow to be in the directory and be checked for their names not
2184 # conflicting with any other files on a DOS 8.3 name filesystem, but to
2185 # not otherwise be processed, and to not raise a warning about not being
2186 # handled. In the constructor call, any value that evaluates to a numeric
2187 # 0 or undef means don't skip. Any other value is a string giving the
2188 # reason it is being skippped, and this will appear in generated pod.
2189 # However, an empty string reason will suppress the pod entry.
2190 # Internally, calls that evaluate to numeric 0 are changed into undef to
2191 # distinguish them from an empty string call.
2192 main::set_access('skip', \%skip, 'c', 'r');
37e2e78e 2193
99870f4d
KW
2194 my %each_line_handler;
2195 # list of subroutines to look at and filter each non-comment line in the
2196 # file. defaults to none. The subroutines are called in order, each is
2197 # to adjust $_ for the next one, and the final one adjusts it for
2198 # 'handler'
2199 main::set_access('each_line_handler', \%each_line_handler, 'c');
2200
83b68635
KW
2201 my %properties; # Optional ordered list of the properties that occur in each
2202 # meaningful line of the input file. If present, an appropriate
2203 # each_line_handler() is automatically generated and pushed onto the stack
2204 # of such handlers. This is useful when a file contains multiple
2205 # proerties per line, but no other special considerations are necessary.
2206 # The special value "<ignored>" means to discard the corresponding input
2207 # field.
2208 # Any @missing lines in the file should also match this syntax; no such
2209 # files exist as of 6.3. But if it happens in a future release, the code
2210 # could be expanded to properly parse them.
2211 main::set_access('properties', \%properties, qw{ c r });
2212
99870f4d
KW
2213 my %has_missings_defaults;
2214 # ? Are there lines in the file giving default values for code points
2215 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
2216 # the norm, but IGNORED means it has such lines, but the handler doesn't
2217 # use them. Having these three states allows us to catch changes to the
83b68635
KW
2218 # UCD that this program should track. XXX This could be expanded to
2219 # specify the syntax for such lines, like %properties above.
99870f4d
KW
2220 main::set_access('has_missings_defaults',
2221 \%has_missings_defaults, qw{ c r });
2222
40045c2d
KW
2223 my %construction_time_handler;
2224 # Subroutine to call at the end of the new method. If undef, no such
2225 # handler is called.
2226 main::set_access('construction_time_handler',
2227 \%construction_time_handler, qw{ c });
2228
99870f4d
KW
2229 my %pre_handler;
2230 # Subroutine to call before doing anything else in the file. If undef, no
2231 # such handler is called.
2232 main::set_access('pre_handler', \%pre_handler, qw{ c });
2233
2234 my %eof_handler;
3848e5b7 2235 # Subroutines to call upon getting an EOF on the input file, but before
99870f4d
KW
2236 # that is returned to the main handler. This is to allow buffers to be
2237 # flushed. The handler is expected to call insert_lines() or
2238 # insert_adjusted() with the buffered material
3848e5b7 2239 main::set_access('eof_handler', \%eof_handler, qw{ c });
99870f4d
KW
2240
2241 my %post_handler;
2242 # Subroutine to call after all the lines of the file are read in and
9be22ca5
KW
2243 # processed. If undef, no such handler is called. Note that this cannot
2244 # add lines to be processed; instead use eof_handler
99870f4d
KW
2245 main::set_access('post_handler', \%post_handler, qw{ c });
2246
2247 my %progress_message;
2248 # Message to print to display progress in lieu of the standard one
2249 main::set_access('progress_message', \%progress_message, qw{ c });
2250
2251 my %handle;
2252 # cache open file handle, internal. Is undef if file hasn't been
2253 # processed at all, empty if has;
2254 main::set_access('handle', \%handle);
2255
2256 my %added_lines;
2257 # cache of lines added virtually to the file, internal
2258 main::set_access('added_lines', \%added_lines);
2259
74cd47d0
KW
2260 my %remapped_lines;
2261 # cache of lines added virtually to the file, internal
2262 main::set_access('remapped_lines', \%remapped_lines);
2263
99870f4d
KW
2264 my %errors;
2265 # cache of errors found, internal
2266 main::set_access('errors', \%errors);
2267
2268 my %missings;
2269 # storage of '@missing' defaults lines
2270 main::set_access('missings', \%missings);
2271
a0204bb1
KW
2272 my %early;
2273 # Used for properties that must be defined (for Perl's purposes) on
2274 # versions of Unicode earlier than Unicode itself defines them. The
2275 # parameter is an array (it would be better to be a hash, but not worth
2276 # bothering about due to its rare use).
2277 #
2278 # The first element is either a code reference to call when in a release
2279 # earlier than the Unicode file is available in, or it is an alternate
2280 # file to use instead of the non-existent one. This file must have been
2281 # plunked down in the same directory as mktables. Should you be compiling
2282 # on a release that needs such a file, mktables will abort the
2283 # compilation, and tell you where to get the necessary file(s), and what
2284 # name(s) to use to store them as.
2285 # In the case of specifying an alternate file, the array must contain two
2286 # further elements:
2287 #
2288 # [1] is the name of the property that will be generated by this file.
2289 # The class automatically takes the input file and excludes any code
2290 # points in it that were not assigned in the Unicode version being
2291 # compiled. It then uses this result to define the property in the given
2292 # version. Since the property doesn't actually exist in the Unicode
2293 # version being compiled, this should be a name accessible only by core
2294 # perl. If it is the same name as the regular property, the constructor
2295 # will mark the output table as a $PLACEHOLDER so that it doesn't actually
2296 # get output, and so will be unusable by non-core code. Otherwise it gets
2297 # marked as $INTERNAL_ONLY.
2298 #
2299 # [2] is a property value to assign (only when compiling Unicode 1.1.5) to
2300 # the Hangul syllables in that release (which were ripped out in version
2301 # 2) for the given property . (Hence it is ignored except when compiling
2302 # version 1. You only get one value that applies to all of them, which
2303 # may not be the actual reality, but probably nobody cares anyway for
2304 # these obsolete characters.)
2305 #
3d36fe12
KW
2306 # [3] if present is the default value for the property to assign for code
2307 # points not given in the input. If not present, the default from the
2308 # normal property is used
2309 #
4ed2b786
KW
2310 # [-1] If there is an extra final element that is the string 'ONLY_EARLY'.
2311 # it means to not add the name in [1] as an alias to the property name
2312 # used for these. Normally, when compiling Unicode versions that don't
2313 # invoke the early handling, the name is added as a synonym.
cbdbe9d4 2314 #
a0204bb1
KW
2315 # Not all files can be handled in the above way, and so the code ref
2316 # alternative is available. It can do whatever it needs to. The other
2317 # array elements are optional in this case, and the code is free to use or
2318 # ignore them if they are present.
2319 #
2320 # Internally, the constructor unshifts a 0 or 1 onto this array to
2321 # indicate if an early alternative is actually being used or not. This
2322 # makes for easier testing later on.
2323 main::set_access('early', \%early, 'c');
2324
4ed2b786
KW
2325 my %only_early;
2326 main::set_access('only_early', \%only_early, 'c');
2327
0dee21d6
KW
2328 my %required_even_in_debug_skip;
2329 # debug_skip is used to speed up compilation during debugging by skipping
2330 # processing files that are not needed for the task at hand. However,
2331 # some files pretty much can never be skipped, and this is used to specify
2332 # that this is one of them. In order to skip this file, the call to the
2333 # constructor must be edited to comment out this parameter.
2334 main::set_access('required_even_in_debug_skip',
2335 \%required_even_in_debug_skip, 'c');
2336
d8062573
KW
2337 my %withdrawn;
2338 # Some files get removed from the Unicode DB. This is a version object
2339 # giving the first release without this file.
2340 main::set_access('withdrawn', \%withdrawn, 'c');
2341
0dee21d6
KW
2342 my %in_this_release;
2343 # Calculated value from %first_released and %withdrawn. Are we compiling
2344 # a Unicode release which includes this file?
2345 main::set_access('in_this_release', \%in_this_release);
2346
74cd47d0
KW
2347 sub _next_line;
2348 sub _next_line_with_remapped_range;
2349
99870f4d
KW
2350 sub new {
2351 my $class = shift;
2352
2353 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2354 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2355
2356 # Set defaults
2357 $handler{$addr} = \&main::process_generic_property_file;
2358 $non_skip{$addr} = 0;
0dee21d6 2359 $skip{$addr} = undef;
99870f4d
KW
2360 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2361 $handle{$addr} = undef;
2362 $added_lines{$addr} = [ ];
74cd47d0 2363 $remapped_lines{$addr} = [ ];
99870f4d 2364 $each_line_handler{$addr} = [ ];
3848e5b7 2365 $eof_handler{$addr} = [ ];
99870f4d
KW
2366 $errors{$addr} = { };
2367 $missings{$addr} = [ ];
a0204bb1 2368 $early{$addr} = [ ];
3274d896 2369 $optional{$addr} = [ ];
99870f4d
KW
2370
2371 # Two positional parameters.
99f78760 2372 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2373 $file{$addr} = main::internal_file_to_platform(shift);
2374 $first_released{$addr} = shift;
2375
2376 # The rest of the arguments are key => value pairs
2377 # %constructor_fields has been set up earlier to list all possible
2378 # ones. Either set or push, depending on how the default has been set
2379 # up just above.
2380 my %args = @_;
2381 foreach my $key (keys %args) {
2382 my $argument = $args{$key};
2383
2384 # Note that the fields are the lower case of the constructor keys
2385 my $hash = $constructor_fields{lc $key};
2386 if (! defined $hash) {
2387 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2388 next;
2389 }
2390 if (ref $hash->{$addr} eq 'ARRAY') {
2391 if (ref $argument eq 'ARRAY') {
2392 foreach my $argument (@{$argument}) {
2393 next if ! defined $argument;
2394 push @{$hash->{$addr}}, $argument;
2395 }
2396 }
2397 else {
2398 push @{$hash->{$addr}}, $argument if defined $argument;
2399 }
2400 }
2401 else {
2402 $hash->{$addr} = $argument;
2403 }
2404 delete $args{$key};
2405 };
2406
0dee21d6
KW
2407 $non_skip{$addr} = 1 if $required_even_in_debug_skip{$addr};
2408
2409 # Convert 0 (meaning don't skip) to undef
2410 undef $skip{$addr} unless $skip{$addr};
2411
3274d896
KW
2412 # Handle the case where this file is optional
2413 my $pod_message_for_non_existent_optional = "";
2414 if ($optional{$addr}->@*) {
2415
2416 # First element is the pod message
2417 $pod_message_for_non_existent_optional
2418 = shift $optional{$addr}->@*;
2419 # Convert a 0 'Optional' argument to an empty list to make later
2420 # code more concise.
2421 if ( $optional{$addr}->@*
2422 && $optional{$addr}->@* == 1
2423 && $optional{$addr}[0] ne ""
2424 && $optional{$addr}[0] !~ /\D/
2425 && $optional{$addr}[0] == 0)
2426 {
2427 $optional{$addr} = [ ];
2428 }
2429 else { # But if the only element doesn't evaluate to 0, make sure
2430 # that this file is indeed considered optional below.
2431 unshift $optional{$addr}->@*, 1;
2432 }
2433 }
2434
0dee21d6 2435 my $progress;
a0204bb1
KW
2436 my $function_instead_of_file = 0;
2437
4ed2b786
KW
2438 if ($early{$addr}->@* && $early{$addr}[-1] eq 'ONLY_EARLY') {
2439 $only_early{$addr} = 1;
2440 pop $early{$addr}->@*;
cbdbe9d4
KW
2441 }
2442
a0204bb1
KW
2443 # If we are compiling a Unicode release earlier than the file became
2444 # available, the constructor may have supplied a substitute
2445 if ($first_released{$addr} gt $v_version && $early{$addr}->@*) {
2446
2447 # Yes, we have a substitute, that we will use; mark it so
2448 unshift $early{$addr}->@*, 1;
2449
2450 # See the definition of %early for what the array elements mean.
4ed2b786
KW
2451 # Note that we have just unshifted onto the array, so the numbers
2452 # below are +1 of those in the %early description.
a0204bb1
KW
2453 # If we have a property this defines, create a table and default
2454 # map for it now (at essentially compile time), so that it will be
2455 # available for the whole of run time. (We will want to add this
2456 # name as an alias when we are using the official property name;
2457 # but this must be deferred until run(), because at construction
2458 # time the official names have yet to be defined.)
2459 if ($early{$addr}[2]) {
2460 my $fate = ($property{$addr}
2461 && $property{$addr} eq $early{$addr}[2])
2462 ? $PLACEHOLDER
2463 : $INTERNAL_ONLY;
2464 my $prop_object = Property->new($early{$addr}[2],
2465 Fate => $fate,
2466 Perl_Extension => 1,
2467 );
2468
3d36fe12
KW
2469 # If not specified by the constructor, use the default mapping
2470 # for the regular property for this substitute one.
4ed2b786
KW
2471 if ($early{$addr}[4]) {
2472 $prop_object->set_default_map($early{$addr}[4]);
3d36fe12
KW
2473 }
2474 elsif ( defined $property{$addr}
2475 && defined $default_mapping{$property{$addr}})
a0204bb1
KW
2476 {
2477 $prop_object
2478 ->set_default_map($default_mapping{$property{$addr}});
2479 }
2480 }
2481
2482 if (ref $early{$addr}[1] eq 'CODE') {
2483 $function_instead_of_file = 1;
2484
2485 # If the first element of the array is a code ref, the others
2486 # are optional.
2487 $handler{$addr} = $early{$addr}[1];
2488 $property{$addr} = $early{$addr}[2]
2489 if defined $early{$addr}[2];
2490 $progress = "substitute $file{$addr}";
2491
2492 undef $file{$addr};
2493 }
2494 else { # Specifying a substitute file
2495
2496 if (! main::file_exists($early{$addr}[1])) {
2497
2498 # If we don't see the substitute file, generate an error
2499 # message giving the needed things, and add it to the list
2500 # of such to output before actual processing happens
2501 # (hence the user finds out all of them in one run).
ce8d64d9
KW
2502 # Instead of creating a general method for NameAliases,
2503 # hard-code it here, as there is unlikely to ever be a
2504 # second one which needs special handling.
2505 my $string_version = ($file{$addr} eq "NameAliases.txt")
2506 ? 'at least 6.1 (the later, the better)'
2507 : sprintf "%vd", $first_released{$addr};
a0204bb1
KW
2508 push @missing_early_files, <<END;
2509'$file{$addr}' version $string_version should be copied to '$early{$addr}[1]'.
2510END
2511 ;
2512 return;
2513 }
2514 $progress = $early{$addr}[1];
2515 $progress .= ", substituting for $file{$addr}" if $file{$addr};
2516 $file{$addr} = $early{$addr}[1];
2517 $property{$addr} = $early{$addr}[2];
0dee21d6 2518
a0204bb1
KW
2519 # Ignore code points not in the version being compiled
2520 push $each_line_handler{$addr}->@*, \&_exclude_unassigned;
2521
2522 if ( $v_version lt v2.0 # Hanguls in this release ...
4ed2b786 2523 && defined $early{$addr}[3]) # ... need special treatment
a0204bb1
KW
2524 {
2525 push $eof_handler{$addr}->@*, \&_fixup_obsolete_hanguls;
2526 }
2527 }
2528
2529 # And this substitute is valid for all releases.
2530 $first_released{$addr} = v0;
2531 }
2532 else { # Normal behavior
0dee21d6 2533 $progress = $file{$addr};
a0204bb1 2534 unshift $early{$addr}->@*, 0; # No substitute
99870f4d
KW
2535 }
2536
0dee21d6
KW
2537 my $file = $file{$addr};
2538 $progress_message{$addr} = "Processing $progress"
2539 unless $progress_message{$addr};
2540
d8062573
KW
2541 # A file should be there if it is within the window of versions for
2542 # which Unicode supplies it
2543 if ($withdrawn{$addr} && $withdrawn{$addr} le $v_version) {
2544 $in_this_release{$addr} = 0;
2545 $skip{$addr} = "";
2546 }
2547 else {
0dee21d6
KW
2548 $in_this_release{$addr} = $first_released{$addr} le $v_version;
2549
a0204bb1
KW
2550 # Check that the file for this object (possibly using a substitute
2551 # for early releases) exists or we have a function alternative
2552 if ( ! $function_instead_of_file
2553 && ! main::file_exists($file))
0dee21d6
KW
2554 {
2555 # Here there is nothing available for this release. This is
2556 # fine if we aren't expecting anything in this release.
2557 if (! $in_this_release{$addr}) {
2558 $skip{$addr} = ""; # Don't remark since we expected
2559 # nothing and got nothing
2560 }
3274d896 2561 elsif ($optional{$addr}->@*) {
0dee21d6 2562
3274d896
KW
2563 # Here the file is optional in this release; Use the
2564 # passed in text to document this case in the pod.
2565 $skip{$addr} = $pod_message_for_non_existent_optional;
0dee21d6
KW
2566 }
2567 elsif ( $in_this_release{$addr}
2568 && ! defined $skip{$addr}
2569 && defined $file)
2570 { # Doesn't exist but should.
2571 $skip{$addr} = "'$file' not found. Possibly Big problems";
2572 Carp::my_carp($skip{$addr});
2573 }
2574 }
2575 elsif ($debug_skip && ! defined $skip{$addr} && ! $non_skip{$addr})
2576 {
2577
2578 # The file exists; if not skipped for another reason, and we are
2579 # skipping most everything during debugging builds, use that as
2580 # the skip reason.
2581 $skip{$addr} = '$debug_skip is on'
2582 }
d8062573 2583 }
0dee21d6
KW
2584
2585 if ( ! $debug_skip
2586 && $non_skip{$addr}
2587 && ! $required_even_in_debug_skip{$addr}
2588 && $verbosity)
2589 {
2590 print "Warning: " . __PACKAGE__ . " constructor for $file has useless 'non_skip' in it\n";
a3a8c5f0 2591 }
99870f4d 2592
0dee21d6 2593 # Here, we have figured out if we will be skipping this file or not.
3274d896
KW
2594 # If so, we add any single property it defines to any passed in
2595 # optional property list. These will be dealt with at run time.
0dee21d6 2596 if (defined $skip{$addr}) {
3274d896
KW
2597 if ($property{$addr}) {
2598 push $optional{$addr}->@*, $property{$addr};
2599 }
2600 } # Otherwise, are going to process the file.
0dee21d6
KW
2601 elsif ($property{$addr}) {
2602
2603 # If the file has a property defined in the constructor for it, it
2604 # means that the property is not listed in the file's entries. So
2605 # add a handler (to the list of line handlers) to insert the
2606 # property name into the lines, to provide a uniform interface to
2607 # the final processing subroutine.
2608 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
09ca89ce 2609 }
83b68635
KW
2610 elsif ($properties{$addr}) {
2611
9be22ca5
KW
2612 # Similarly, there may be more than one property represented on
2613 # each line, with no clue but the constructor input what those
2614 # might be. Add a handler for each line in the input so that it
2615 # creates a separate input line for each property in those input
2616 # lines, thus making them suitable to handle generically.
83b68635
KW
2617
2618 push @{$each_line_handler{$addr}},
2619 sub {
2620 my $file = shift;
2621 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2622
2623 my @fields = split /\s*;\s*/, $_, -1;
2624
2625 if (@fields - 1 > @{$properties{$addr}}) {
2626 $file->carp_bad_line('Extra fields');
2627 $_ = "";
2628 return;
2629 }
2630 my $range = shift @fields; # 0th element is always the
2631 # range
2632
2633 # The next fields in the input line correspond
2634 # respectively to the stored properties.
2635 for my $i (0 .. @{$properties{$addr}} - 1) {
2636 my $property_name = $properties{$addr}[$i];
2637 next if $property_name eq '<ignored>';
2638 $file->insert_adjusted_lines(
2639 "$range; $property_name; $fields[$i]");
2640 }
2641 $_ = "";
2642
2643 return;
2644 };
2645 }
37e2e78e 2646
9be22ca5 2647 { # On non-ascii platforms, we use a special pre-handler
74cd47d0
KW
2648 no strict;
2649 no warnings 'once';
2650 *next_line = (main::NON_ASCII_PLATFORM)
2651 ? *_next_line_with_remapped_range
2652 : *_next_line;
2653 }
2654
40045c2d
KW
2655 &{$construction_time_handler{$addr}}($self)
2656 if $construction_time_handler{$addr};
2657
99870f4d 2658 return $self;
d73e5302
JH
2659 }
2660
cf25bb62 2661
99870f4d
KW
2662 use overload
2663 fallback => 0,
2664 qw("") => "_operator_stringify",
2665 "." => \&main::_operator_dot,
1285127e 2666 ".=" => \&main::_operator_dot_equal,
99870f4d 2667 ;
cf25bb62 2668
99870f4d
KW
2669 sub _operator_stringify {
2670 my $self = shift;
cf25bb62 2671
99870f4d 2672 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2673 }
d73e5302 2674
99870f4d
KW
2675 sub run {
2676 # Process the input object $self. This opens and closes the file and
2677 # calls all the handlers for it. Currently, this can only be called
3848e5b7 2678 # once per file, as it destroy's the EOF handlers
d73e5302 2679
0dee21d6 2680 # flag to make sure extracted files are processed early
f8b69004 2681 state $seen_non_extracted = 0;
0dee21d6 2682
99870f4d
KW
2683 my $self = shift;
2684 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2685
ffe43484 2686 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2687
99870f4d 2688 my $file = $file{$addr};
d73e5302 2689
0dee21d6
KW
2690 if (! $file) {
2691 $handle{$addr} = 'pretend_is_open';
99870f4d
KW
2692 }
2693 else {
f8b69004 2694 if ($seen_non_extracted) {
9be22ca5
KW
2695 if ($file =~ /$EXTRACTED/i) # Some platforms may change the
2696 # case of the file's name
2697 {
1675ea0d 2698 Carp::my_carp_bug(main::join_lines(<<END
99f78760 2699$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2700anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2701have subtle problems
2702END
2703 ));
2704 }
2705 }
2706 elsif ($EXTRACTED_DIR
0dee21d6
KW
2707
2708 # We only do this check for generic property files
2709 && $handler{$addr} == \&main::process_generic_property_file
2710
f8b69004 2711 && $file !~ /$EXTRACTED/i)
99870f4d
KW
2712 {
2713 # We don't set this (by the 'if' above) if we have no
2714 # extracted directory, so if running on an early version,
2715 # this test won't work. Not worth worrying about.
f8b69004 2716 $seen_non_extracted = 1;
99870f4d
KW
2717 }
2718
9be22ca5 2719 # Mark the file as having being processed, and warn if it
99870f4d
KW
2720 # isn't a file we are expecting. As we process the files,
2721 # they are deleted from the hash, so any that remain at the
2722 # end of the program are files that we didn't process.
517956bf 2723 my $fkey = File::Spec->rel2abs($file);
0dee21d6 2724 my $exists = delete $potential_files{lc($fkey)};
faf3cf6b 2725
0dee21d6
KW
2726 Carp::my_carp("Was not expecting '$file'.")
2727 if $exists && ! $in_this_release{$addr};
2728
a0204bb1
KW
2729 # If there is special handling for compiling Unicode releases
2730 # earlier than the first one in which Unicode defines this
2731 # property ...
2732 if ($early{$addr}->@* > 1) {
2733
2734 # Mark as processed any substitute file that would be used in
2735 # such a release
2736 $fkey = File::Spec->rel2abs($early{$addr}[1]);
2737 delete $potential_files{lc($fkey)};
2738
2739 # As commented in the constructor code, when using the
2740 # official property, we still have to allow the publicly
2741 # inaccessible early name so that the core code which uses it
2742 # will work regardless.
4ed2b786
KW
2743 if ( ! $only_early{$addr}
2744 && ! $early{$addr}[0]
2745 && $early{$addr}->@* > 2)
2746 {
a0204bb1 2747 my $early_property_name = $early{$addr}[2];
4ed2b786 2748 if ($property{$addr} ne $early_property_name) {
a0204bb1
KW
2749 main::property_ref($property{$addr})
2750 ->add_alias($early_property_name);
2751 }
2752 }
2753 }
2754
0dee21d6
KW
2755 # We may be skipping this file ...
2756 if (defined $skip{$addr}) {
2757
2758 # If the file isn't supposed to be in this release, there is
2759 # nothing to do
2760 if ($in_this_release{$addr}) {
2761
2762 # But otherwise, we may print a message
2763 if ($debug_skip) {
2764 print STDERR "Skipping input file '$file'",
2765 " because '$skip{$addr}'\n";
2766 }
2767
2768 # And add it to the list of skipped files, which is later
2769 # used to make the pod
2770 $skipped_files{$file} = $skip{$addr};
3274d896
KW
2771
2772 # The 'optional' list contains properties that are also to
2773 # be skipped along with the file. (There may also be
2774 # digits which are just placeholders to make sure it isn't
2775 # an empty list
2776 foreach my $property ($optional{$addr}->@*) {
2777 next unless $property =~ /\D/;
2778 my $prop_object = main::property_ref($property);
2779 next unless defined $prop_object;
2780 $prop_object->set_fate($SUPPRESSED, $skip{$addr});
2781 }
0dee21d6 2782 }
99870f4d 2783
37e2e78e
KW
2784 return;
2785 }
2786
9be22ca5
KW
2787 # Here, we are going to process the file. Open it, converting the
2788 # slashes used in this program into the proper form for the OS
99870f4d
KW
2789 my $file_handle;
2790 if (not open $file_handle, "<", $file) {
2791 Carp::my_carp("Can't open $file. Skipping: $!");
d3664eb9 2792 return;
99870f4d
KW
2793 }
2794 $handle{$addr} = $file_handle; # Cache the open file handle
9e65c3f4 2795
0dee21d6
KW
2796 # If possible, make sure that the file is the correct version.
2797 # (This data isn't available on early Unicode releases or in
2798 # UnicodeData.txt.) We don't do this check if we are using a
2799 # substitute file instead of the official one (though the code
2800 # could be extended to do so).
2801 if ($in_this_release{$addr}
a0204bb1 2802 && ! $early{$addr}[0]
0dee21d6
KW
2803 && lc($file) ne 'unicodedata.txt')
2804 {
96f226dc 2805 if ($file !~ /^Unihan/i) {
0dee21d6
KW
2806
2807 # The non-Unihan files started getting version numbers in
2808 # 3.2, but some files in 4.0 are unchanged from 3.2, and
2809 # marked as 3.2. 4.0.1 is the first version where there
2810 # are no files marked as being from less than 4.0, though
2811 # some are marked as 4.0. In versions after that, the
2812 # numbers are correct.
2813 if ($v_version ge v4.0.1) {
2814 $_ = <$file_handle>; # The version number is in the
2815 # very first line
2816 if ($_ !~ / - $string_version \. /x) {
2817 chomp;
2818 $_ =~ s/^#\s*//;
2819
2820 # 4.0.1 had some valid files that weren't updated.
2821 if (! ($v_version eq v4.0.1 && $_ =~ /4\.0\.0/)) {
2822 die Carp::my_carp("File '$file' is version "
2823 . "'$_'. It should be "
2824 . "version $string_version");
2825 }
2826 }
cafe9cf0 2827 }
96f226dc 2828 }
0dee21d6
KW
2829 elsif ($v_version ge v6.0.0) { # Unihan
2830
2831 # Unihan files didn't get accurate version numbers until
2832 # 6.0. The version is somewhere in the first comment
2833 # block
96f226dc
KW
2834 while (<$file_handle>) {
2835 if ($_ !~ /^#/) {
0dee21d6
KW
2836 Carp::my_carp_bug("Could not find the expected "
2837 . "version info in file '$file'");
96f226dc
KW
2838 last;
2839 }
2840 chomp;
2841 $_ =~ s/^#\s*//;
2842 next if $_ !~ / version: /x;
2843 last if $_ =~ /$string_version/;
0dee21d6
KW
2844 die Carp::my_carp("File '$file' is version "
2845 . "'$_'. It should be "
2846 . "version $string_version");
96f226dc
KW
2847 }
2848 }
9e65c3f4 2849 }
99870f4d
KW
2850 }
2851
0dee21d6 2852 print "$progress_message{$addr}\n" if $verbosity >= $PROGRESS;
99870f4d
KW
2853
2854 # Call any special handler for before the file.
2855 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2856
2857 # Then the main handler
2858 &{$handler{$addr}}($self);
2859
2860 # Then any special post-file handler.
2861 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2862
2863 # If any errors have been accumulated, output the counts (as the first
2864 # error message in each class was output when it was encountered).
2865 if ($errors{$addr}) {
2866 my $total = 0;
2867 my $types = 0;
2868 foreach my $error (keys %{$errors{$addr}}) {
2869 $total += $errors{$addr}->{$error};
2870 delete $errors{$addr}->{$error};
2871 $types++;
2872 }
2873 if ($total > 1) {
2874 my $message
2875 = "A total of $total lines had errors in $file. ";
2876
2877 $message .= ($types == 1)
2878 ? '(Only the first one was displayed.)'
2879 : '(Only the first of each type was displayed.)';
2880 Carp::my_carp($message);
2881 }
2882 }
2883
2884 if (@{$missings{$addr}}) {
2885 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2886 }
2887
2888 # If a real file handle, close it.
2889 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2890 ref $handle{$addr};
2891 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2892 # the file, as opposed to undef
2893 return;
2894 }
2895
74cd47d0 2896 sub _next_line {
99870f4d
KW
2897 # Sets $_ to be the next logical input line, if any. Returns non-zero
2898 # if such a line exists. 'logical' means that any lines that have
2899 # been added via insert_lines() will be returned in $_ before the file
2900 # is read again.
2901
2902 my $self = shift;
2903 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2904
ffe43484 2905 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2906
2907 # Here the file is open (or if the handle is not a ref, is an open
2908 # 'virtual' file). Get the next line; any inserted lines get priority
2909 # over the file itself.
2910 my $adjusted;
2911
2912 LINE:
2913 while (1) { # Loop until find non-comment, non-empty line
2914 #local $to_trace = 1 if main::DEBUG;
2915 my $inserted_ref = shift @{$added_lines{$addr}};
2916 if (defined $inserted_ref) {
2917 ($adjusted, $_) = @{$inserted_ref};
2918 trace $adjusted, $_ if main::DEBUG && $to_trace;
2919 return 1 if $adjusted;
2920 }
2921 else {
2922 last if ! ref $handle{$addr}; # Don't read unless is real file
2923 last if ! defined ($_ = readline $handle{$addr});
2924 }
2925 chomp;
2926 trace $_ if main::DEBUG && $to_trace;
2927
2928 # See if this line is the comment line that defines what property
2929 # value that code points that are not listed in the file should
2930 # have. The format or existence of these lines is not guaranteed
2931 # by Unicode since they are comments, but the documentation says
2932 # that this was added for machine-readability, so probably won't
2933 # change. This works starting in Unicode Version 5.0. They look
2934 # like:
2935 #
2936 # @missing: 0000..10FFFF; Not_Reordered
2937 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2938 # @missing: 0000..10FFFF; ; NaN
2939 #
2940 # Save the line for a later get_missings() call.
2941 if (/$missing_defaults_prefix/) {
2942 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2943 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2944 }
2945 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2946 my @defaults = split /\s* ; \s*/x, $_;
2947
2948 # The first field is the @missing, which ends in a
2949 # semi-colon, so can safely shift.
2950 shift @defaults;
2951
2952 # Some of these lines may have empty field placeholders
2953 # which get in the way. An example is:
2954 # @missing: 0000..10FFFF; ; NaN
2955 # Remove them. Process starting from the top so the
2956 # splice doesn't affect things still to be looked at.
2957 for (my $i = @defaults - 1; $i >= 0; $i--) {
2958 next if $defaults[$i] ne "";
2959 splice @defaults, $i, 1;
2960 }
2961
2962 # What's left should be just the property (maybe) and the
2963 # default. Having only one element means it doesn't have
2964 # the property.
2965 my $default;
2966 my $property;
2967 if (@defaults >= 1) {
2968 if (@defaults == 1) {
2969 $default = $defaults[0];
2970 }
2971 else {
2972 $property = $defaults[0];
2973 $default = $defaults[1];
2974 }
2975 }
2976
2977 if (@defaults < 1
2978 || @defaults > 2
2979 || ($default =~ /^</
2980 && $default !~ /^<code *point>$/i
09f8d0ac
KW
2981 && $default !~ /^<none>$/i
2982 && $default !~ /^<script>$/i))
99870f4d
KW
2983 {
2984 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2985 }
2986 else {
2987
2988 # If the property is missing from the line, it should
2989 # be the one for the whole file
2990 $property = $property{$addr} if ! defined $property;
2991
2992 # Change <none> to the null string, which is what it
2993 # really means. If the default is the code point
2994 # itself, set it to <code point>, which is what
2995 # Unicode uses (but sometimes they've forgotten the
2996 # space)
2997 if ($default =~ /^<none>$/i) {
2998 $default = "";
2999 }
3000 elsif ($default =~ /^<code *point>$/i) {
3001 $default = $CODE_POINT;
3002 }
09f8d0ac
KW
3003 elsif ($default =~ /^<script>$/i) {
3004
3005 # Special case this one. Currently is from
3006 # ScriptExtensions.txt, and means for all unlisted
3007 # code points, use their Script property values.
3008 # For the code points not listed in that file, the
3009 # default value is 'Unknown'.
3010 $default = "Unknown";
3011 }
99870f4d
KW
3012
3013 # Store them as a sub-arrays with both components.
3014 push @{$missings{$addr}}, [ $default, $property ];
3015 }
3016 }
3017
3018 # There is nothing for the caller to process on this comment
3019 # line.
3020 next;
3021 }
3022
3023 # Remove comments and trailing space, and skip this line if the
3024 # result is empty
3025 s/#.*//;
3026 s/\s+$//;
3027 next if /^$/;
3028
3029 # Call any handlers for this line, and skip further processing of
3030 # the line if the handler sets the line to null.
3031 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
3032 &{$sub_ref}($self);
3033 next LINE if /^$/;
3034 }
3035
3036 # Here the line is ok. return success.
3037 return 1;
3038 } # End of looping through lines.
3039
3848e5b7 3040 # If there are EOF handlers, call each (only once) and if it generates
99870f4d 3041 # more lines to process go back in the loop to handle them.
3848e5b7
KW
3042 while ($eof_handler{$addr}->@*) {
3043 &{$eof_handler{$addr}[0]}($self);
3044 shift $eof_handler{$addr}->@*; # Currently only get one shot at it.
99870f4d
KW
3045 goto LINE if $added_lines{$addr};
3046 }
3047
3048 # Return failure -- no more lines.
3049 return 0;
3050
3051 }
3052
74cd47d0
KW
3053 sub _next_line_with_remapped_range {
3054 my $self = shift;
3055 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3056
3057 # like _next_line(), but for use on non-ASCII platforms. It sets $_
3058 # to be the next logical input line, if any. Returns non-zero if such
3059 # a line exists. 'logical' means that any lines that have been added
3060 # via insert_lines() will be returned in $_ before the file is read
3061 # again.
3062 #
3063 # The difference from _next_line() is that this remaps the Unicode
3064 # code points in the input to those of the native platform. Each
3065 # input line contains a single code point, or a single contiguous
3066 # range of them This routine splits each range into its individual
3067 # code points and caches them. It returns the cached values,
3068 # translated into their native equivalents, one at a time, for each
3069 # call, before reading the next line. Since native values can only be
3070 # a single byte wide, no translation is needed for code points above
3071 # 0xFF, and ranges that are entirely above that number are not split.
3072 # If an input line contains the range 254-1000, it would be split into
3073 # three elements: 254, 255, and 256-1000. (The downstream table
3074 # insertion code will sort and coalesce the individual code points
3075 # into appropriate ranges.)
3076
3077 my $addr = do { no overloading; pack 'J', $self; };
3078
3079 while (1) {
3080
3081 # Look in cache before reading the next line. Return any cached
3082 # value, translated
3083 my $inserted = shift @{$remapped_lines{$addr}};
3084 if (defined $inserted) {
3085 trace $inserted if main::DEBUG && $to_trace;
3086 $_ = $inserted =~ s/^ ( \d+ ) /sprintf("%04X", utf8::unicode_to_native($1))/xer;
3087 trace $_ if main::DEBUG && $to_trace;
3088 return 1;
3089 }
3090
3091 # Get the next line.
3092 return 0 unless _next_line($self);
3093
3094 # If there is a special handler for it, return the line,
3095 # untranslated. This should happen only for files that are
3096 # special, not being code-point related, such as property names.
3097 return 1 if $handler{$addr}
3098 != \&main::process_generic_property_file;
3099
3100 my ($range, $property_name, $map, @remainder)
3101 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3102
3103 if (@remainder
3104 || ! defined $property_name
3105 || $range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3106 {
3107 Carp::my_carp_bug("Unrecognized input line '$_'. Ignored");
3108 }
3109
3110 my $low = hex $1;
3111 my $high = (defined $2) ? hex $2 : $low;
3112
3113 # If the input maps the range to another code point, remap the
3114 # target if it is between 0 and 255.
3115 my $tail;
3116 if (defined $map) {
3117 $map =~ s/\b 00 ( [0-9A-F]{2} ) \b/sprintf("%04X", utf8::unicode_to_native(hex $1))/gxe;
3118 $tail = "$property_name; $map";
3119 $_ = "$range; $tail";
3120 }
3121 else {
3122 $tail = $property_name;
3123 }
3124
3125 # If entire range is above 255, just return it, unchanged (except
3126 # any mapped-to code point, already changed above)
3127 return 1 if $low > 255;
3128
3129 # Cache an entry for every code point < 255. For those in the
3130 # range above 255, return a dummy entry for just that portion of
3131 # the range. Note that this will be out-of-order, but that is not
3132 # a problem.
3133 foreach my $code_point ($low .. $high) {
3134 if ($code_point > 255) {
3135 $_ = sprintf "%04X..%04X; $tail", $code_point, $high;
3136 return 1;
3137 }
3138 push @{$remapped_lines{$addr}}, "$code_point; $tail";
3139 }
3140 } # End of looping through lines.
3141
3142 # NOTREACHED
3143 }
3144
99870f4d
KW
3145# Not currently used, not fully tested.
3146# sub peek {
f67a5002 3147# # Non-destructive lookahead one non-adjusted, non-comment, non-blank
99870f4d
KW
3148# # record. Not callable from an each_line_handler(), nor does it call
3149# # an each_line_handler() on the line.
3150#
3151# my $self = shift;
ffe43484 3152# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3153#
3154# foreach my $inserted_ref (@{$added_lines{$addr}}) {
3155# my ($adjusted, $line) = @{$inserted_ref};
3156# next if $adjusted;
3157#
3158# # Remove comments and trailing space, and return a non-empty
3159# # resulting line
3160# $line =~ s/#.*//;
3161# $line =~ s/\s+$//;
3162# return $line if $line ne "";
3163# }
3164#
3165# return if ! ref $handle{$addr}; # Don't read unless is real file
3166# while (1) { # Loop until find non-comment, non-empty line
3167# local $to_trace = 1 if main::DEBUG;
3168# trace $_ if main::DEBUG && $to_trace;
3169# return if ! defined (my $line = readline $handle{$addr});
3170# chomp $line;
3171# push @{$added_lines{$addr}}, [ 0, $line ];
3172#
3173# $line =~ s/#.*//;
3174# $line =~ s/\s+$//;
3175# return $line if $line ne "";
3176# }
3177#
3178# return;
3179# }
3180
3181
3182 sub insert_lines {
3183 # Lines can be inserted so that it looks like they were in the input
3184 # file at the place it was when this routine is called. See also
3185 # insert_adjusted_lines(). Lines inserted via this routine go through
3186 # any each_line_handler()
3187
3188 my $self = shift;
3189
3190 # Each inserted line is an array, with the first element being 0 to
3191 # indicate that this line hasn't been adjusted, and needs to be
3192 # processed.
f998e60c 3193 no overloading;
051df77b 3194 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
3195 return;
3196 }
3197
3198 sub insert_adjusted_lines {
3199 # Lines can be inserted so that it looks like they were in the input
3200 # file at the place it was when this routine is called. See also
3201 # insert_lines(). Lines inserted via this routine are already fully
3202 # adjusted, ready to be processed; each_line_handler()s handlers will
3203 # not be called. This means this is not a completely general
3204 # facility, as only the last each_line_handler on the stack should
3205 # call this. It could be made more general, by passing to each of the
3206 # line_handlers their position on the stack, which they would pass on
3207 # to this routine, and that would replace the boolean first element in
3208 # the anonymous array pushed here, so that the next_line routine could
3209 # use that to call only those handlers whose index is after it on the
3210 # stack. But this is overkill for what is needed now.
3211
3212 my $self = shift;
3213 trace $_[0] if main::DEBUG && $to_trace;
3214
3215 # Each inserted line is an array, with the first element being 1 to
3216 # indicate that this line has been adjusted
f998e60c 3217 no overloading;
051df77b 3218 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
3219 return;
3220 }
3221
3222 sub get_missings {
3223 # Returns the stored up @missings lines' values, and clears the list.
3224 # The values are in an array, consisting of the default in the first
3225 # element, and the property in the 2nd. However, since these lines
3226 # can be stacked up, the return is an array of all these arrays.
3227
3228 my $self = shift;
3229 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3230
ffe43484 3231 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3232
3233 # If not accepting a list return, just return the first one.
3234 return shift @{$missings{$addr}} unless wantarray;
3235
3236 my @return = @{$missings{$addr}};
3237 undef @{$missings{$addr}};
3238 return @return;
3239 }
3240
a0204bb1
KW
3241 sub _exclude_unassigned {
3242
3243 # Takes the range in $_ and excludes code points that aren't assigned
3244 # in this release
3245
3246 state $skip_inserted_count = 0;
3247
3248 # Ignore recursive calls.
3249 if ($skip_inserted_count) {
3250 $skip_inserted_count--;
3251 return;
3252 }
3253
3254 # Find what code points are assigned in this release
3255 main::calculate_Assigned() if ! defined $Assigned;
3256
3257 my $self = shift;
3258 my $addr = do { no overloading; pack 'J', $self; };
3259 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3260
3261 my ($range, @remainder)
3262 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3263
3264 # Examine the range.
3265 if ($range =~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3266 {
3267 my $low = hex $1;
3268 my $high = (defined $2) ? hex $2 : $low;
3269
3270 # Split the range into subranges of just those code points in it
3271 # that are assigned.
3272 my @ranges = (Range_List->new(Initialize
3273 => Range->new($low, $high)) & $Assigned)->ranges;
3274
3275 # Do nothing if nothing in the original range is assigned in this
3276 # release; handle normally if everything is in this release.
3277 if (! @ranges) {
3278 $_ = "";
3279 }
3280 elsif (@ranges != 1) {
3281
3282 # Here, some code points in the original range aren't in this
3283 # release; @ranges gives the ones that are. Create fake input
3284 # lines for each of the ranges, and set things up so that when
3285 # this routine is called on that fake input, it will do
3286 # nothing.
3287 $skip_inserted_count = @ranges;
3288 my $remainder = join ";", @remainder;
3289 for my $range (@ranges) {
3290 $self->insert_lines(sprintf("%04X..%04X;%s",
3291 $range->start, $range->end, $remainder));
3292 }
3293 $_ = ""; # The original range is now defunct.
3294 }
3295 }
3296
3297 return;
3298 }
3299
3300 sub _fixup_obsolete_hanguls {
3301
3302 # This is called only when compiling Unicode version 1. All Unicode
3303 # data for subsequent releases assumes that the code points that were
3304 # Hangul syllables in this release only are something else, so if
3305 # using such data, we have to override it
3306
3307 my $self = shift;
3308 my $addr = do { no overloading; pack 'J', $self; };
3309 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3310
3311 my $object = main::property_ref($property{$addr});
53adf6a2
KW
3312 $object->add_map($FIRST_REMOVED_HANGUL_SYLLABLE,
3313 $FINAL_REMOVED_HANGUL_SYLLABLE,
a0204bb1
KW
3314 $early{$addr}[3], # Passed-in value for these
3315 Replace => $UNCONDITIONALLY);
3316 }
3317
99870f4d
KW
3318 sub _insert_property_into_line {
3319 # Add a property field to $_, if this file requires it.
3320
f998e60c 3321 my $self = shift;
ffe43484 3322 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 3323 my $property = $property{$addr};
99870f4d
KW
3324 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3325
3326 $_ =~ s/(;|$)/; $property$1/;
3327 return;
3328 }
3329
3330 sub carp_bad_line {
3331 # Output consistent error messages, using either a generic one, or the
3332 # one given by the optional parameter. To avoid gazillions of the
3333 # same message in case the syntax of a file is way off, this routine
3334 # only outputs the first instance of each message, incrementing a
3335 # count so the totals can be output at the end of the file.
3336
3337 my $self = shift;
3338 my $message = shift;
3339 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3340
ffe43484 3341 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3342
3343 $message = 'Unexpected line' unless $message;
3344
3345 # No trailing punctuation so as to fit with our addenda.
3346 $message =~ s/[.:;,]$//;
3347
3348 # If haven't seen this exact message before, output it now. Otherwise
3349 # increment the count of how many times it has occurred
3350 unless ($errors{$addr}->{$message}) {
3351 Carp::my_carp("$message in '$_' in "
f998e60c 3352 . $file{$addr}
99870f4d
KW
3353 . " at line $.. Skipping this line;");
3354 $errors{$addr}->{$message} = 1;
3355 }
3356 else {
3357 $errors{$addr}->{$message}++;
3358 }
3359
3360 # Clear the line to prevent any further (meaningful) processing of it.
3361 $_ = "";
3362
3363 return;
3364 }
3365} # End closure
3366
3367package Multi_Default;
3368
3369# Certain properties in early versions of Unicode had more than one possible
3370# default for code points missing from the files. In these cases, one
3371# default applies to everything left over after all the others are applied,
3372# and for each of the others, there is a description of which class of code
3373# points applies to it. This object helps implement this by storing the
3374# defaults, and for all but that final default, an eval string that generates
3375# the class that it applies to.
3376
3377
3378{ # Closure
3379
3380 main::setup_package();
3381
3382 my %class_defaults;
3383 # The defaults structure for the classes
3384 main::set_access('class_defaults', \%class_defaults);
3385
3386 my %other_default;
3387 # The default that applies to everything left over.
3388 main::set_access('other_default', \%other_default, 'r');
3389
3390
3391 sub new {
3392 # The constructor is called with default => eval pairs, terminated by
3393 # the left-over default. e.g.
3394 # Multi_Default->new(
3395 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
3396 # - 0x200D',
3397 # 'R' => 'some other expression that evaluates to code points',
3398 # .
3399 # .
3400 # .
3401 # 'U'));
68c3923f
KW
3402 # It is best to leave the final value be the one that matches the
3403 # above-Unicode code points.
99870f4d
KW
3404
3405 my $class = shift;
3406
3407 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 3408 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3409
3410 while (@_ > 1) {
3411 my $default = shift;
3412 my $eval = shift;
3413 $class_defaults{$addr}->{$default} = $eval;
3414 }
3415
3416 $other_default{$addr} = shift;
3417
3418 return $self;
3419 }
3420
3421 sub get_next_defaults {
3422 # Iterates and returns the next class of defaults.
3423 my $self = shift;
3424 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3425
ffe43484 3426 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3427
3428 return each %{$class_defaults{$addr}};
3429 }
3430}
3431
3432package Alias;
3433
3434# An alias is one of the names that a table goes by. This class defines them
3435# including some attributes. Everything is currently setup in the
3436# constructor.
3437
3438
3439{ # Closure
3440
3441 main::setup_package();
3442
3443 my %name;
3444 main::set_access('name', \%name, 'r');
3445
3446 my %loose_match;
c12f2655 3447 # Should this name match loosely or not.
99870f4d
KW
3448 main::set_access('loose_match', \%loose_match, 'r');
3449
33e96e72
KW
3450 my %make_re_pod_entry;
3451 # Some aliases should not get their own entries in the re section of the
3452 # pod, because they are covered by a wild-card, and some we want to
3453 # discourage use of. Binary
f82fe4ba 3454 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r', 's');
99870f4d 3455
fd1e3e84
KW
3456 my %ucd;
3457 # Is this documented to be accessible via Unicode::UCD
3458 main::set_access('ucd', \%ucd, 'r', 's');
3459
99870f4d
KW
3460 my %status;
3461 # Aliases have a status, like deprecated, or even suppressed (which means
3462 # they don't appear in documentation). Enum
3463 main::set_access('status', \%status, 'r');
3464
0eac1e20 3465 my %ok_as_filename;
99870f4d
KW
3466 # Similarly, some aliases should not be considered as usable ones for
3467 # external use, such as file names, or we don't want documentation to
3468 # recommend them. Boolean
0eac1e20 3469 main::set_access('ok_as_filename', \%ok_as_filename, 'r');
99870f4d
KW
3470
3471 sub new {
3472 my $class = shift;
3473
3474 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3475 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3476
3477 $name{$addr} = shift;
3478 $loose_match{$addr} = shift;
33e96e72 3479 $make_re_pod_entry{$addr} = shift;
0eac1e20 3480 $ok_as_filename{$addr} = shift;
99870f4d 3481 $status{$addr} = shift;
fd1e3e84 3482 $ucd{$addr} = shift;
99870f4d
KW
3483
3484 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3485
3486 # Null names are never ok externally
0eac1e20 3487 $ok_as_filename{$addr} = 0 if $name{$addr} eq "";
99870f4d
KW
3488
3489 return $self;
3490 }
3491}
3492
3493package Range;
3494
3495# A range is the basic unit for storing code points, and is described in the
3496# comments at the beginning of the program. Each range has a starting code
3497# point; an ending code point (not less than the starting one); a value
3498# that applies to every code point in between the two end-points, inclusive;
3499# and an enum type that applies to the value. The type is for the user's
3500# convenience, and has no meaning here, except that a non-zero type is
3501# considered to not obey the normal Unicode rules for having standard forms.
3502#
3503# The same structure is used for both map and match tables, even though in the
3504# latter, the value (and hence type) is irrelevant and could be used as a
3505# comment. In map tables, the value is what all the code points in the range
3506# map to. Type 0 values have the standardized version of the value stored as
3507# well, so as to not have to recalculate it a lot.
3508
3509sub trace { return main::trace(@_); }
3510
3511{ # Closure
3512
3513 main::setup_package();
3514
3515 my %start;
3516 main::set_access('start', \%start, 'r', 's');
3517
3518 my %end;
3519 main::set_access('end', \%end, 'r', 's');
3520
3521 my %value;
3522 main::set_access('value', \%value, 'r');
3523
3524 my %type;
3525 main::set_access('type', \%type, 'r');
3526
3527 my %standard_form;
3528 # The value in internal standard form. Defined only if the type is 0.
3529 main::set_access('standard_form', \%standard_form);
3530
3531 # Note that if these fields change, the dump() method should as well
3532
3533 sub new {
3534 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
3535 my $class = shift;
3536
3537 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3538 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3539
3540 $start{$addr} = shift;
3541 $end{$addr} = shift;
3542
3543 my %args = @_;
3544
3545 my $value = delete $args{'Value'}; # Can be 0
3546 $value = "" unless defined $value;
3547 $value{$addr} = $value;
3548
3549 $type{$addr} = delete $args{'Type'} || 0;
3550
3551 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3552
99870f4d
KW
3553 return $self;
3554 }
3555
3556 use overload
3557 fallback => 0,
3558 qw("") => "_operator_stringify",
3559 "." => \&main::_operator_dot,
1285127e 3560 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3561 ;
3562
3563 sub _operator_stringify {
3564 my $self = shift;
ffe43484 3565 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3566
3567 # Output it like '0041..0065 (value)'
3568 my $return = sprintf("%04X", $start{$addr})
3569 . '..'
3570 . sprintf("%04X", $end{$addr});
3571 my $value = $value{$addr};
3572 my $type = $type{$addr};
3573 $return .= ' (';
3574 $return .= "$value";
3575 $return .= ", Type=$type" if $type != 0;
3576 $return .= ')';
3577
3578 return $return;
3579 }
3580
3581 sub standard_form {
c292d35a
NC
3582 # Calculate the standard form only if needed, and cache the result.
3583 # The standard form is the value itself if the type is special.
3584 # This represents a considerable CPU and memory saving - at the time
3585 # of writing there are 368676 non-special objects, but the standard
3586 # form is only requested for 22047 of them - ie about 6%.
99870f4d
KW
3587
3588 my $self = shift;
3589 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3590
ffe43484 3591 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3592
3593 return $standard_form{$addr} if defined $standard_form{$addr};
c292d35a
NC
3594
3595 my $value = $value{$addr};
3596 return $value if $type{$addr};
3597 return $standard_form{$addr} = main::standardize($value);
99870f4d
KW
3598 }
3599
3600 sub dump {
3601 # Human, not machine readable. For machine readable, comment out this
3602 # entire routine and let the standard one take effect.
3603 my $self = shift;
3604 my $indent = shift;
3605 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3606
ffe43484 3607 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3608
3609 my $return = $indent
3610 . sprintf("%04X", $start{$addr})
3611 . '..'
3612 . sprintf("%04X", $end{$addr})
3613 . " '$value{$addr}';";
3614 if (! defined $standard_form{$addr}) {
3615 $return .= "(type=$type{$addr})";
3616 }
3617 elsif ($standard_form{$addr} ne $value{$addr}) {
3618 $return .= "(standard '$standard_form{$addr}')";
3619 }
3620 return $return;
3621 }
3622} # End closure
3623
3624package _Range_List_Base;
3625
3626# Base class for range lists. A range list is simply an ordered list of
3627# ranges, so that the ranges with the lowest starting numbers are first in it.
3628#
3629# When a new range is added that is adjacent to an existing range that has the
3630# same value and type, it merges with it to form a larger range.
3631#
3632# Ranges generally do not overlap, except that there can be multiple entries
3633# of single code point ranges. This is because of NameAliases.txt.
3634#
3635# In this program, there is a standard value such that if two different
3636# values, have the same standard value, they are considered equivalent. This
3637# value was chosen so that it gives correct results on Unicode data
3638
3639# There are a number of methods to manipulate range lists, and some operators
3640# are overloaded to handle them.
3641
99870f4d
KW
3642sub trace { return main::trace(@_); }
3643
3644{ # Closure
3645
3646 our $addr;
3647
5b348b71
KW
3648 # Max is initialized to a negative value that isn't adjacent to 0, for
3649 # simpler tests
3650 my $max_init = -2;
3651
99870f4d
KW
3652 main::setup_package();
3653
3654 my %ranges;
3655 # The list of ranges
3656 main::set_access('ranges', \%ranges, 'readable_array');
3657
3658 my %max;
3659 # The highest code point in the list. This was originally a method, but
3660 # actual measurements said it was used a lot.
3661 main::set_access('max', \%max, 'r');
3662
3663 my %each_range_iterator;
3664 # Iterator position for each_range()
3665 main::set_access('each_range_iterator', \%each_range_iterator);
3666
3667 my %owner_name_of;
3668 # Name of parent this is attached to, if any. Solely for better error
3669 # messages.
3670 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
3671
3672 my %_search_ranges_cache;
3673 # A cache of the previous result from _search_ranges(), for better
3674 # performance
3675 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
3676
3677 sub new {
3678 my $class = shift;
3679 my %args = @_;
3680
3681 # Optional initialization data for the range list.
3682 my $initialize = delete $args{'Initialize'};
3683
3684 my $self;
3685
3686 # Use _union() to initialize. _union() returns an object of this
3687 # class, which means that it will call this constructor recursively.
3688 # But it won't have this $initialize parameter so that it won't
3689 # infinitely loop on this.
3690 return _union($class, $initialize, %args) if defined $initialize;
3691
3692 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3693 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3694
3695 # Optional parent object, only for debug info.
3696 $owner_name_of{$addr} = delete $args{'Owner'};
3697 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
3