This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
pods: Discourage use of 'In' prefix for Unicode Block property
[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
cdcef19a 11my $start_time;
98dc9551 12BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
13 # get it as close as possible
14 $start_time= time;
15}
16
23e33b60 17require 5.010_001;
d73e5302 18use strict;
99870f4d 19use warnings;
cf25bb62 20use Carp;
bd9ebcfd 21use Config;
99870f4d
KW
22use File::Find;
23use File::Path;
d07a55ed 24use File::Spec;
99870f4d 25use Text::Tabs;
6b64c11c 26use re "/aa";
0dee21d6 27use feature 'state';
99870f4d
KW
28
29sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 30my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d 31
74cd47d0
KW
32sub NON_ASCII_PLATFORM { ord("A") != 65 }
33
99870f4d
KW
34##########################################################################
35#
36# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
37# from the Unicode database files (lib/unicore/.../*.txt), It also generates
232ed87f 38# a pod file and .t files, depending on option parameters.
99870f4d
KW
39#
40# The structure of this file is:
41# First these introductory comments; then
42# code needed for everywhere, such as debugging stuff; then
43# code to handle input parameters; then
44# data structures likely to be of external interest (some of which depend on
45# the input parameters, so follows them; then
46# more data structures and subroutine and package (class) definitions; then
47# the small actual loop to process the input files and finish up; then
48# a __DATA__ section, for the .t tests
49#
232ed87f
KW
50# This program works on all releases of Unicode so far. The outputs have been
51# scrutinized most intently for release 5.1. The others have been checked for
52# somewhat more than just sanity. It can handle all non-provisional Unicode
53# character properties in those releases.
99870f4d 54#
99870f4d
KW
55# This program is mostly about Unicode character (or code point) properties.
56# A property describes some attribute or quality of a code point, like if it
57# is lowercase or not, its name, what version of Unicode it was first defined
58# in, or what its uppercase equivalent is. Unicode deals with these disparate
59# possibilities by making all properties into mappings from each code point
60# into some corresponding value. In the case of it being lowercase or not,
61# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
62# property maps each Unicode code point to a single value, called a "property
232ed87f
KW
63# value". (Some more recently defined properties, map a code point to a set
64# of values.)
99870f4d
KW
65#
66# When using a property in a regular expression, what is desired isn't the
67# mapping of the code point to its property's value, but the reverse (or the
68# mathematical "inverse relation"): starting with the property value, "Does a
69# code point map to it?" These are written in a "compound" form:
70# \p{property=value}, e.g., \p{category=punctuation}. This program generates
71# files containing the lists of code points that map to each such regular
72# expression property value, one file per list
73#
74# There is also a single form shortcut that Perl adds for many of the commonly
75# used properties. This happens for all binary properties, plus script,
76# general_category, and block properties.
77#
78# Thus the outputs of this program are files. There are map files, mostly in
79# the 'To' directory; and there are list files for use in regular expression
80# matching, all in subdirectories of the 'lib' directory, with each
81# subdirectory being named for the property that the lists in it are for.
82# Bookkeeping, test, and documentation files are also generated.
83
84my $matches_directory = 'lib'; # Where match (\p{}) files go.
85my $map_directory = 'To'; # Where map files go.
86
87# DATA STRUCTURES
88#
89# The major data structures of this program are Property, of course, but also
90# Table. There are two kinds of tables, very similar to each other.
91# "Match_Table" is the data structure giving the list of code points that have
92# a particular property value, mentioned above. There is also a "Map_Table"
93# data structure which gives the property's mapping from code point to value.
94# There are two structures because the match tables need to be combined in
95# various ways, such as constructing unions, intersections, complements, etc.,
96# and the map ones don't. And there would be problems, perhaps subtle, if
97# a map table were inadvertently operated on in some of those ways.
98# The use of separate classes with operations defined on one but not the other
99# prevents accidentally confusing the two.
100#
101# At the heart of each table's data structure is a "Range_List", which is just
102# an ordered list of "Ranges", plus ancillary information, and methods to
103# operate on them. A Range is a compact way to store property information.
104# Each range has a starting code point, an ending code point, and a value that
105# is meant to apply to all the code points between the two end points,
106# inclusive. For a map table, this value is the property value for those
107# code points. Two such ranges could be written like this:
108# 0x41 .. 0x5A, 'Upper',
109# 0x61 .. 0x7A, 'Lower'
110#
111# Each range also has a type used as a convenience to classify the values.
112# Most ranges in this program will be Type 0, or normal, but there are some
113# ranges that have a non-zero type. These are used only in map tables, and
114# are for mappings that don't fit into the normal scheme of things. Mappings
115# that require a hash entry to communicate with utf8.c are one example;
116# another example is mappings for charnames.pm to use which indicate a name
232ed87f 117# that is algorithmically determinable from its code point (and the reverse).
99870f4d
KW
118# These are used to significantly compact these tables, instead of listing
119# each one of the tens of thousands individually.
120#
121# In a match table, the value of a range is irrelevant (and hence the type as
122# well, which will always be 0), and arbitrarily set to the null string.
123# Using the example above, there would be two match tables for those two
124# entries, one named Upper would contain the 0x41..0x5A range, and the other
125# named Lower would contain 0x61..0x7A.
126#
127# Actually, there are two types of range lists, "Range_Map" is the one
128# associated with map tables, and "Range_List" with match tables.
232ed87f
KW
129# Again, this is so that methods can be defined on one and not the others so
130# as to prevent operating on them in incorrect ways.
99870f4d
KW
131#
132# Eventually, most tables are written out to files to be read by utf8_heavy.pl
133# in the perl core. All tables could in theory be written, but some are
134# suppressed because there is no current practical use for them. It is easy
135# to change which get written by changing various lists that are near the top
136# of the actual code in this file. The table data structures contain enough
137# ancillary information to allow them to be treated as separate entities for
138# writing, such as the path to each one's file. There is a heading in each
139# map table that gives the format of its entries, and what the map is for all
140# the code points missing from it. (This allows tables to be more compact.)
678f13d5 141#
99870f4d
KW
142# The Property data structure contains one or more tables. All properties
143# contain a map table (except the $perl property which is a
144# pseudo-property containing only match tables), and any properties that
145# are usable in regular expression matches also contain various matching
146# tables, one for each value the property can have. A binary property can
147# have two values, True and False (or Y and N, which are preferred by Unicode
148# terminology). Thus each of these properties will have a map table that
149# takes every code point and maps it to Y or N (but having ranges cuts the
150# number of entries in that table way down), and two match tables, one
151# which has a list of all the code points that map to Y, and one for all the
232ed87f 152# code points that map to N. (For each binary property, a third table is also
99870f4d 153# generated for the pseudo Perl property. It contains the identical code
232ed87f
KW
154# points as the Y table, but can be written in regular expressions, not in the
155# compound form, but in a "single" form like \p{IsUppercase}.) Many
156# properties are binary, but some properties have several possible values,
157# some have many, and properties like Name have a different value for every
158# named code point. Those will not, unless the controlling lists are changed,
159# have their match tables written out. But all the ones which can be used in
160# regular expression \p{} and \P{} constructs will. Prior to 5.14, generally
161# a property would have either its map table or its match tables written but
162# not both. Again, what gets written is controlled by lists which can easily
163# be changed. Starting in 5.14, advantage was taken of this, and all the map
164# tables needed to reconstruct the Unicode db are now written out, while
165# suppressing the Unicode .txt files that contain the data. Our tables are
166# much more compact than the .txt files, so a significant space savings was
167# achieved. Also, tables are not written out that are trivially derivable
168# from tables that do get written. So, there typically is no file containing
169# the code points not matched by a binary property (the table for \P{} versus
170# lowercase \p{}), since you just need to invert the True table to get the
171# False table.
172
173# Properties have a 'Type', like 'binary', or 'string', or 'enum' depending on
174# how many match tables there are and the content of the maps. This 'Type' is
c12f2655
KW
175# different than a range 'Type', so don't get confused by the two concepts
176# having the same name.
678f13d5 177#
99870f4d
KW
178# For information about the Unicode properties, see Unicode's UAX44 document:
179
180my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
181
182# As stated earlier, this program will work on any release of Unicode so far.
183# Most obvious problems in earlier data have NOT been corrected except when
be864b6c 184# necessary to make Perl or this program work reasonably, and to keep out
232ed87f
KW
185# potential security issues. For example, no folding information was given in
186# early releases, so this program substitutes lower case instead, just so that
187# a regular expression with the /i option will do something that actually
188# gives the right results in many cases. There are also a couple other
189# corrections for version 1.1.5, commented at the point they are made. As an
190# example of corrections that weren't made (but could be) is this statement
191# from DerivedAge.txt: "The supplementary private use code points and the
192# non-character code points were assigned in version 2.0, but not specifically
193# listed in the UCD until versions 3.0 and 3.1 respectively." (To be precise
194# it was 3.0.1 not 3.0.0) More information on Unicode version glitches is
195# further down in these introductory comments.
99870f4d 196#
232ed87f
KW
197# This program works on all non-provisional properties as of the current
198# Unicode release, though the files for some are suppressed for various
199# reasons. You can change which are output by changing lists in this program.
678f13d5 200#
dc85bd38 201# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
202# loose matchings rules (from Unicode TR18):
203#
204# The recommended names for UCD properties and property values are in
205# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
206# [PropValue]. There are both abbreviated names and longer, more
207# descriptive names. It is strongly recommended that both names be
208# recognized, and that loose matching of property names be used,
209# whereby the case distinctions, whitespace, hyphens, and underbar
210# are ignored.
232ed87f 211#
99870f4d
KW
212# The program still allows Fuzzy to override its determination of if loose
213# matching should be used, but it isn't currently used, as it is no longer
214# needed; the calculations it makes are good enough.
678f13d5 215#
99870f4d
KW
216# SUMMARY OF HOW IT WORKS:
217#
218# Process arguments
219#
220# A list is constructed containing each input file that is to be processed
221#
222# Each file on the list is processed in a loop, using the associated handler
223# code for each:
224# The PropertyAliases.txt and PropValueAliases.txt files are processed
225# first. These files name the properties and property values.
226# Objects are created of all the property and property value names
227# that the rest of the input should expect, including all synonyms.
228# The other input files give mappings from properties to property
229# values. That is, they list code points and say what the mapping
230# is under the given property. Some files give the mappings for
231# just one property; and some for many. This program goes through
232ed87f
KW
232# each file and populates the properties and their map tables from
233# them. Some properties are listed in more than one file, and
234# Unicode has set up a precedence as to which has priority if there
235# is a conflict. Thus the order of processing matters, and this
236# program handles the conflict possibility by processing the
237# overriding input files last, so that if necessary they replace
238# earlier values.
99870f4d
KW
239# After this is all done, the program creates the property mappings not
240# furnished by Unicode, but derivable from what it does give.
241# The tables of code points that match each property value in each
242# property that is accessible by regular expressions are created.
243# The Perl-defined properties are created and populated. Many of these
244# require data determined from the earlier steps
245# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 246# and Unicode are reconciled and warned about.
99870f4d
KW
247# All the properties are written to files
248# Any other files are written, and final warnings issued.
678f13d5 249#
99870f4d
KW
250# For clarity, a number of operators have been overloaded to work on tables:
251# ~ means invert (take all characters not in the set). The more
252# conventional '!' is not used because of the possibility of confusing
253# it with the actual boolean operation.
254# + means union
255# - means subtraction
256# & means intersection
257# The precedence of these is the order listed. Parentheses should be
258# copiously used. These are not a general scheme. The operations aren't
259# defined for a number of things, deliberately, to avoid getting into trouble.
260# Operations are done on references and affect the underlying structures, so
261# that the copy constructors for them have been overloaded to not return a new
262# clone, but the input object itself.
678f13d5 263#
99870f4d
KW
264# The bool operator is deliberately not overloaded to avoid confusion with
265# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
266#
267# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
268#
269# This program needs to be able to run under miniperl. Therefore, it uses a
270# minimum of other modules, and hence implements some things itself that could
271# be gotten from CPAN
272#
273# This program uses inputs published by the Unicode Consortium. These can
274# change incompatibly between releases without the Perl maintainers realizing
275# it. Therefore this program is now designed to try to flag these. It looks
276# at the directories where the inputs are, and flags any unrecognized files.
277# It keeps track of all the properties in the files it handles, and flags any
278# that it doesn't know how to handle. It also flags any input lines that
279# don't match the expected syntax, among other checks.
280#
281# It is also designed so if a new input file matches one of the known
282# templates, one hopefully just needs to add it to a list to have it
283# processed.
284#
285# As mentioned earlier, some properties are given in more than one file. In
286# particular, the files in the extracted directory are supposedly just
287# reformattings of the others. But they contain information not easily
9be22ca5
KW
288# derivable from the other files, including results for Unihan (which isn't
289# usually available to this program) and for unassigned code points. They
678f13d5
KW
290# also have historically had errors or been incomplete. In an attempt to
291# create the best possible data, this program thus processes them first to
292# glean information missing from the other files; then processes those other
293# files to override any errors in the extracted ones. Much of the design was
294# driven by this need to store things and then possibly override them.
295#
296# It tries to keep fatal errors to a minimum, to generate something usable for
297# testing purposes. It always looks for files that could be inputs, and will
298# warn about any that it doesn't know how to handle (the -q option suppresses
299# the warning).
99870f4d 300#
678f13d5
KW
301# Why is there more than one type of range?
302# This simplified things. There are some very specialized code points that
303# have to be handled specially for output, such as Hangul syllable names.
304# By creating a range type (done late in the development process), it
305# allowed this to be stored with the range, and overridden by other input.
306# Originally these were stored in another data structure, and it became a
307# mess trying to decide if a second file that was for the same property was
308# overriding the earlier one or not.
309#
310# Why are there two kinds of tables, match and map?
311# (And there is a base class shared by the two as well.) As stated above,
312# they actually are for different things. Development proceeded much more
313# smoothly when I (khw) realized the distinction. Map tables are used to
314# give the property value for every code point (actually every code point
315# that doesn't map to a default value). Match tables are used for regular
316# expression matches, and are essentially the inverse mapping. Separating
317# the two allows more specialized methods, and error checks so that one
318# can't just take the intersection of two map tables, for example, as that
319# is nonsensical.
99870f4d 320#
232ed87f
KW
321# What about 'fate' and 'status'. The concept of a table's fate was created
322# late when it became clear that something more was needed. The difference
323# between this and 'status' is unclean, and could be improved if someone
324# wanted to spend the effort.
325#
23e33b60
KW
326# DEBUGGING
327#
678f13d5
KW
328# This program is written so it will run under miniperl. Occasionally changes
329# will cause an error where the backtrace doesn't work well under miniperl.
330# To diagnose the problem, you can instead run it under regular perl, if you
331# have one compiled.
332#
333# There is a good trace facility. To enable it, first sub DEBUG must be set
334# to return true. Then a line like
335#
336# local $to_trace = 1 if main::DEBUG;
337#
232ed87f
KW
338# can be added to enable tracing in its lexical scope (plus dynamic) or until
339# you insert another line:
678f13d5
KW
340#
341# local $to_trace = 0 if main::DEBUG;
342#
232ed87f 343# To actually trace, use a line like "trace $a, @b, %c, ...;
678f13d5
KW
344#
345# Some of the more complex subroutines already have trace statements in them.
346# Permanent trace statements should be like:
347#
348# trace ... if main::DEBUG && $to_trace;
349#
350# If there is just one or a few files that you're debugging, you can easily
351# cause most everything else to be skipped. Change the line
352#
353# my $debug_skip = 0;
354#
355# to 1, and every file whose object is in @input_file_objects and doesn't have
232ed87f
KW
356# a, 'non_skip => 1,' in its constructor will be skipped. However, skipping
357# Jamo.txt or UnicodeData.txt will likely cause fatal errors.
678f13d5 358#
b4a0206c 359# To compare the output tables, it may be useful to specify the -annotate
97a8a595
KW
360# flag. (As of this writing, this can't be done on a clean workspace, due to
361# requirements in Text::Tabs used in this option; so first run mktables
362# without this option.) This option adds comment lines to each table, one for
363# each non-algorithmically named character giving, currently its code point,
364# name, and graphic representation if printable (and you have a font that
365# knows about it). This makes it easier to see what the particular code
366# points are in each output table. Non-named code points are annotated with a
367# description of their status, and contiguous ones with the same description
368# will be output as a range rather than individually. Algorithmically named
369# characters are also output as ranges, except when there are just a few
370# contiguous ones.
c4019d52 371#
99870f4d
KW
372# FUTURE ISSUES
373#
374# The program would break if Unicode were to change its names so that
375# interior white space, underscores, or dashes differences were significant
376# within property and property value names.
377#
378# It might be easier to use the xml versions of the UCD if this program ever
379# would need heavy revision, and the ability to handle old versions was not
380# required.
381#
382# There is the potential for name collisions, in that Perl has chosen names
383# that Unicode could decide it also likes. There have been such collisions in
384# the past, with mostly Perl deciding to adopt the Unicode definition of the
385# name. However in the 5.2 Unicode beta testing, there were a number of such
386# collisions, which were withdrawn before the final release, because of Perl's
387# and other's protests. These all involved new properties which began with
388# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
389# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
390# Unicode document, so they are unlikely to be used by Unicode for another
391# purpose. However, they might try something beginning with 'In', or use any
392# of the other Perl-defined properties. This program will warn you of name
393# collisions, and refuse to generate tables with them, but manual intervention
394# will be required in this event. One scheme that could be implemented, if
395# necessary, would be to have this program generate another file, or add a
396# field to mktables.lst that gives the date of first definition of a property.
397# Each new release of Unicode would use that file as a basis for the next
398# iteration. And the Perl synonym addition code could sort based on the age
399# of the property, so older properties get priority, and newer ones that clash
400# would be refused; hence existing code would not be impacted, and some other
401# synonym would have to be used for the new property. This is ugly, and
402# manual intervention would certainly be easier to do in the short run; lets
403# hope it never comes to this.
678f13d5 404#
99870f4d
KW
405# A NOTE ON UNIHAN
406#
b9e57660
KW
407# This program can generate tables from the Unihan database. But that db
408# isn't normally available, so it is marked as optional. Prior to version
409# 5.2, this database was in a single file, Unihan.txt. In 5.2 the database
410# was split into 8 different files, all beginning with the letters 'Unihan'.
411# If you plunk those files down into the directory mktables ($0) is in, this
412# program will read them and automatically create tables for the properties
413# from it that are listed in PropertyAliases.txt and PropValueAliases.txt,
414# plus any you add to the @cjk_properties array and the @cjk_property_values
415# array, being sure to add necessary '# @missings' lines to the latter. For
416# Unicode versions earlier than 5.2, most of the Unihan properties are not
417# listed at all in PropertyAliases nor PropValueAliases. This program assumes
418# for these early releases that you want the properties that are specified in
419# the 5.2 release.
99870f4d
KW
420#
421# You may need to adjust the entries to suit your purposes. setup_unihan(),
422# and filter_unihan_line() are the functions where this is done. This program
423# already does some adjusting to make the lines look more like the rest of the
424# Unicode DB; You can see what that is in filter_unihan_line()
425#
426# There is a bug in the 3.2 data file in which some values for the
427# kPrimaryNumeric property have commas and an unexpected comment. A filter
9be22ca5
KW
428# could be added to correct these; or for a particular installation, the
429# Unihan.txt file could be edited to fix them.
99870f4d 430#
678f13d5
KW
431# HOW TO ADD A FILE TO BE PROCESSED
432#
433# A new file from Unicode needs to have an object constructed for it in
434# @input_file_objects, probably at the end or at the end of the extracted
435# ones. The program should warn you if its name will clash with others on
436# restrictive file systems, like DOS. If so, figure out a better name, and
437# add lines to the README.perl file giving that. If the file is a character
232ed87f 438# property, it should be in the format that Unicode has implicitly
678f13d5
KW
439# standardized for such files for the more recently introduced ones.
440# If so, the Input_file constructor for @input_file_objects can just be the
441# file name and release it first appeared in. If not, then it should be
442# possible to construct an each_line_handler() to massage the line into the
443# standardized form.
444#
445# For non-character properties, more code will be needed. You can look at
446# the existing entries for clues.
447#
448# UNICODE VERSIONS NOTES
449#
450# The Unicode UCD has had a number of errors in it over the versions. And
451# these remain, by policy, in the standard for that version. Therefore it is
452# risky to correct them, because code may be expecting the error. So this
453# program doesn't generally make changes, unless the error breaks the Perl
454# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
455# for U+1105, which causes real problems for the algorithms for Jamo
456# calculations, so it is changed here.
457#
458# But it isn't so clear cut as to what to do about concepts that are
459# introduced in a later release; should they extend back to earlier releases
460# where the concept just didn't exist? It was easier to do this than to not,
461# so that's what was done. For example, the default value for code points not
462# in the files for various properties was probably undefined until changed by
463# some version. No_Block for blocks is such an example. This program will
464# assign No_Block even in Unicode versions that didn't have it. This has the
465# benefit that code being written doesn't have to special case earlier
466# versions; and the detriment that it doesn't match the Standard precisely for
467# the affected versions.
468#
469# Here are some observations about some of the issues in early versions:
470#
232ed87f
KW
471# Prior to version 3.0, there were 3 character decompositions. These are not
472# handled by Unicode::Normalize, nor will it compile when presented a version
473# that has them. However, you can trivially get it to compile by simply
474# ignoring those decompositions, by changing the croak to a carp. At the time
9be22ca5
KW
475# of this writing, the line (in cpan/Unicode-Normalize/Normalize.pm or
476# cpan/Unicode-Normalize/mkheader) reads
232ed87f
KW
477#
478# croak("Weird Canonical Decomposition of U+$h");
479#
28807e1d 480# Simply comment it out. It will compile, but will not know about any three
9be22ca5 481# character decompositions.
232ed87f
KW
482
483# The number of code points in \p{alpha=True} halved in 2.1.9. It turns out
484# that the reason is that the CJK block starting at 4E00 was removed from
485# PropList, and was not put back in until 3.1.0. The Perl extension (the
486# single property name \p{alpha}) has the correct values. But the compound
487# form is simply not generated until 3.1, as it can be argued that prior to
488# this release, this was not an official property. The comments for
489# filter_old_style_proplist() give more details.
678f13d5
KW
490#
491# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
492# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
493# reason is that 3.2 introduced U+205F=medium math space, which was not
494# classed as white space, but Perl figured out that it should have been. 4.0
495# reclassified it correctly.
496#
497# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
232ed87f
KW
498# this was erroneously a synonym for 202 (it should be 200). In 4.0, ATB
499# became 202, and ATBL was left with no code points, as all the ones that
500# mapped to 202 stayed mapped to 202. Thus if your program used the numeric
501# name for the class, it would not have been affected, but if it used the
502# mnemonic, it would have been.
678f13d5 503#
9be22ca5 504# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that, code
678f13d5
KW
505# points which eventually came to have this script property value, instead
506# mapped to "Unknown". But in the next release all these code points were
507# moved to \p{sc=common} instead.
9be22ca5
KW
508
509# The tests furnished by Unicode for testing WordBreak and SentenceBreak
510# generate errors in 5.0 and earlier.
99870f4d
KW
511#
512# The default for missing code points for BidiClass is complicated. Starting
513# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
514# tries to do the best it can for earlier releases. It is done in
515# process_PropertyAliases()
516#
232ed87f
KW
517# In version 2.1.2, the entry in UnicodeData.txt:
518# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;;019F;
519# should instead be
520# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;019F;;019F
521# Without this change, there are casing problems for this character.
522#
7803ad2d
KW
523# Search for $string_compare_versions to see how to compare changes to
524# properties between Unicode versions
525#
99870f4d
KW
526##############################################################################
527
528my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
529 # and errors
530my $MAX_LINE_WIDTH = 78;
531
532# Debugging aid to skip most files so as to not be distracted by them when
533# concentrating on the ones being debugged. Add
534# non_skip => 1,
535# to the constructor for those files you want processed when you set this.
536# Files with a first version number of 0 are special: they are always
c12f2655
KW
537# processed regardless of the state of this flag. Generally, Jamo.txt and
538# UnicodeData.txt must not be skipped if you want this program to not die
539# before normal completion.
99870f4d
KW
540my $debug_skip = 0;
541
e9c4b4f8
KW
542
543# Normally these are suppressed.
544my $write_Unicode_deprecated_tables = 0;
545
99870f4d
KW
546# Set to 1 to enable tracing.
547our $to_trace = 0;
548
549{ # Closure for trace: debugging aid
550 my $print_caller = 1; # ? Include calling subroutine name
551 my $main_with_colon = 'main::';
552 my $main_colon_length = length($main_with_colon);
553
554 sub trace {
555 return unless $to_trace; # Do nothing if global flag not set
556
557 my @input = @_;
558
559 local $DB::trace = 0;
560 $DB::trace = 0; # Quiet 'used only once' message
561
562 my $line_number;
563
564 # Loop looking up the stack to get the first non-trace caller
565 my $caller_line;
566 my $caller_name;
567 my $i = 0;
568 do {
569 $line_number = $caller_line;
570 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
571 $caller = $main_with_colon unless defined $caller;
572
573 $caller_name = $caller;
574
575 # get rid of pkg
576 $caller_name =~ s/.*:://;
577 if (substr($caller_name, 0, $main_colon_length)
578 eq $main_with_colon)
579 {
580 $caller_name = substr($caller_name, $main_colon_length);
581 }
582
583 } until ($caller_name ne 'trace');
584
585 # If the stack was empty, we were called from the top level
586 $caller_name = 'main' if ($caller_name eq ""
587 || $caller_name eq 'trace');
588
589 my $output = "";
9346f59a 590 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
99870f4d 591 foreach my $string (@input) {
99870f4d
KW
592 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
593 $output .= simple_dumper($string);
594 }
595 else {
596 $string = "$string" if ref $string;
597 $string = $UNDEF unless defined $string;
598 chomp $string;
599 $string = '""' if $string eq "";
600 $output .= " " if $output ne ""
601 && $string ne ""
602 && substr($output, -1, 1) ne " "
603 && substr($string, 0, 1) ne " ";
604 $output .= $string;
605 }
606 }
607
99f78760
KW
608 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
609 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
610 print STDERR $output, "\n";
611 return;
612 }
613}
614
615# This is for a rarely used development feature that allows you to compare two
616# versions of the Unicode standard without having to deal with changes caused
9be22ca5 617# by the code points introduced in the later version. You probably also want
8f21889c
KW
618# to use the -annotate option when using this. Run this program on a unicore
619# containing the starting release you want to compare. Save that output
620# structrue. Then, switching to a unicore with the ending release, change the
621# 0 in the $string_compare_versions definition just below to a string
622# containing a SINGLE dotted Unicode release number (e.g. "2.1") corresponding
623# to the starting release. This program will then compile, but throw away all
624# code points introduced after the starting release. Finally use a diff tool
625# to compare the two directory structures. They include only the code points
626# common to both releases, and you can see the changes caused just by the
627# underlying release semantic changes. For versions earlier than 3.2, you
628# must copy a version of DAge.txt into the directory.
c12f2655 629my $string_compare_versions = DEBUG && 0; # e.g., "2.1";
99870f4d
KW
630my $compare_versions = DEBUG
631 && $string_compare_versions
632 && pack "C*", split /\./, $string_compare_versions;
633
634sub uniques {
635 # Returns non-duplicated input values. From "Perl Best Practices:
636 # Encapsulated Cleverness". p. 455 in first edition.
637
638 my %seen;
0e407844
NC
639 # Arguably this breaks encapsulation, if the goal is to permit multiple
640 # distinct objects to stringify to the same value, and be interchangeable.
641 # However, for this program, no two objects stringify identically, and all
642 # lists passed to this function are either objects or strings. So this
643 # doesn't affect correctness, but it does give a couple of percent speedup.
644 no overloading;
99870f4d
KW
645 return grep { ! $seen{$_}++ } @_;
646}
647
648$0 = File::Spec->canonpath($0);
649
650my $make_test_script = 0; # ? Should we output a test script
6b5ab373 651my $make_norm_test_script = 0; # ? Should we output a normalization test script
99870f4d
KW
652my $write_unchanged_files = 0; # ? Should we update the output files even if
653 # we don't think they have changed
654my $use_directory = ""; # ? Should we chdir somewhere.
655my $pod_directory; # input directory to store the pod file.
656my $pod_file = 'perluniprops';
657my $t_path; # Path to the .t test file
658my $file_list = 'mktables.lst'; # File to store input and output file names.
659 # This is used to speed up the build, by not
660 # executing the main body of the program if
661 # nothing on the list has changed since the
662 # previous build
663my $make_list = 1; # ? Should we write $file_list. Set to always
664 # make a list so that when the pumpking is
665 # preparing a release, s/he won't have to do
666 # special things
667my $glob_list = 0; # ? Should we try to include unknown .txt files
668 # in the input.
bd9ebcfd
KW
669my $output_range_counts = $debugging_build; # ? Should we include the number
670 # of code points in ranges in
671 # the output
558712cf 672my $annotate = 0; # ? Should character names be in the output
9ef2b94f 673
99870f4d
KW
674# Verbosity levels; 0 is quiet
675my $NORMAL_VERBOSITY = 1;
676my $PROGRESS = 2;
677my $VERBOSE = 3;
678
679my $verbosity = $NORMAL_VERBOSITY;
680
0458fbc1
KW
681# Stored in mktables.lst so that if this program is called with different
682# options, will regenerate even if the files otherwise look like they're
683# up-to-date.
684my $command_line_arguments = join " ", @ARGV;
685
99870f4d
KW
686# Process arguments
687while (@ARGV) {
cf25bb62
JH
688 my $arg = shift @ARGV;
689 if ($arg eq '-v') {
99870f4d
KW
690 $verbosity = $VERBOSE;
691 }
692 elsif ($arg eq '-p') {
693 $verbosity = $PROGRESS;
694 $| = 1; # Flush buffers as we go.
695 }
696 elsif ($arg eq '-q') {
697 $verbosity = 0;
698 }
699 elsif ($arg eq '-w') {
700 $write_unchanged_files = 1; # update the files even if havent changed
701 }
702 elsif ($arg eq '-check') {
6ae7e459
YO
703 my $this = shift @ARGV;
704 my $ok = shift @ARGV;
705 if ($this ne $ok) {
706 print "Skipping as check params are not the same.\n";
707 exit(0);
708 }
00a8df5c 709 }
99870f4d
KW
710 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
711 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
712 }
3df51b85
KW
713 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
714 {
99870f4d 715 $make_test_script = 1;
99870f4d 716 }
6b5ab373
KW
717 elsif ($arg eq '-makenormtest')
718 {
719 $make_norm_test_script = 1;
720 }
99870f4d
KW
721 elsif ($arg eq '-makelist') {
722 $make_list = 1;
723 }
724 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
725 -d $use_directory or croak "Unknown directory '$use_directory'";
726 }
727 elsif ($arg eq '-L') {
728
729 # Existence not tested until have chdir'd
730 $file_list = shift;
731 }
732 elsif ($arg eq '-globlist') {
733 $glob_list = 1;
734 }
735 elsif ($arg eq '-c') {
736 $output_range_counts = ! $output_range_counts
737 }
b4a0206c 738 elsif ($arg eq '-annotate') {
558712cf 739 $annotate = 1;
bd9ebcfd
KW
740 $debugging_build = 1;
741 $output_range_counts = 1;
9ef2b94f 742 }
99870f4d
KW
743 else {
744 my $with_c = 'with';
745 $with_c .= 'out' if $output_range_counts; # Complements the state
746 croak <<END;
747usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
748 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
749 [-check A B ]
750 -c : Output comments $with_c number of code points in ranges
751 -q : Quiet Mode: Only output serious warnings.
752 -p : Set verbosity level to normal plus show progress.
753 -v : Set Verbosity level high: Show progress and non-serious
754 warnings
755 -w : Write files regardless
756 -C dir : Change to this directory before proceeding. All relative paths
757 except those specified by the -P and -T options will be done
758 with respect to this directory.
759 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 760 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
761 -L filelist : Use alternate 'filelist' instead of standard one
762 -globlist : Take as input all non-Test *.txt files in current and sub
763 directories
3df51b85
KW
764 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
765 overrides -T
99870f4d 766 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 767 -annotate : Output an annotation for each character in the table files;
97a8a595
KW
768 useful for debugging mktables, looking at diffs; but is slow
769 and memory intensive
99870f4d
KW
770 -check A B : Executes $0 only if A and B are the same
771END
772 }
773}
774
775# Stores the most-recently changed file. If none have changed, can skip the
776# build
aeab6150 777my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
778
779# Change directories now, because need to read 'version' early.
780if ($use_directory) {
3df51b85 781 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
782 $pod_directory = File::Spec->rel2abs($pod_directory);
783 }
3df51b85 784 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 785 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 786 }
99870f4d 787 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 788 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 789 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 790 }
3df51b85 791 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 792 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 793 }
00a8df5c
YO
794}
795
99870f4d
KW
796# Get Unicode version into regular and v-string. This is done now because
797# various tables below get populated based on it. These tables are populated
798# here to be near the top of the file, and so easily seeable by those needing
799# to modify things.
800open my $VERSION, "<", "version"
801 or croak "$0: can't open required file 'version': $!\n";
802my $string_version = <$VERSION>;
803close $VERSION;
804chomp $string_version;
805my $v_version = pack "C*", split /\./, $string_version; # v string
806
8f21889c
KW
807my $unicode_version = ($compare_versions)
808 ? ( "$string_compare_versions (using "
809 . "$string_version rules)")
810 : $string_version;
811
99870f4d
KW
812# The following are the complete names of properties with property values that
813# are known to not match any code points in some versions of Unicode, but that
814# may change in the future so they should be matchable, hence an empty file is
815# generated for them.
a9c9e371
KW
816my @tables_that_may_be_empty;
817push @tables_that_may_be_empty, 'Joining_Type=Left_Joining'
818 if $v_version lt v6.3.0;
99870f4d
KW
819push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
820push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
821push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
822 if $v_version ge v4.1.0;
82aed44a
KW
823push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
824 if $v_version ge v6.0.0;
f583b44c
KW
825push @tables_that_may_be_empty, 'Grapheme_Cluster_Break=Prepend'
826 if $v_version ge v6.1.0;
1e958ea9
KW
827push @tables_that_may_be_empty, 'Canonical_Combining_Class=CCC133'
828 if $v_version ge v6.2.0;
99870f4d
KW
829
830# The lists below are hashes, so the key is the item in the list, and the
831# value is the reason why it is in the list. This makes generation of
832# documentation easier.
833
834my %why_suppressed; # No file generated for these.
835
836# Files aren't generated for empty extraneous properties. This is arguable.
837# Extraneous properties generally come about because a property is no longer
838# used in a newer version of Unicode. If we generated a file without code
839# points, programs that used to work on that property will still execute
840# without errors. It just won't ever match (or will always match, with \P{}).
841# This means that the logic is now likely wrong. I (khw) think its better to
842# find this out by getting an error message. Just move them to the table
843# above to change this behavior
844my %why_suppress_if_empty_warn_if_not = (
845
846 # It is the only property that has ever officially been removed from the
847 # Standard. The database never contained any code points for it.
848 'Special_Case_Condition' => 'Obsolete',
849
850 # Apparently never official, but there were code points in some versions of
851 # old-style PropList.txt
852 'Non_Break' => 'Obsolete',
853);
854
855# These would normally go in the warn table just above, but they were changed
856# a long time before this program was written, so warnings about them are
857# moot.
858if ($v_version gt v3.2.0) {
859 push @tables_that_may_be_empty,
860 'Canonical_Combining_Class=Attached_Below_Left'
861}
862
9be22ca5
KW
863# Enum values for to_output_map() method in the Map_Table package. (0 is don't
864# output)
272501f6
KW
865my $EXTERNAL_MAP = 1;
866my $INTERNAL_MAP = 2;
ce712c88 867my $OUTPUT_ADJUSTED = 3;
272501f6 868
fcf1973c
KW
869# To override computed values for writing the map tables for these properties.
870# The default for enum map tables is to write them out, so that the Unicode
871# .txt files can be removed, but all the data to compute any property value
872# for any code point is available in a more compact form.
873my %global_to_output_map = (
874 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
875 # get stuck supporting it if things change. Since it is a STRING
876 # property, it normally would be listed in the pod, but INTERNAL_MAP
877 # suppresses that.
fcf1973c
KW
878 Unicode_1_Name => $INTERNAL_MAP,
879
880 Present_In => 0, # Suppress, as easily computed from Age
74cd47d0
KW
881 Block => (NON_ASCII_PLATFORM) ? 1 : 0, # Suppress, as Blocks.txt is
882 # retained, but needed for
883 # non-ASCII
53d34b6c
KW
884
885 # Suppress, as mapping can be found instead from the
886 # Perl_Decomposition_Mapping file
887 Decomposition_Type => 0,
fcf1973c
KW
888);
889
99870f4d
KW
890# There are several types of obsolete properties defined by Unicode. These
891# must be hand-edited for every new Unicode release.
892my %why_deprecated; # Generates a deprecated warning message if used.
893my %why_stabilized; # Documentation only
894my %why_obsolete; # Documentation only
895
896{ # Closure
8364f9eb 897 my $simple = 'Perl uses the more complete version';
99870f4d
KW
898 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
899
900 my $other_properties = 'other properties';
901 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 902 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
903
904 %why_deprecated = (
5f7264c7 905 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
906 'Jamo_Short_Name' => $contributory,
907 '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',
908 'Other_Alphabetic' => $contributory,
909 'Other_Default_Ignorable_Code_Point' => $contributory,
910 'Other_Grapheme_Extend' => $contributory,
911 'Other_ID_Continue' => $contributory,
912 'Other_ID_Start' => $contributory,
913 'Other_Lowercase' => $contributory,
914 'Other_Math' => $contributory,
915 'Other_Uppercase' => $contributory,
e22aaf5c
KW
916 'Expands_On_NFC' => $why_no_expand,
917 'Expands_On_NFD' => $why_no_expand,
918 'Expands_On_NFKC' => $why_no_expand,
919 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
920 );
921
922 %why_suppressed = (
5f7264c7 923 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
924 # contains the same information, but without the algorithmically
925 # determinable Hangul syllables'. This file is not published, so it's
926 # existence is not noted in the comment.
12fee290 927 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize or prop_invmap() or charprop() in Unicode::UCD::',
99870f4d 928
5f8d1a89
KW
929 # Don't suppress ISO_Comment, as otherwise special handling is needed
930 # to differentiate between it and gc=c, which can be written as 'isc',
931 # which is the same characters as ISO_Comment's short name.
99870f4d 932
12fee290 933 'Name' => "Accessible via \\N{...} or 'use charnames;' or charprop() or prop_invmap() in Unicode::UCD::",
e0b29447 934
12fee290
KW
935 'Simple_Case_Folding' => "$simple. Can access this through casefold(), charprop(), or prop_invmap() in Unicode::UCD",
936 'Simple_Lowercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
937 'Simple_Titlecase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
938 'Simple_Uppercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
99870f4d 939
dac6f618 940 FC_NFKC_Closure => 'Deprecated by Unicode, and supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
941 );
942
1704a0ea
KW
943 foreach my $property (
944
945 # The following are suppressed because they were made contributory
946 # or deprecated by Unicode before Perl ever thought about
947 # supporting them.
948 'Jamo_Short_Name',
949 'Grapheme_Link',
950 'Expands_On_NFC',
951 'Expands_On_NFD',
952 'Expands_On_NFKC',
953 'Expands_On_NFKD',
954
955 # The following are suppressed because they have been marked
956 # as deprecated for a sufficient amount of time
957 'Other_Alphabetic',
958 'Other_Default_Ignorable_Code_Point',
959 'Other_Grapheme_Extend',
960 'Other_ID_Continue',
961 'Other_ID_Start',
962 'Other_Lowercase',
963 'Other_Math',
964 'Other_Uppercase',
e22aaf5c 965 ) {
99870f4d
KW
966 $why_suppressed{$property} = $why_deprecated{$property};
967 }
cf25bb62 968
99870f4d
KW
969 # Customize the message for all the 'Other_' properties
970 foreach my $property (keys %why_deprecated) {
971 next if (my $main_property = $property) !~ s/^Other_//;
972 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
973 }
974}
975
e9c4b4f8
KW
976if ($write_Unicode_deprecated_tables) {
977 foreach my $property (keys %why_suppressed) {
978 delete $why_suppressed{$property} if $property =~
979 / ^ Other | Grapheme /x;
980 }
981}
982
99870f4d
KW
983if ($v_version ge 4.0.0) {
984 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
985 if ($v_version ge 6.0.0) {
986 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
987 }
99870f4d 988}
5f7264c7 989if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 990 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 991 if ($v_version ge 6.0.0) {
63f74647 992 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 993 }
99870f4d
KW
994}
995
996# Probably obsolete forever
997if ($v_version ge v4.1.0) {
82aed44a
KW
998 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
999}
1000if ($v_version ge v6.0.0) {
caa75395 1001 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana" (or both)';
2b352efd 1002 $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
1003}
1004
1005# This program can create files for enumerated-like properties, such as
1006# 'Numeric_Type'. This file would be the same format as for a string
1007# property, with a mapping from code point to its value, so you could look up,
1008# for example, the script a code point is in. But no one so far wants this
1009# mapping, or they have found another way to get it since this is a new
1010# feature. So no file is generated except if it is in this list.
1011my @output_mapped_properties = split "\n", <<END;
1012END
1013
b9e57660
KW
1014# If you want more Unihan properties than the default, you need to add them to
1015# these arrays. Depending on the property type, @missing lines might have to
1016# be added to the second array. A sample entry would be (including the '#'):
1017# @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
99870f4d 1018my @cjk_properties = split "\n", <<'END';
99870f4d 1019END
99870f4d 1020my @cjk_property_values = split "\n", <<'END';
99870f4d
KW
1021END
1022
1023# The input files don't list every code point. Those not listed are to be
1024# defaulted to some value. Below are hard-coded what those values are for
1025# non-binary properties as of 5.1. Starting in 5.0, there are
caa75395 1026# machine-parsable comment lines in the files that give the defaults; so this
99870f4d
KW
1027# list shouldn't have to be extended. The claim is that all missing entries
1028# for binary properties will default to 'N'. Unicode tried to change that in
1029# 5.2, but the beta period produced enough protest that they backed off.
1030#
1031# The defaults for the fields that appear in UnicodeData.txt in this hash must
1032# be in the form that it expects. The others may be synonyms.
1033my $CODE_POINT = '<code point>';
1034my %default_mapping = (
1035 Age => "Unassigned",
1036 # Bidi_Class => Complicated; set in code
1037 Bidi_Mirroring_Glyph => "",
1038 Block => 'No_Block',
1039 Canonical_Combining_Class => 0,
1040 Case_Folding => $CODE_POINT,
1041 Decomposition_Mapping => $CODE_POINT,
1042 Decomposition_Type => 'None',
1043 East_Asian_Width => "Neutral",
1044 FC_NFKC_Closure => $CODE_POINT,
ec2cd995 1045 General_Category => ($v_version le 6.3.0) ? 'Cn' : 'Unassigned',
99870f4d
KW
1046 Grapheme_Cluster_Break => 'Other',
1047 Hangul_Syllable_Type => 'NA',
1048 ISO_Comment => "",
1049 Jamo_Short_Name => "",
1050 Joining_Group => "No_Joining_Group",
1051 # Joining_Type => Complicated; set in code
1052 kIICore => 'N', # Is converted to binary
1053 #Line_Break => Complicated; set in code
1054 Lowercase_Mapping => $CODE_POINT,
1055 Name => "",
1056 Name_Alias => "",
1057 NFC_QC => 'Yes',
1058 NFD_QC => 'Yes',
1059 NFKC_QC => 'Yes',
1060 NFKD_QC => 'Yes',
1061 Numeric_Type => 'None',
1062 Numeric_Value => 'NaN',
1063 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1064 Sentence_Break => 'Other',
1065 Simple_Case_Folding => $CODE_POINT,
1066 Simple_Lowercase_Mapping => $CODE_POINT,
1067 Simple_Titlecase_Mapping => $CODE_POINT,
1068 Simple_Uppercase_Mapping => $CODE_POINT,
1069 Titlecase_Mapping => $CODE_POINT,
1070 Unicode_1_Name => "",
1071 Unicode_Radical_Stroke => "",
1072 Uppercase_Mapping => $CODE_POINT,
1073 Word_Break => 'Other',
1074);
1075
678f13d5 1076### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1077
1078my $HEADER=<<"EOF";
1079# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85 1080# This file is machine-generated by $0 from the Unicode
8f21889c 1081# database, Version $unicode_version. Any changes made here will be lost!
cf25bb62
JH
1082EOF
1083
126c3d4e 1084my $INTERNAL_ONLY_HEADER = <<"EOF";
99870f4d
KW
1085
1086# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1087# This file is for internal use by core Perl only. The format and even the
1088# name or existence of this file are subject to change without notice. Don't
d9ae3878
KW
1089# use it directly. Use Unicode::UCD to access the Unicode character data
1090# base.
99870f4d
KW
1091EOF
1092
1093my $DEVELOPMENT_ONLY=<<"EOF";
1094# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1095# This file contains information artificially constrained to code points
1096# present in Unicode release $string_compare_versions.
1097# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1098# not be used for production.
b6922eda
KW
1099
1100EOF
1101
346a20cf
KW
1102my $MAX_UNICODE_CODEPOINT_STRING = ($v_version ge v2.0.0)
1103 ? "10FFFF"
1104 : "FFFF";
6189eadc
KW
1105my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1106my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d 1107
2d88a86a
KW
1108# We work with above-Unicode code points, up to UV_MAX. But when you get
1109# that high, above IV_MAX, some operations don't work, and you can easily get
1110# overflow. Therefore for internal use, we use a much smaller number,
1111# translating it to UV_MAX only for output. The exact number is immaterial
1112# (all Unicode code points are treated exactly the same), but the algorithm
1113# requires it to be at least 2 * $MAX_UNICODE_CODEPOINTS + 1;
1114my $MAX_WORKING_CODEPOINTS= $MAX_UNICODE_CODEPOINT * 8;
1115my $MAX_WORKING_CODEPOINT = $MAX_WORKING_CODEPOINTS - 1;
1116my $MAX_WORKING_CODEPOINT_STRING = sprintf("%X", $MAX_WORKING_CODEPOINT);
1117
1118my $MAX_PLATFORM_CODEPOINT = ~0;
1119
99870f4d
KW
1120# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1121# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1122# to decrease backtracking. The first regex allows the code point to be at
1123# the end of a word, but to work properly, the word shouldn't end with a valid
1124# hex character. The second one won't match a code point at the end of a
1125# word, and doesn't have the run-on issue
8c32d378
KW
1126my $run_on_code_point_re =
1127 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1128my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d
KW
1129
1130# This matches the beginning of the line in the Unicode db files that give the
1131# defaults for code points not listed (i.e., missing) in the file. The code
1132# depends on this ending with a semi-colon, so it can assume it is a valid
1133# field when the line is split() by semi-colons
346a20cf 1134my $missing_defaults_prefix = qr/^#\s+\@missing:\s+0000\.\.10FFFF\s*;/;
99870f4d
KW
1135
1136# Property types. Unicode has more types, but these are sufficient for our
1137# purposes.
1138my $UNKNOWN = -1; # initialized to illegal value
1139my $NON_STRING = 1; # Either binary or enum
1140my $BINARY = 2;
06f26c45
KW
1141my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1142 # tables, additional true and false tables are
1143 # generated so that false is anything matching the
1144 # default value, and true is everything else.
1145my $ENUM = 4; # Include catalog
1146my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1147
1148# Some input files have lines that give default values for code points not
1149# contained in the file. Sometimes these should be ignored.
1150my $NO_DEFAULTS = 0; # Must evaluate to false
1151my $NOT_IGNORED = 1;
1152my $IGNORED = 2;
1153
1154# Range types. Each range has a type. Most ranges are type 0, for normal,
1155# and will appear in the main body of the tables in the output files, but
1156# there are other types of ranges as well, listed below, that are specially
1157# handled. There are pseudo-types as well that will never be stored as a
1158# type, but will affect the calculation of the type.
1159
1160# 0 is for normal, non-specials
1161my $MULTI_CP = 1; # Sequence of more than code point
1162my $HANGUL_SYLLABLE = 2;
1163my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1164my $NULL = 4; # The map is to the null string; utf8.c can't
1165 # handle these, nor is there an accepted syntax
1166 # for them in \p{} constructs
f86864ac 1167my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1168 # otherwise be $MULTI_CP type are instead type 0
1169
1170# process_generic_property_file() can accept certain overrides in its input.
1171# Each of these must begin AND end with $CMD_DELIM.
1172my $CMD_DELIM = "\a";
1173my $REPLACE_CMD = 'replace'; # Override the Replace
1174my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1175
1176my $NO = 0;
1177my $YES = 1;
1178
1179# Values for the Replace argument to add_range.
1180# $NO # Don't replace; add only the code points not
1181 # already present.
1182my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1183 # the comments at the subroutine definition.
1184my $UNCONDITIONALLY = 2; # Replace without conditions.
9470941f 1185my $MULTIPLE_BEFORE = 4; # Don't replace, but add a duplicate record if
99870f4d 1186 # already there
7f4b1e25
KW
1187my $MULTIPLE_AFTER = 5; # Don't replace, but add a duplicate record if
1188 # already there
1189my $CROAK = 6; # Die with an error if is already there
99870f4d
KW
1190
1191# Flags to give property statuses. The phrases are to remind maintainers that
1192# if the flag is changed, the indefinite article referring to it in the
1193# documentation may need to be as well.
1194my $NORMAL = "";
99870f4d
KW
1195my $DEPRECATED = 'D';
1196my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1197my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1198my $DISCOURAGED = 'X';
1199my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1200my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1201my $STRICTER = 'T';
1202my $a_bold_stricter = "a 'B<$STRICTER>'";
1203my $A_bold_stricter = "A 'B<$STRICTER>'";
1204my $STABILIZED = 'S';
1205my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1206my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1207my $OBSOLETE = 'O';
1208my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1209my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1210
5c6d9947
KW
1211# Aliases can also have an extra status:
1212my $INTERNAL_ALIAS = 'P';
1213
99870f4d
KW
1214my %status_past_participles = (
1215 $DISCOURAGED => 'discouraged',
99870f4d
KW
1216 $STABILIZED => 'stabilized',
1217 $OBSOLETE => 'obsolete',
37e2e78e 1218 $DEPRECATED => 'deprecated',
5c6d9947 1219 $INTERNAL_ALIAS => 'reserved for Perl core internal use only',
99870f4d
KW
1220);
1221
395dfc19
KW
1222# Table fates. These are somewhat ordered, so that fates < $MAP_PROXIED should be
1223# externally documented.
301ba948 1224my $ORDINARY = 0; # The normal fate.
395dfc19
KW
1225my $MAP_PROXIED = 1; # The map table for the property isn't written out,
1226 # but there is a file written that can be used to
1227 # reconstruct this table
3cdaf629 1228my $INTERNAL_ONLY = 2; # The file for this table is written out, but it is
301ba948 1229 # for Perl's internal use only
277b7b16
KW
1230my $LEGACY_ONLY = 3; # Like $INTERNAL_ONLY, but not actually used by Perl.
1231 # Is for backwards compatibility for applications that
1232 # read the file directly, so it's format is
1233 # unchangeable.
1234my $SUPPRESSED = 4; # The file for this table is not written out, and as a
3cdaf629
KW
1235 # result, we don't bother to do many computations on
1236 # it.
277b7b16 1237my $PLACEHOLDER = 5; # Like $SUPPRESSED, but we go through all the
3cdaf629
KW
1238 # computations anyway, as the values are needed for
1239 # things to work. This happens when we have Perl
1240 # extensions that depend on Unicode tables that
1241 # wouldn't normally be in a given Unicode version.
301ba948 1242
f5817e0a
KW
1243# The format of the values of the tables:
1244my $EMPTY_FORMAT = "";
99870f4d
KW
1245my $BINARY_FORMAT = 'b';
1246my $DECIMAL_FORMAT = 'd';
1247my $FLOAT_FORMAT = 'f';
1248my $INTEGER_FORMAT = 'i';
1249my $HEX_FORMAT = 'x';
1250my $RATIONAL_FORMAT = 'r';
1251my $STRING_FORMAT = 's';
d11155ec 1252my $ADJUST_FORMAT = 'a';
24303724 1253my $HEX_ADJUST_FORMAT = 'ax';
a14f3cb1 1254my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1255my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1256
1257my %map_table_formats = (
1258 $BINARY_FORMAT => 'binary',
1259 $DECIMAL_FORMAT => 'single decimal digit',
1260 $FLOAT_FORMAT => 'floating point number',
1261 $INTEGER_FORMAT => 'integer',
add63c13 1262 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1263 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1264 $STRING_FORMAT => 'string',
d11155ec 1265 $ADJUST_FORMAT => 'some entries need adjustment',
24303724 1266 $HEX_ADJUST_FORMAT => 'mapped value in hex; some entries need adjustment',
92f9d56c 1267 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1268 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1269);
1270
1271# Unicode didn't put such derived files in a separate directory at first.
1272my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1273my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1274my $AUXILIARY = 'auxiliary';
1275
3854b4b8
KW
1276# Hashes and arrays that will eventually go into Heavy.pl for the use of
1277# utf8_heavy.pl and into UCD.pl for the use of UCD.pm
99870f4d
KW
1278my %loose_to_file_of; # loosely maps table names to their respective
1279 # files
1280my %stricter_to_file_of; # same; but for stricter mapping.
315bfd4e 1281my %loose_property_to_file_of; # Maps a loose property name to its map file
e1dc048d 1282my %strict_property_to_file_of; # Same, but strict
3854b4b8
KW
1283my @inline_definitions = "V0"; # Each element gives a definition of a unique
1284 # inversion list. When a definition is inlined,
1285 # its value in the hash it's in (one of the two
1286 # defined just above) will include an index into
1287 # this array. The 0th element is initialized to
9be22ca5 1288 # the definition for a zero length inversion list
89cf10cc
KW
1289my %file_to_swash_name; # Maps the file name to its corresponding key name
1290 # in the hash %utf8::SwashInfo
99870f4d
KW
1291my %nv_floating_to_rational; # maps numeric values floating point numbers to
1292 # their rational equivalent
c12f2655
KW
1293my %loose_property_name_of; # Loosely maps (non_string) property names to
1294 # standard form
e1dc048d
KW
1295my %strict_property_name_of; # Strictly maps (non_string) property names to
1296 # standard form
86a52d1e 1297my %string_property_loose_to_name; # Same, for string properties.
c15fda25
KW
1298my %loose_defaults; # keys are of form "prop=value", where 'prop' is
1299 # the property name in standard loose form, and
1300 # 'value' is the default value for that property,
1301 # also in standard loose form.
9e4a1e86
KW
1302my %loose_to_standard_value; # loosely maps table names to the canonical
1303 # alias for them
2df7880f
KW
1304my %ambiguous_names; # keys are alias names (in standard form) that
1305 # have more than one possible meaning.
d963b40d
KW
1306my %combination_property; # keys are alias names (in standard form) that
1307 # have both a map table, and a binary one that
1308 # yields true for all non-null maps.
5d1df013
KW
1309my %prop_aliases; # Keys are standard property name; values are each
1310 # one's aliases
1e863613
KW
1311my %prop_value_aliases; # Keys of top level are standard property name;
1312 # values are keys to another hash, Each one is
1313 # one of the property's values, in standard form.
1314 # The values are that prop-val's aliases.
b9e57660 1315my %skipped_files; # List of files that we skip
2df7880f 1316my %ucd_pod; # Holds entries that will go into the UCD section of the pod
99870f4d 1317
d867ccfb
KW
1318# Most properties are immune to caseless matching, otherwise you would get
1319# nonsensical results, as properties are a function of a code point, not
1320# everything that is caselessly equivalent to that code point. For example,
1321# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1322# be true because 's' and 'S' are equivalent caselessly. However,
1323# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1324# extend that concept to those very few properties that are like this. Each
1325# such property will match the full range caselessly. They are hard-coded in
1326# the program; it's not worth trying to make it general as it's extremely
1327# unlikely that they will ever change.
1328my %caseless_equivalent_to;
1329
53adf6a2
KW
1330# This is the range of characters that were in Release 1 of Unicode, and
1331# removed in Release 2 (replaced with the current Hangul syllables starting at
1332# U+AC00). The range was reused starting in Release 3 for other purposes.
1333my $FIRST_REMOVED_HANGUL_SYLLABLE = 0x3400;
1334my $FINAL_REMOVED_HANGUL_SYLLABLE = 0x4DFF;
1335
99870f4d
KW
1336# These constants names and values were taken from the Unicode standard,
1337# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1338# syllables. The '_string' versions are so generated tables can retain the
1339# hex format, which is the more familiar value
1340my $SBase_string = "0xAC00";
1341my $SBase = CORE::hex $SBase_string;
1342my $LBase_string = "0x1100";
1343my $LBase = CORE::hex $LBase_string;
1344my $VBase_string = "0x1161";
1345my $VBase = CORE::hex $VBase_string;
1346my $TBase_string = "0x11A7";
1347my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1348my $SCount = 11172;
1349my $LCount = 19;
1350my $VCount = 21;
1351my $TCount = 28;
1352my $NCount = $VCount * $TCount;
1353
1354# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1355# with the above published constants.
1356my %Jamo;
1357my %Jamo_L; # Leading consonants
1358my %Jamo_V; # Vowels
1359my %Jamo_T; # Trailing consonants
1360
bb1dd3da
KW
1361# For code points whose name contains its ordinal as a '-ABCD' suffix.
1362# The key is the base name of the code point, and the value is an
1363# array giving all the ranges that use this base name. Each range
1364# is actually a hash giving the 'low' and 'high' values of it.
1365my %names_ending_in_code_point;
1366my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1367 # removed from the names
1368# Inverse mapping. The list of ranges that have these kinds of
1369# names. Each element contains the low, high, and base names in an
1370# anonymous hash.
1371my @code_points_ending_in_code_point;
1372
6b5ab373
KW
1373# To hold Unicode's normalization test suite
1374my @normalization_tests;
1375
bb1dd3da
KW
1376# Boolean: does this Unicode version have the hangul syllables, and are we
1377# writing out a table for them?
1378my $has_hangul_syllables = 0;
1379
1380# Does this Unicode version have code points whose names end in their
1381# respective code points, and are we writing out a table for them? 0 for no;
1382# otherwise points to first property that a table is needed for them, so that
1383# if multiple tables are needed, we don't create duplicates
1384my $needing_code_points_ending_in_code_point = 0;
1385
37e2e78e 1386my @backslash_X_tests; # List of tests read in for testing \X
06ae2722 1387my @SB_tests; # List of tests read in for testing \b{sb}
ae3bb8ea 1388my @WB_tests; # List of tests read in for testing \b{wb}
99870f4d
KW
1389my @unhandled_properties; # Will contain a list of properties found in
1390 # the input that we didn't process.
f86864ac 1391my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1392 # listed in the pod
1393my @map_properties; # Properties that get map files written
1394my @named_sequences; # NamedSequences.txt contents.
1395my %potential_files; # Generated list of all .txt files in the directory
1396 # structure so we can warn if something is being
1397 # ignored.
a0204bb1
KW
1398my @missing_early_files; # Generated list of absent files that we need to
1399 # proceed in compiling this early Unicode version
99870f4d
KW
1400my @files_actually_output; # List of files we generated.
1401my @more_Names; # Some code point names are compound; this is used
1402 # to store the extra components of them.
1403my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1404 # the minimum before we consider it equivalent to a
1405 # candidate rational
1406my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1407
1408# These store references to certain commonly used property objects
1254636b 1409my $age;
3c88a801 1410my $ccc;
99870f4d
KW
1411my $gc;
1412my $perl;
1413my $block;
3e20195b
KW
1414my $perl_charname;
1415my $print;
ac7dbdb6 1416my $All;
bc0c431d 1417my $Assigned; # All assigned characters in this Unicode release
e47e66b9
KW
1418my $DI; # Default_Ignorable_Code_Point property
1419my $NChar; # Noncharacter_Code_Point property
359523e2 1420my $script;
99870f4d
KW
1421
1422# Are there conflicting names because of beginning with 'In_', or 'Is_'
1423my $has_In_conflicts = 0;
1424my $has_Is_conflicts = 0;
1425
1426sub internal_file_to_platform ($) {
1427 # Convert our file paths which have '/' separators to those of the
1428 # platform.
1429
1430 my $file = shift;
1431 return undef unless defined $file;
1432
1433 return File::Spec->join(split '/', $file);
d07a55ed 1434}
5beb625e 1435
99870f4d
KW
1436sub file_exists ($) { # platform independent '-e'. This program internally
1437 # uses slash as a path separator.
1438 my $file = shift;
1439 return 0 if ! defined $file;
1440 return -e internal_file_to_platform($file);
1441}
5beb625e 1442
99870f4d 1443sub objaddr($) {
23e33b60
KW
1444 # Returns the address of the blessed input object.
1445 # It doesn't check for blessedness because that would do a string eval
1446 # every call, and the program is structured so that this is never called
1447 # for a non-blessed object.
99870f4d 1448
23e33b60 1449 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1450
1451 # Numifying a ref gives its address.
051df77b 1452 return pack 'J', $_[0];
99870f4d
KW
1453}
1454
558712cf 1455# These are used only if $annotate is true.
c4019d52
KW
1456# The entire range of Unicode characters is examined to populate these
1457# after all the input has been processed. But most can be skipped, as they
1458# have the same descriptive phrases, such as being unassigned
1459my @viacode; # Contains the 1 million character names
4b06b8de 1460my @age; # And their ages ("" if none)
c4019d52
KW
1461my @printable; # boolean: And are those characters printable?
1462my @annotate_char_type; # Contains a type of those characters, specifically
1463 # for the purposes of annotation.
1464my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1465 # name for the purposes of annotation. They map to the
c4019d52
KW
1466 # upper edge of the range, so that the end point can
1467 # be immediately found. This is used to skip ahead to
1468 # the end of a range, and avoid processing each
1469 # individual code point in it.
1470my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1471 # characters, but excluding those which are
1472 # also noncharacter code points
1473
1474# The annotation types are an extension of the regular range types, though
1475# some of the latter are folded into one. Make the new types negative to
1476# avoid conflicting with the regular types
1477my $SURROGATE_TYPE = -1;
1478my $UNASSIGNED_TYPE = -2;
1479my $PRIVATE_USE_TYPE = -3;
1480my $NONCHARACTER_TYPE = -4;
1481my $CONTROL_TYPE = -5;
2d88a86a
KW
1482my $ABOVE_UNICODE_TYPE = -6;
1483my $UNKNOWN_TYPE = -7; # Used only if there is a bug in this program
c4019d52
KW
1484
1485sub populate_char_info ($) {
558712cf 1486 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1487 # input code point's info that are needed for outputting more detailed
1488 # comments. If calling context wants a return, it is the end point of
1489 # any contiguous range of characters that share essentially the same info
1490
1491 my $i = shift;
1492 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1493
1494 $viacode[$i] = $perl_charname->value_of($i) || "";
1254636b
KW
1495 $age[$i] = (defined $age)
1496 ? (($age->value_of($i) =~ / ^ \d \. \d $ /x)
1497 ? $age->value_of($i)
1498 : "")
1499 : "";
c4019d52
KW
1500
1501 # A character is generally printable if Unicode says it is,
1502 # but below we make sure that most Unicode general category 'C' types
1503 # aren't.
1504 $printable[$i] = $print->contains($i);
1505
1254636b
KW
1506 # But the characters in this range were removed in v2.0 and replaced by
1507 # different ones later. Modern fonts will be for the replacement
1508 # characters, so suppress printing them.
1509 if (($v_version lt v2.0
1510 || ($compare_versions && $compare_versions lt v2.0))
1511 && ( $i >= $FIRST_REMOVED_HANGUL_SYLLABLE
1512 && $i <= $FINAL_REMOVED_HANGUL_SYLLABLE))
1513 {
1514 $printable[$i] = 0;
1515 }
1516
c4019d52
KW
1517 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1518
1519 # Only these two regular types are treated specially for annotations
1520 # purposes
1521 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1522 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1523
1524 # Give a generic name to all code points that don't have a real name.
1525 # We output ranges, if applicable, for these. Also calculate the end
1526 # point of the range.
1527 my $end;
1528 if (! $viacode[$i]) {
2d88a86a
KW
1529 if ($i > $MAX_UNICODE_CODEPOINT) {
1530 $viacode[$i] = 'Above-Unicode';
1531 $annotate_char_type[$i] = $ABOVE_UNICODE_TYPE;
1532 $printable[$i] = 0;
1533 $end = $MAX_WORKING_CODEPOINT;
1534 }
1535 elsif ($gc-> table('Private_use')->contains($i)) {
c4019d52
KW
1536 $viacode[$i] = 'Private Use';
1537 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1538 $printable[$i] = 0;
1539 $end = $gc->table('Private_Use')->containing_range($i)->end;
1540 }
e47e66b9 1541 elsif ($NChar->contains($i)) {
c4019d52
KW
1542 $viacode[$i] = 'Noncharacter';
1543 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1544 $printable[$i] = 0;
e47e66b9 1545 $end = $NChar->containing_range($i)->end;
c4019d52
KW
1546 }
1547 elsif ($gc-> table('Control')->contains($i)) {
1254636b
KW
1548 my $name_ref = property_ref('Name_Alias');
1549 $name_ref = property_ref('Unicode_1_Name') if ! defined $name_ref;
1550 $viacode[$i] = (defined $name_ref)
1551 ? $name_ref->value_of($i)
1552 : 'Control';
c4019d52
KW
1553 $annotate_char_type[$i] = $CONTROL_TYPE;
1554 $printable[$i] = 0;
c4019d52
KW
1555 }
1556 elsif ($gc-> table('Unassigned')->contains($i)) {
c4019d52
KW
1557 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1558 $printable[$i] = 0;
1254636b
KW
1559 $viacode[$i] = 'Unassigned';
1560
349edad4 1561 if (defined $block) { # No blocks in earliest releases
1254636b 1562 $viacode[$i] .= ', block=' . $block-> value_of($i);
1d025d66 1563 $end = $gc-> table('Unassigned')->containing_range($i)->end;
c4019d52 1564
bf06c733
KW
1565 # Because we name the unassigned by the blocks they are in, it
1566 # can't go past the end of that block, and it also can't go
1567 # past the unassigned range it is in. The special table makes
1568 # sure that the non-characters, which are unassigned, are
1569 # separated out.
1570 $end = min($block->containing_range($i)->end,
1571 $unassigned_sans_noncharacters->
1572 containing_range($i)->end);
1d025d66 1573 }
1254636b
KW
1574 else {
1575 $end = $i + 1;
1576 while ($unassigned_sans_noncharacters->contains($end)) {
1577 $end++;
1578 }
1579 $end--;
1580 }
1d025d66 1581 }
35e2b416 1582 elsif ($perl->table('_Perl_Surrogate')->contains($i)) {
1d025d66
KW
1583 $viacode[$i] = 'Surrogate';
1584 $annotate_char_type[$i] = $SURROGATE_TYPE;
1585 $printable[$i] = 0;
1586 $end = $gc->table('Surrogate')->containing_range($i)->end;
13ca76ff
KW
1587 }
1588 else {
1589 Carp::my_carp_bug("Can't figure out how to annotate "
1590 . sprintf("U+%04X", $i)
1591 . ". Proceeding anyway.");
c4019d52
KW
1592 $viacode[$i] = 'UNKNOWN';
1593 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1594 $printable[$i] = 0;
1595 }
1596 }
1597
1598 # Here, has a name, but if it's one in which the code point number is
1599 # appended to the name, do that.
1600 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1601 $viacode[$i] .= sprintf("-%04X", $i);
4b06b8de 1602
4b06b8de 1603 my $limit = $perl_charname->containing_range($i)->end;
1254636b
KW
1604 if (defined $age) {
1605 # Do all these as groups of the same age, instead of individually,
1606 # because their names are so meaningless, and there are typically
1607 # large quantities of them.
1608 $end = $i + 1;
1609 while ($end <= $limit && $age->value_of($end) == $age[$i]) {
1610 $end++;
1611 }
1612 $end--;
1613 }
1614 else {
1615 $end = $limit;
4b06b8de 1616 }
c4019d52
KW
1617 }
1618
1619 # And here, has a name, but if it's a hangul syllable one, replace it with
1620 # the correct name from the Unicode algorithm
1621 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1622 use integer;
1623 my $SIndex = $i - $SBase;
1624 my $L = $LBase + $SIndex / $NCount;
1625 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1626 my $T = $TBase + $SIndex % $TCount;
1627 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1628 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1629 $end = $perl_charname->containing_range($i)->end;
1630 }
1631
1632 return if ! defined wantarray;
1633 return $i if ! defined $end; # If not a range, return the input
1634
1635 # Save this whole range so can find the end point quickly
1636 $annotate_ranges->add_map($i, $end, $end);
1637
1638 return $end;
1639}
1640
23e33b60
KW
1641# Commented code below should work on Perl 5.8.
1642## This 'require' doesn't necessarily work in miniperl, and even if it does,
1643## the native perl version of it (which is what would operate under miniperl)
1644## is extremely slow, as it does a string eval every call.
7e017d6d 1645#my $has_fast_scalar_util = $^X !~ /miniperl/
23e33b60
KW
1646# && defined eval "require Scalar::Util";
1647#
1648#sub objaddr($) {
1649# # Returns the address of the blessed input object. Uses the XS version if
1650# # available. It doesn't check for blessedness because that would do a
1651# # string eval every call, and the program is structured so that this is
1652# # never called for a non-blessed object.
1653#
1654# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1655#
1656# # Check at least that is a ref.
1657# my $pkg = ref($_[0]) or return undef;
1658#
1659# # Change to a fake package to defeat any overloaded stringify
1660# bless $_[0], 'main::Fake';
1661#
1662# # Numifying a ref gives its address.
051df77b 1663# my $addr = pack 'J', $_[0];
23e33b60
KW
1664#
1665# # Return to original class
1666# bless $_[0], $pkg;
1667# return $addr;
1668#}
1669
99870f4d
KW
1670sub max ($$) {
1671 my $a = shift;
1672 my $b = shift;
1673 return $a if $a >= $b;
1674 return $b;
1675}
1676
1677sub min ($$) {
1678 my $a = shift;
1679 my $b = shift;
1680 return $a if $a <= $b;
1681 return $b;
1682}
1683
1684sub clarify_number ($) {
1685 # This returns the input number with underscores inserted every 3 digits
1686 # in large (5 digits or more) numbers. Input must be entirely digits, not
1687 # checked.
1688
1689 my $number = shift;
1690 my $pos = length($number) - 3;
1691 return $number if $pos <= 1;
1692 while ($pos > 0) {
1693 substr($number, $pos, 0) = '_';
1694 $pos -= 3;
5beb625e 1695 }
99870f4d 1696 return $number;
99598c8c
JH
1697}
1698
731cb813
KW
1699sub clarify_code_point_count ($) {
1700 # This is like clarify_number(), but the input is assumed to be a count of
1701 # code points, rather than a generic number.
1702
2d88a86a
KW
1703 my $append = "";
1704
1705 my $number = shift;
1706 if ($number > $MAX_UNICODE_CODEPOINTS) {
1707 $number -= ($MAX_WORKING_CODEPOINTS - $MAX_UNICODE_CODEPOINTS);
1708 return "All above-Unicode code points" if $number == 0;
1709 $append = " + all above-Unicode code points";
1710 }
1711 return clarify_number($number) . $append;
731cb813 1712}
12ac2576 1713
99870f4d 1714package Carp;
7ebf06b3 1715
99870f4d
KW
1716# These routines give a uniform treatment of messages in this program. They
1717# are placed in the Carp package to cause the stack trace to not include them,
1718# although an alternative would be to use another package and set @CARP_NOT
1719# for it.
12ac2576 1720
99870f4d 1721our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1722
99f78760
KW
1723# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1724# and overload trying to load Scalar:Util under miniperl. See
1725# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1726undef $overload::VERSION;
1727
99870f4d
KW
1728sub my_carp {
1729 my $message = shift || "";
1730 my $nofold = shift || 0;
7ebf06b3 1731
99870f4d
KW
1732 if ($message) {
1733 $message = main::join_lines($message);
1734 $message =~ s/^$0: *//; # Remove initial program name
1735 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1736 $message = "\n$0: $message;";
12ac2576 1737
99870f4d
KW
1738 # Fold the message with program name, semi-colon end punctuation
1739 # (which looks good with the message that carp appends to it), and a
1740 # hanging indent for continuation lines.
1741 $message = main::simple_fold($message, "", 4) unless $nofold;
1742 $message =~ s/\n$//; # Remove the trailing nl so what carp
1743 # appends is to the same line
1744 }
12ac2576 1745
99870f4d 1746 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1747
99870f4d
KW
1748 carp $message;
1749 return;
1750}
7ebf06b3 1751
99870f4d
KW
1752sub my_carp_bug {
1753 # This is called when it is clear that the problem is caused by a bug in
1754 # this program.
7ebf06b3 1755
99870f4d
KW
1756 my $message = shift;
1757 $message =~ s/^$0: *//;
1758 $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");
1759 carp $message;
1760 return;
1761}
7ebf06b3 1762
99870f4d
KW
1763sub carp_too_few_args {
1764 if (@_ != 2) {
1765 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1766 return;
12ac2576 1767 }
7ebf06b3 1768
99870f4d
KW
1769 my $args_ref = shift;
1770 my $count = shift;
7ebf06b3 1771
99870f4d
KW
1772 my_carp_bug("Need at least $count arguments to "
1773 . (caller 1)[3]
1774 . ". Instead got: '"
1775 . join ', ', @$args_ref
1776 . "'. No action taken.");
1777 return;
12ac2576
JP
1778}
1779
99870f4d
KW
1780sub carp_extra_args {
1781 my $args_ref = shift;
1782 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1783
99870f4d
KW
1784 unless (ref $args_ref) {
1785 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1786 return;
1787 }
1788 my ($package, $file, $line) = caller;
1789 my $subroutine = (caller 1)[3];
cf25bb62 1790
99870f4d
KW
1791 my $list;
1792 if (ref $args_ref eq 'HASH') {
1793 foreach my $key (keys %$args_ref) {
1794 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1795 }
99870f4d 1796 $list = join ', ', each %{$args_ref};
cf25bb62 1797 }
99870f4d
KW
1798 elsif (ref $args_ref eq 'ARRAY') {
1799 foreach my $arg (@$args_ref) {
1800 $arg = $UNDEF unless defined $arg;
1801 }
1802 $list = join ', ', @$args_ref;
1803 }
1804 else {
1805 my_carp_bug("Can't cope with ref "
1806 . ref($args_ref)
1807 . " . argument to 'carp_extra_args'. Not checking arguments.");
1808 return;
1809 }
1810
1811 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1812 return;
d73e5302
JH
1813}
1814
99870f4d
KW
1815package main;
1816
1817{ # Closure
1818
1819 # This program uses the inside-out method for objects, as recommended in
1b6b3fa9
KW
1820 # "Perl Best Practices". (This is the best solution still, since this has
1821 # to run under miniperl.) This closure aids in generating those. There
99870f4d
KW
1822 # are two routines. setup_package() is called once per package to set
1823 # things up, and then set_access() is called for each hash representing a
1824 # field in the object. These routines arrange for the object to be
1825 # properly destroyed when no longer used, and for standard accessor
1826 # functions to be generated. If you need more complex accessors, just
1827 # write your own and leave those accesses out of the call to set_access().
1828 # More details below.
1829
1830 my %constructor_fields; # fields that are to be used in constructors; see
1831 # below
1832
1833 # The values of this hash will be the package names as keys to other
1834 # hashes containing the name of each field in the package as keys, and
1835 # references to their respective hashes as values.
1836 my %package_fields;
1837
1838 sub setup_package {
1839 # Sets up the package, creating standard DESTROY and dump methods
1840 # (unless already defined). The dump method is used in debugging by
1841 # simple_dumper().
1842 # The optional parameters are:
1843 # a) a reference to a hash, that gets populated by later
1844 # set_access() calls with one of the accesses being
1845 # 'constructor'. The caller can then refer to this, but it is
1846 # not otherwise used by these two routines.
1847 # b) a reference to a callback routine to call during destruction
1848 # of the object, before any fields are actually destroyed
1849
1850 my %args = @_;
1851 my $constructor_ref = delete $args{'Constructor_Fields'};
1852 my $destroy_callback = delete $args{'Destroy_Callback'};
1853 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1854
1855 my %fields;
1856 my $package = (caller)[0];
1857
1858 $package_fields{$package} = \%fields;
1859 $constructor_fields{$package} = $constructor_ref;
1860
1861 unless ($package->can('DESTROY')) {
1862 my $destroy_name = "${package}::DESTROY";
1863 no strict "refs";
1864
1865 # Use typeglob to give the anonymous subroutine the name we want
1866 *$destroy_name = sub {
1867 my $self = shift;
ffe43484 1868 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1869
1870 $self->$destroy_callback if $destroy_callback;
1871 foreach my $field (keys %{$package_fields{$package}}) {
1872 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1873 delete $package_fields{$package}{$field}{$addr};
1874 }
1875 return;
1876 }
1877 }
1878
1879 unless ($package->can('dump')) {
1880 my $dump_name = "${package}::dump";
1881 no strict "refs";
1882 *$dump_name = sub {
1883 my $self = shift;
1884 return dump_inside_out($self, $package_fields{$package}, @_);
1885 }
1886 }
1887 return;
1888 }
1889
1890 sub set_access {
1891 # Arrange for the input field to be garbage collected when no longer
1892 # needed. Also, creates standard accessor functions for the field
1893 # based on the optional parameters-- none if none of these parameters:
1894 # 'addable' creates an 'add_NAME()' accessor function.
1895 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1896 # function.
1897 # 'settable' creates a 'set_NAME()' accessor function.
1898 # 'constructor' doesn't create an accessor function, but adds the
1899 # field to the hash that was previously passed to
1900 # setup_package();
1901 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1902 # 'add' etc. all mean 'addable'.
1903 # The read accessor function will work on both array and scalar
1904 # values. If another accessor in the parameter list is 'a', the read
1905 # access assumes an array. You can also force it to be array access
1906 # by specifying 'readable_array' instead of 'readable'
1907 #
1908 # A sort-of 'protected' access can be set-up by preceding the addable,
1909 # readable or settable with some initial portion of 'protected_' (but,
1910 # the underscore is required), like 'p_a', 'pro_set', etc. The
1911 # "protection" is only by convention. All that happens is that the
1912 # accessor functions' names begin with an underscore. So instead of
1913 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1914 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1915 # end of each package, and then storing the __LINE__ ranges and
1916 # checking them on every accessor. But that is way overkill.)
1917
1918 # We create anonymous subroutines as the accessors and then use
1919 # typeglobs to assign them to the proper package and name
1920
1921 my $name = shift; # Name of the field
1922 my $field = shift; # Reference to the inside-out hash containing the
1923 # field
1924
1925 my $package = (caller)[0];
1926
1927 if (! exists $package_fields{$package}) {
1928 croak "$0: Must call 'setup_package' before 'set_access'";
1929 }
d73e5302 1930
99870f4d
KW
1931 # Stash the field so DESTROY can get it.
1932 $package_fields{$package}{$name} = $field;
cf25bb62 1933
99870f4d
KW
1934 # Remaining arguments are the accessors. For each...
1935 foreach my $access (@_) {
1936 my $access = lc $access;
cf25bb62 1937
99870f4d 1938 my $protected = "";
cf25bb62 1939
99870f4d
KW
1940 # Match the input as far as it goes.
1941 if ($access =~ /^(p[^_]*)_/) {
1942 $protected = $1;
1943 if (substr('protected_', 0, length $protected)
1944 eq $protected)
1945 {
1946
1947 # Add 1 for the underscore not included in $protected
1948 $access = substr($access, length($protected) + 1);
1949 $protected = '_';
1950 }
1951 else {
1952 $protected = "";
1953 }
1954 }
1955
1956 if (substr('addable', 0, length $access) eq $access) {
1957 my $subname = "${package}::${protected}add_$name";
1958 no strict "refs";
1959
1960 # add_ accessor. Don't add if already there, which we
1961 # determine using 'eq' for scalars and '==' otherwise.
1962 *$subname = sub {
1963 use strict "refs";
1964 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1965 my $self = shift;
1966 my $value = shift;
ffe43484 1967 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1968 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1969 if (ref $value) {
f998e60c 1970 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1971 }
1972 else {
f998e60c 1973 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1974 }
f998e60c 1975 push @{$field->{$addr}}, $value;
99870f4d
KW
1976 return;
1977 }
1978 }
1979 elsif (substr('constructor', 0, length $access) eq $access) {
1980 if ($protected) {
1981 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1982 }
1983 else {
1984 $constructor_fields{$package}{$name} = $field;
1985 }
1986 }
1987 elsif (substr('readable_array', 0, length $access) eq $access) {
1988
1989 # Here has read access. If one of the other parameters for
1990 # access is array, or this one specifies array (by being more
1991 # than just 'readable_'), then create a subroutine that
1992 # assumes the data is an array. Otherwise just a scalar
1993 my $subname = "${package}::${protected}$name";
1994 if (grep { /^a/i } @_
1995 or length($access) > length('readable_'))
1996 {
1997 no strict "refs";
1998 *$subname = sub {
1999 use strict "refs";
23e33b60 2000 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 2001 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
2002 if (ref $field->{$addr} ne 'ARRAY') {
2003 my $type = ref $field->{$addr};
2004 $type = 'scalar' unless $type;
2005 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
2006 return;
2007 }
2008 return scalar @{$field->{$addr}} unless wantarray;
2009
2010 # Make a copy; had problems with caller modifying the
2011 # original otherwise
2012 my @return = @{$field->{$addr}};
2013 return @return;
2014 }
2015 }
2016 else {
2017
2018 # Here not an array value, a simpler function.
2019 no strict "refs";
2020 *$subname = sub {
2021 use strict "refs";
23e33b60 2022 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 2023 no overloading;
051df77b 2024 return $field->{pack 'J', $_[0]};
99870f4d
KW
2025 }
2026 }
2027 }
2028 elsif (substr('settable', 0, length $access) eq $access) {
2029 my $subname = "${package}::${protected}set_$name";
2030 no strict "refs";
2031 *$subname = sub {
2032 use strict "refs";
23e33b60
KW
2033 if (main::DEBUG) {
2034 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
2035 Carp::carp_extra_args(\@_) if @_ > 2;
2036 }
2037 # $self is $_[0]; $value is $_[1]
f998e60c 2038 no overloading;
051df77b 2039 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
2040 return;
2041 }
2042 }
2043 else {
2044 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
2045 }
cf25bb62 2046 }
99870f4d 2047 return;
cf25bb62 2048 }
99870f4d
KW
2049}
2050
2051package Input_file;
2052
2053# All input files use this object, which stores various attributes about them,
2054# and provides for convenient, uniform handling. The run method wraps the
2055# processing. It handles all the bookkeeping of opening, reading, and closing
2056# the file, returning only significant input lines.
2057#
2058# Each object gets a handler which processes the body of the file, and is
74cd47d0
KW
2059# called by run(). All character property files must use the generic,
2060# default handler, which has code scrubbed to handle things you might not
2061# expect, including automatic EBCDIC handling. For files that don't deal with
2062# mapping code points to a property value, such as test files,
2063# PropertyAliases, PropValueAliases, and named sequences, you can override the
2064# handler to be a custom one. Such a handler should basically be a
2065# while(next_line()) {...} loop.
99870f4d
KW
2066#
2067# You can also set up handlers to
40045c2d 2068# 0) call during object construction time, after everything else is done
537124e4 2069# 1) call before the first line is read, for pre processing
83b68635
KW
2070# 2) call to adjust each line of the input before the main handler gets
2071# them. This can be automatically generated, if appropriately simple
2072# enough, by specifiying a Properties parameter in the constructor.
99870f4d 2073# 3) call upon EOF before the main handler exits its loop
537124e4 2074# 4) call at the end, for post processing
99870f4d
KW
2075#
2076# $_ is used to store the input line, and is to be filtered by the
2077# each_line_handler()s. So, if the format of the line is not in the desired
2078# format for the main handler, these are used to do that adjusting. They can
2079# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
3848e5b7
KW
2080# so the $_ output of one is used as the input to the next. The eof handler
2081# is also stackable, but none of the others are, but could easily be changed
2082# to be so.
99870f4d 2083#
a0204bb1
KW
2084# Some properties are used by the Perl core but aren't defined until later
2085# Unicode releases. The perl interpreter would have problems working when
2086# compiled with an earlier Unicode version that doesn't have them, so we need
2087# to define them somehow for those releases. The 'Early' constructor
2088# parameter can be used to automatically handle this. It is essentially
2089# ignored if the Unicode version being compiled has a data file for this
2090# property. Either code to execute or a file to read can be specified.
2091# Details are at the %early definition.
2092#
99870f4d
KW
2093# Most of the handlers can call insert_lines() or insert_adjusted_lines()
2094# which insert the parameters as lines to be processed before the next input
3848e5b7 2095# file line is read. This allows the EOF handler(s) to flush buffers, for
99870f4d
KW
2096# example. The difference between the two routines is that the lines inserted
2097# by insert_lines() are subjected to the each_line_handler()s. (So if you
9be22ca5
KW
2098# called it from such a handler, you would get infinite recursion without some
2099# mechanism to prevent that.) Lines inserted by insert_adjusted_lines() go
2100# directly to the main handler without any adjustments. If the
2101# post-processing handler calls any of these, there will be no effect. Some
2102# error checking for these conditions could be added, but it hasn't been done.
99870f4d
KW
2103#
2104# carp_bad_line() should be called to warn of bad input lines, which clears $_
2105# to prevent further processing of the line. This routine will output the
2106# message as a warning once, and then keep a count of the lines that have the
2107# same message, and output that count at the end of the file's processing.
2108# This keeps the number of messages down to a manageable amount.
2109#
2110# get_missings() should be called to retrieve any @missing input lines.
2111# Messages will be raised if this isn't done if the options aren't to ignore
2112# missings.
2113
2114sub trace { return main::trace(@_); }
2115
99870f4d
KW
2116{ # Closure
2117 # Keep track of fields that are to be put into the constructor.
2118 my %constructor_fields;
2119
2120 main::setup_package(Constructor_Fields => \%constructor_fields);
2121
2122 my %file; # Input file name, required
2123 main::set_access('file', \%file, qw{ c r });
2124
2125 my %first_released; # Unicode version file was first released in, required
2126 main::set_access('first_released', \%first_released, qw{ c r });
2127
2128 my %handler; # Subroutine to process the input file, defaults to
2129 # 'process_generic_property_file'
2130 main::set_access('handler', \%handler, qw{ c });
2131
2132 my %property;
2133 # name of property this file is for. defaults to none, meaning not
2134 # applicable, or is otherwise determinable, for example, from each line.
696609bf 2135 main::set_access('property', \%property, qw{ c r });
99870f4d
KW
2136
2137 my %optional;
3274d896
KW
2138 # This is either an unsigned number, or a list of property names. In the
2139 # former case, if it is non-zero, it means the file is optional, so if the
2140 # file is absent, no warning about that is output. In the latter case, it
2141 # is a list of properties that the file (exclusively) defines. If the
2142 # file is present, tables for those properties will be produced; if
2143 # absent, none will, even if they are listed elsewhere (namely
2144 # PropertyAliases.txt and PropValueAliases.txt) as being in this release,
2145 # and no warnings will be raised about them not being available. (And no
2146 # warning about the file itself will be raised.)
2147 main::set_access('optional', \%optional, qw{ c readable_array } );
99870f4d
KW
2148
2149 my %non_skip;
2150 # This is used for debugging, to skip processing of all but a few input
2151 # files. Add 'non_skip => 1' to the constructor for those files you want
2152 # processed when you set the $debug_skip global.
2153 main::set_access('non_skip', \%non_skip, 'c');
2154
37e2e78e 2155 my %skip;
0dee21d6
KW
2156 # This is used to skip processing of this input file (semi-) permanently.
2157 # The value should be the reason the file is being skipped. It is used
2158 # for files that we aren't planning to process anytime soon, but want to
2159 # allow to be in the directory and be checked for their names not
2160 # conflicting with any other files on a DOS 8.3 name filesystem, but to
2161 # not otherwise be processed, and to not raise a warning about not being
2162 # handled. In the constructor call, any value that evaluates to a numeric
2163 # 0 or undef means don't skip. Any other value is a string giving the
2164 # reason it is being skippped, and this will appear in generated pod.
2165 # However, an empty string reason will suppress the pod entry.
2166 # Internally, calls that evaluate to numeric 0 are changed into undef to
2167 # distinguish them from an empty string call.
2168 main::set_access('skip', \%skip, 'c', 'r');
37e2e78e 2169
99870f4d
KW
2170 my %each_line_handler;
2171 # list of subroutines to look at and filter each non-comment line in the
2172 # file. defaults to none. The subroutines are called in order, each is
2173 # to adjust $_ for the next one, and the final one adjusts it for
2174 # 'handler'
2175 main::set_access('each_line_handler', \%each_line_handler, 'c');
2176
83b68635
KW
2177 my %properties; # Optional ordered list of the properties that occur in each
2178 # meaningful line of the input file. If present, an appropriate
2179 # each_line_handler() is automatically generated and pushed onto the stack
2180 # of such handlers. This is useful when a file contains multiple
2181 # proerties per line, but no other special considerations are necessary.
2182 # The special value "<ignored>" means to discard the corresponding input
2183 # field.
2184 # Any @missing lines in the file should also match this syntax; no such
2185 # files exist as of 6.3. But if it happens in a future release, the code
2186 # could be expanded to properly parse them.
2187 main::set_access('properties', \%properties, qw{ c r });
2188
99870f4d
KW
2189 my %has_missings_defaults;
2190 # ? Are there lines in the file giving default values for code points
2191 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
2192 # the norm, but IGNORED means it has such lines, but the handler doesn't
2193 # use them. Having these three states allows us to catch changes to the
83b68635
KW
2194 # UCD that this program should track. XXX This could be expanded to
2195 # specify the syntax for such lines, like %properties above.
99870f4d
KW
2196 main::set_access('has_missings_defaults',
2197 \%has_missings_defaults, qw{ c r });
2198
40045c2d
KW
2199 my %construction_time_handler;
2200 # Subroutine to call at the end of the new method. If undef, no such
2201 # handler is called.
2202 main::set_access('construction_time_handler',
2203 \%construction_time_handler, qw{ c });
2204
99870f4d
KW
2205 my %pre_handler;
2206 # Subroutine to call before doing anything else in the file. If undef, no
2207 # such handler is called.
2208 main::set_access('pre_handler', \%pre_handler, qw{ c });
2209
2210 my %eof_handler;
3848e5b7 2211 # Subroutines to call upon getting an EOF on the input file, but before
99870f4d
KW
2212 # that is returned to the main handler. This is to allow buffers to be
2213 # flushed. The handler is expected to call insert_lines() or
2214 # insert_adjusted() with the buffered material
3848e5b7 2215 main::set_access('eof_handler', \%eof_handler, qw{ c });
99870f4d
KW
2216
2217 my %post_handler;
2218 # Subroutine to call after all the lines of the file are read in and
9be22ca5
KW
2219 # processed. If undef, no such handler is called. Note that this cannot
2220 # add lines to be processed; instead use eof_handler
99870f4d
KW
2221 main::set_access('post_handler', \%post_handler, qw{ c });
2222
2223 my %progress_message;
2224 # Message to print to display progress in lieu of the standard one
2225 main::set_access('progress_message', \%progress_message, qw{ c });
2226
2227 my %handle;
2228 # cache open file handle, internal. Is undef if file hasn't been
2229 # processed at all, empty if has;
2230 main::set_access('handle', \%handle);
2231
2232 my %added_lines;
2233 # cache of lines added virtually to the file, internal
2234 main::set_access('added_lines', \%added_lines);
2235
74cd47d0
KW
2236 my %remapped_lines;
2237 # cache of lines added virtually to the file, internal
2238 main::set_access('remapped_lines', \%remapped_lines);
2239
99870f4d
KW
2240 my %errors;
2241 # cache of errors found, internal
2242 main::set_access('errors', \%errors);
2243
2244 my %missings;
2245 # storage of '@missing' defaults lines
2246 main::set_access('missings', \%missings);
2247
a0204bb1
KW
2248 my %early;
2249 # Used for properties that must be defined (for Perl's purposes) on
2250 # versions of Unicode earlier than Unicode itself defines them. The
2251 # parameter is an array (it would be better to be a hash, but not worth
2252 # bothering about due to its rare use).
2253 #
2254 # The first element is either a code reference to call when in a release
2255 # earlier than the Unicode file is available in, or it is an alternate
2256 # file to use instead of the non-existent one. This file must have been
2257 # plunked down in the same directory as mktables. Should you be compiling
2258 # on a release that needs such a file, mktables will abort the
2259 # compilation, and tell you where to get the necessary file(s), and what
2260 # name(s) to use to store them as.
2261 # In the case of specifying an alternate file, the array must contain two
2262 # further elements:
2263 #
2264 # [1] is the name of the property that will be generated by this file.
2265 # The class automatically takes the input file and excludes any code
2266 # points in it that were not assigned in the Unicode version being
2267 # compiled. It then uses this result to define the property in the given
2268 # version. Since the property doesn't actually exist in the Unicode
2269 # version being compiled, this should be a name accessible only by core
2270 # perl. If it is the same name as the regular property, the constructor
2271 # will mark the output table as a $PLACEHOLDER so that it doesn't actually
2272 # get output, and so will be unusable by non-core code. Otherwise it gets
2273 # marked as $INTERNAL_ONLY.
2274 #
2275 # [2] is a property value to assign (only when compiling Unicode 1.1.5) to
2276 # the Hangul syllables in that release (which were ripped out in version
2277 # 2) for the given property . (Hence it is ignored except when compiling
2278 # version 1. You only get one value that applies to all of them, which
2279 # may not be the actual reality, but probably nobody cares anyway for
2280 # these obsolete characters.)
2281 #
2282 # Not all files can be handled in the above way, and so the code ref
2283 # alternative is available. It can do whatever it needs to. The other
2284 # array elements are optional in this case, and the code is free to use or
2285 # ignore them if they are present.
2286 #
2287 # Internally, the constructor unshifts a 0 or 1 onto this array to
2288 # indicate if an early alternative is actually being used or not. This
2289 # makes for easier testing later on.
2290 main::set_access('early', \%early, 'c');
2291
0dee21d6
KW
2292 my %required_even_in_debug_skip;
2293 # debug_skip is used to speed up compilation during debugging by skipping
2294 # processing files that are not needed for the task at hand. However,
2295 # some files pretty much can never be skipped, and this is used to specify
2296 # that this is one of them. In order to skip this file, the call to the
2297 # constructor must be edited to comment out this parameter.
2298 main::set_access('required_even_in_debug_skip',
2299 \%required_even_in_debug_skip, 'c');
2300
d8062573
KW
2301 my %withdrawn;
2302 # Some files get removed from the Unicode DB. This is a version object
2303 # giving the first release without this file.
2304 main::set_access('withdrawn', \%withdrawn, 'c');
2305
0dee21d6
KW
2306 my %in_this_release;
2307 # Calculated value from %first_released and %withdrawn. Are we compiling
2308 # a Unicode release which includes this file?
2309 main::set_access('in_this_release', \%in_this_release);
2310
74cd47d0
KW
2311 sub _next_line;
2312 sub _next_line_with_remapped_range;
2313
99870f4d
KW
2314 sub new {
2315 my $class = shift;
2316
2317 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2318 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2319
2320 # Set defaults
2321 $handler{$addr} = \&main::process_generic_property_file;
2322 $non_skip{$addr} = 0;
0dee21d6 2323 $skip{$addr} = undef;
99870f4d
KW
2324 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2325 $handle{$addr} = undef;
2326 $added_lines{$addr} = [ ];
74cd47d0 2327 $remapped_lines{$addr} = [ ];
99870f4d 2328 $each_line_handler{$addr} = [ ];
3848e5b7 2329 $eof_handler{$addr} = [ ];
99870f4d
KW
2330 $errors{$addr} = { };
2331 $missings{$addr} = [ ];
a0204bb1 2332 $early{$addr} = [ ];
3274d896 2333 $optional{$addr} = [ ];
99870f4d
KW
2334
2335 # Two positional parameters.
99f78760 2336 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2337 $file{$addr} = main::internal_file_to_platform(shift);
2338 $first_released{$addr} = shift;
2339
2340 # The rest of the arguments are key => value pairs
2341 # %constructor_fields has been set up earlier to list all possible
2342 # ones. Either set or push, depending on how the default has been set
2343 # up just above.
2344 my %args = @_;
2345 foreach my $key (keys %args) {
2346 my $argument = $args{$key};
2347
2348 # Note that the fields are the lower case of the constructor keys
2349 my $hash = $constructor_fields{lc $key};
2350 if (! defined $hash) {
2351 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2352 next;
2353 }
2354 if (ref $hash->{$addr} eq 'ARRAY') {
2355 if (ref $argument eq 'ARRAY') {
2356 foreach my $argument (@{$argument}) {
2357 next if ! defined $argument;
2358 push @{$hash->{$addr}}, $argument;
2359 }
2360 }
2361 else {
2362 push @{$hash->{$addr}}, $argument if defined $argument;
2363 }
2364 }
2365 else {
2366 $hash->{$addr} = $argument;
2367 }
2368 delete $args{$key};
2369 };
2370
0dee21d6
KW
2371 $non_skip{$addr} = 1 if $required_even_in_debug_skip{$addr};
2372
2373 # Convert 0 (meaning don't skip) to undef
2374 undef $skip{$addr} unless $skip{$addr};
2375
3274d896
KW
2376 # Handle the case where this file is optional
2377 my $pod_message_for_non_existent_optional = "";
2378 if ($optional{$addr}->@*) {
2379
2380 # First element is the pod message
2381 $pod_message_for_non_existent_optional
2382 = shift $optional{$addr}->@*;
2383 # Convert a 0 'Optional' argument to an empty list to make later
2384 # code more concise.
2385 if ( $optional{$addr}->@*
2386 && $optional{$addr}->@* == 1
2387 && $optional{$addr}[0] ne ""
2388 && $optional{$addr}[0] !~ /\D/
2389 && $optional{$addr}[0] == 0)
2390 {
2391 $optional{$addr} = [ ];
2392 }
2393 else { # But if the only element doesn't evaluate to 0, make sure
2394 # that this file is indeed considered optional below.
2395 unshift $optional{$addr}->@*, 1;
2396 }
2397 }
2398
0dee21d6 2399 my $progress;
a0204bb1
KW
2400 my $function_instead_of_file = 0;
2401
2402 # If we are compiling a Unicode release earlier than the file became
2403 # available, the constructor may have supplied a substitute
2404 if ($first_released{$addr} gt $v_version && $early{$addr}->@*) {
2405
2406 # Yes, we have a substitute, that we will use; mark it so
2407 unshift $early{$addr}->@*, 1;
2408
2409 # See the definition of %early for what the array elements mean.
2410 # If we have a property this defines, create a table and default
2411 # map for it now (at essentially compile time), so that it will be
2412 # available for the whole of run time. (We will want to add this
2413 # name as an alias when we are using the official property name;
2414 # but this must be deferred until run(), because at construction
2415 # time the official names have yet to be defined.)
2416 if ($early{$addr}[2]) {
2417 my $fate = ($property{$addr}
2418 && $property{$addr} eq $early{$addr}[2])
2419 ? $PLACEHOLDER
2420 : $INTERNAL_ONLY;
2421 my $prop_object = Property->new($early{$addr}[2],
2422 Fate => $fate,
2423 Perl_Extension => 1,
2424 );
2425
2426 # Use the default mapping for the regular property for this
2427 # substitute one.
2428 if ( defined $property{$addr}
2429 && defined $default_mapping{$property{$addr}})
2430 {
2431 $prop_object
2432 ->set_default_map($default_mapping{$property{$addr}});
2433 }
2434 }
2435
2436 if (ref $early{$addr}[1] eq 'CODE') {
2437 $function_instead_of_file = 1;
2438
2439 # If the first element of the array is a code ref, the others
2440 # are optional.
2441 $handler{$addr} = $early{$addr}[1];
2442 $property{$addr} = $early{$addr}[2]
2443 if defined $early{$addr}[2];
2444 $progress = "substitute $file{$addr}";
2445
2446 undef $file{$addr};
2447 }
2448 else { # Specifying a substitute file
2449
2450 if (! main::file_exists($early{$addr}[1])) {
2451
2452 # If we don't see the substitute file, generate an error
2453 # message giving the needed things, and add it to the list
2454 # of such to output before actual processing happens
2455 # (hence the user finds out all of them in one run).
ce8d64d9
KW
2456 # Instead of creating a general method for NameAliases,
2457 # hard-code it here, as there is unlikely to ever be a
2458 # second one which needs special handling.
2459 my $string_version = ($file{$addr} eq "NameAliases.txt")
2460 ? 'at least 6.1 (the later, the better)'
2461 : sprintf "%vd", $first_released{$addr};
a0204bb1
KW
2462 push @missing_early_files, <<END;
2463'$file{$addr}' version $string_version should be copied to '$early{$addr}[1]'.
2464END
2465 ;
2466 return;
2467 }
2468 $progress = $early{$addr}[1];
2469 $progress .= ", substituting for $file{$addr}" if $file{$addr};
2470 $file{$addr} = $early{$addr}[1];
2471 $property{$addr} = $early{$addr}[2];
0dee21d6 2472
a0204bb1
KW
2473 # Ignore code points not in the version being compiled
2474 push $each_line_handler{$addr}->@*, \&_exclude_unassigned;
2475
2476 if ( $v_version lt v2.0 # Hanguls in this release ...
2477 && defined $early{$addr}[3]) # ... need special treatment
2478 {
2479 push $eof_handler{$addr}->@*, \&_fixup_obsolete_hanguls;
2480 }
2481 }
2482
2483 # And this substitute is valid for all releases.
2484 $first_released{$addr} = v0;
2485 }
2486 else { # Normal behavior
0dee21d6 2487 $progress = $file{$addr};
a0204bb1 2488 unshift $early{$addr}->@*, 0; # No substitute
99870f4d
KW
2489 }
2490
0dee21d6
KW
2491 my $file = $file{$addr};
2492 $progress_message{$addr} = "Processing $progress"
2493 unless $progress_message{$addr};
2494
d8062573
KW
2495 # A file should be there if it is within the window of versions for
2496 # which Unicode supplies it
2497 if ($withdrawn{$addr} && $withdrawn{$addr} le $v_version) {
2498 $in_this_release{$addr} = 0;
2499 $skip{$addr} = "";
2500 }
2501 else {
0dee21d6
KW
2502 $in_this_release{$addr} = $first_released{$addr} le $v_version;
2503
a0204bb1
KW
2504 # Check that the file for this object (possibly using a substitute
2505 # for early releases) exists or we have a function alternative
2506 if ( ! $function_instead_of_file
2507 && ! main::file_exists($file))
0dee21d6
KW
2508 {
2509 # Here there is nothing available for this release. This is
2510 # fine if we aren't expecting anything in this release.
2511 if (! $in_this_release{$addr}) {
2512 $skip{$addr} = ""; # Don't remark since we expected
2513 # nothing and got nothing
2514 }
3274d896 2515 elsif ($optional{$addr}->@*) {
0dee21d6 2516
3274d896
KW
2517 # Here the file is optional in this release; Use the
2518 # passed in text to document this case in the pod.
2519 $skip{$addr} = $pod_message_for_non_existent_optional;
0dee21d6
KW
2520 }
2521 elsif ( $in_this_release{$addr}
2522 && ! defined $skip{$addr}
2523 && defined $file)
2524 { # Doesn't exist but should.
2525 $skip{$addr} = "'$file' not found. Possibly Big problems";
2526 Carp::my_carp($skip{$addr});
2527 }
2528 }
2529 elsif ($debug_skip && ! defined $skip{$addr} && ! $non_skip{$addr})
2530 {
2531
2532 # The file exists; if not skipped for another reason, and we are
2533 # skipping most everything during debugging builds, use that as
2534 # the skip reason.
2535 $skip{$addr} = '$debug_skip is on'
2536 }
d8062573 2537 }
0dee21d6
KW
2538
2539 if ( ! $debug_skip
2540 && $non_skip{$addr}
2541 && ! $required_even_in_debug_skip{$addr}
2542 && $verbosity)
2543 {
2544 print "Warning: " . __PACKAGE__ . " constructor for $file has useless 'non_skip' in it\n";
a3a8c5f0 2545 }
99870f4d 2546
0dee21d6 2547 # Here, we have figured out if we will be skipping this file or not.
3274d896
KW
2548 # If so, we add any single property it defines to any passed in
2549 # optional property list. These will be dealt with at run time.
0dee21d6 2550 if (defined $skip{$addr}) {
3274d896
KW
2551 if ($property{$addr}) {
2552 push $optional{$addr}->@*, $property{$addr};
2553 }
2554 } # Otherwise, are going to process the file.
0dee21d6
KW
2555 elsif ($property{$addr}) {
2556
2557 # If the file has a property defined in the constructor for it, it
2558 # means that the property is not listed in the file's entries. So
2559 # add a handler (to the list of line handlers) to insert the
2560 # property name into the lines, to provide a uniform interface to
2561 # the final processing subroutine.
2562 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
09ca89ce 2563 }
83b68635
KW
2564 elsif ($properties{$addr}) {
2565
9be22ca5
KW
2566 # Similarly, there may be more than one property represented on
2567 # each line, with no clue but the constructor input what those
2568 # might be. Add a handler for each line in the input so that it
2569 # creates a separate input line for each property in those input
2570 # lines, thus making them suitable to handle generically.
83b68635
KW
2571
2572 push @{$each_line_handler{$addr}},
2573 sub {
2574 my $file = shift;
2575 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2576
2577 my @fields = split /\s*;\s*/, $_, -1;
2578
2579 if (@fields - 1 > @{$properties{$addr}}) {
2580 $file->carp_bad_line('Extra fields');
2581 $_ = "";
2582 return;
2583 }
2584 my $range = shift @fields; # 0th element is always the
2585 # range
2586
2587 # The next fields in the input line correspond
2588 # respectively to the stored properties.
2589 for my $i (0 .. @{$properties{$addr}} - 1) {
2590 my $property_name = $properties{$addr}[$i];
2591 next if $property_name eq '<ignored>';
2592 $file->insert_adjusted_lines(
2593 "$range; $property_name; $fields[$i]");
2594 }
2595 $_ = "";
2596
2597 return;
2598 };
2599 }
37e2e78e 2600
9be22ca5 2601 { # On non-ascii platforms, we use a special pre-handler
74cd47d0
KW
2602 no strict;
2603 no warnings 'once';
2604 *next_line = (main::NON_ASCII_PLATFORM)
2605 ? *_next_line_with_remapped_range
2606 : *_next_line;
2607 }
2608
40045c2d
KW
2609 &{$construction_time_handler{$addr}}($self)
2610 if $construction_time_handler{$addr};
2611
99870f4d 2612 return $self;
d73e5302
JH
2613 }
2614
cf25bb62 2615
99870f4d
KW
2616 use overload
2617 fallback => 0,
2618 qw("") => "_operator_stringify",
2619 "." => \&main::_operator_dot,
1285127e 2620 ".=" => \&main::_operator_dot_equal,
99870f4d 2621 ;
cf25bb62 2622
99870f4d
KW
2623 sub _operator_stringify {
2624 my $self = shift;
cf25bb62 2625
99870f4d 2626 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2627 }
d73e5302 2628
99870f4d
KW
2629 sub run {
2630 # Process the input object $self. This opens and closes the file and
2631 # calls all the handlers for it. Currently, this can only be called
3848e5b7 2632 # once per file, as it destroy's the EOF handlers
d73e5302 2633
0dee21d6 2634 # flag to make sure extracted files are processed early
f8b69004 2635 state $seen_non_extracted = 0;
0dee21d6 2636
99870f4d
KW
2637 my $self = shift;
2638 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2639
ffe43484 2640 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2641
99870f4d 2642 my $file = $file{$addr};
d73e5302 2643
0dee21d6
KW
2644 if (! $file) {
2645 $handle{$addr} = 'pretend_is_open';
99870f4d
KW
2646 }
2647 else {
f8b69004 2648 if ($seen_non_extracted) {
9be22ca5
KW
2649 if ($file =~ /$EXTRACTED/i) # Some platforms may change the
2650 # case of the file's name
2651 {
1675ea0d 2652 Carp::my_carp_bug(main::join_lines(<<END
99f78760 2653$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2654anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2655have subtle problems
2656END
2657 ));
2658 }
2659 }
2660 elsif ($EXTRACTED_DIR
0dee21d6
KW
2661
2662 # We only do this check for generic property files
2663 && $handler{$addr} == \&main::process_generic_property_file
2664
f8b69004 2665 && $file !~ /$EXTRACTED/i)
99870f4d
KW
2666 {
2667 # We don't set this (by the 'if' above) if we have no
2668 # extracted directory, so if running on an early version,
2669 # this test won't work. Not worth worrying about.
f8b69004 2670 $seen_non_extracted = 1;
99870f4d
KW
2671 }
2672
9be22ca5 2673 # Mark the file as having being processed, and warn if it
99870f4d
KW
2674 # isn't a file we are expecting. As we process the files,
2675 # they are deleted from the hash, so any that remain at the
2676 # end of the program are files that we didn't process.
517956bf 2677 my $fkey = File::Spec->rel2abs($file);
0dee21d6 2678 my $exists = delete $potential_files{lc($fkey)};
faf3cf6b 2679
0dee21d6
KW
2680 Carp::my_carp("Was not expecting '$file'.")
2681 if $exists && ! $in_this_release{$addr};
2682
a0204bb1
KW
2683 # If there is special handling for compiling Unicode releases
2684 # earlier than the first one in which Unicode defines this
2685 # property ...
2686 if ($early{$addr}->@* > 1) {
2687
2688 # Mark as processed any substitute file that would be used in
2689 # such a release
2690 $fkey = File::Spec->rel2abs($early{$addr}[1]);
2691 delete $potential_files{lc($fkey)};
2692
2693 # As commented in the constructor code, when using the
2694 # official property, we still have to allow the publicly
2695 # inaccessible early name so that the core code which uses it
2696 # will work regardless.
2697 if (! $early{$addr}[0] && $early{$addr}->@* > 2) {
2698 my $early_property_name = $early{$addr}[2];
2699 if ($property{$addr} ne $early_property_name) {
2700 main::property_ref($property{$addr})
2701 ->add_alias($early_property_name);
2702 }
2703 }
2704 }
2705
0dee21d6
KW
2706 # We may be skipping this file ...
2707 if (defined $skip{$addr}) {
2708
2709 # If the file isn't supposed to be in this release, there is
2710 # nothing to do
2711 if ($in_this_release{$addr}) {
2712
2713 # But otherwise, we may print a message
2714 if ($debug_skip) {
2715 print STDERR "Skipping input file '$file'",
2716 " because '$skip{$addr}'\n";
2717 }
2718
2719 # And add it to the list of skipped files, which is later
2720 # used to make the pod
2721 $skipped_files{$file} = $skip{$addr};
3274d896
KW
2722
2723 # The 'optional' list contains properties that are also to
2724 # be skipped along with the file. (There may also be
2725 # digits which are just placeholders to make sure it isn't
2726 # an empty list
2727 foreach my $property ($optional{$addr}->@*) {
2728 next unless $property =~ /\D/;
2729 my $prop_object = main::property_ref($property);
2730 next unless defined $prop_object;
2731 $prop_object->set_fate($SUPPRESSED, $skip{$addr});
2732 }
0dee21d6 2733 }
99870f4d 2734
37e2e78e
KW
2735 return;
2736 }
2737
9be22ca5
KW
2738 # Here, we are going to process the file. Open it, converting the
2739 # slashes used in this program into the proper form for the OS
99870f4d
KW
2740 my $file_handle;
2741 if (not open $file_handle, "<", $file) {
2742 Carp::my_carp("Can't open $file. Skipping: $!");
d3664eb9 2743 return;
99870f4d
KW
2744 }
2745 $handle{$addr} = $file_handle; # Cache the open file handle
9e65c3f4 2746
0dee21d6
KW
2747 # If possible, make sure that the file is the correct version.
2748 # (This data isn't available on early Unicode releases or in
2749 # UnicodeData.txt.) We don't do this check if we are using a
2750 # substitute file instead of the official one (though the code
2751 # could be extended to do so).
2752 if ($in_this_release{$addr}
a0204bb1 2753 && ! $early{$addr}[0]
0dee21d6
KW
2754 && lc($file) ne 'unicodedata.txt')
2755 {
96f226dc 2756 if ($file !~ /^Unihan/i) {
0dee21d6
KW
2757
2758 # The non-Unihan files started getting version numbers in
2759 # 3.2, but some files in 4.0 are unchanged from 3.2, and
2760 # marked as 3.2. 4.0.1 is the first version where there
2761 # are no files marked as being from less than 4.0, though
2762 # some are marked as 4.0. In versions after that, the
2763 # numbers are correct.
2764 if ($v_version ge v4.0.1) {
2765 $_ = <$file_handle>; # The version number is in the
2766 # very first line
2767 if ($_ !~ / - $string_version \. /x) {
2768 chomp;
2769 $_ =~ s/^#\s*//;
2770
2771 # 4.0.1 had some valid files that weren't updated.
2772 if (! ($v_version eq v4.0.1 && $_ =~ /4\.0\.0/)) {
2773 die Carp::my_carp("File '$file' is version "
2774 . "'$_'. It should be "
2775 . "version $string_version");
2776 }
2777 }
cafe9cf0 2778 }
96f226dc 2779 }
0dee21d6
KW
2780 elsif ($v_version ge v6.0.0) { # Unihan
2781
2782 # Unihan files didn't get accurate version numbers until
2783 # 6.0. The version is somewhere in the first comment
2784 # block
96f226dc
KW
2785 while (<$file_handle>) {
2786 if ($_ !~ /^#/) {
0dee21d6
KW
2787 Carp::my_carp_bug("Could not find the expected "
2788 . "version info in file '$file'");
96f226dc
KW
2789 last;
2790 }
2791 chomp;
2792 $_ =~ s/^#\s*//;
2793 next if $_ !~ / version: /x;
2794 last if $_ =~ /$string_version/;
0dee21d6
KW
2795 die Carp::my_carp("File '$file' is version "
2796 . "'$_'. It should be "
2797 . "version $string_version");
96f226dc
KW
2798 }
2799 }
9e65c3f4 2800 }
99870f4d
KW
2801 }
2802
0dee21d6 2803 print "$progress_message{$addr}\n" if $verbosity >= $PROGRESS;
99870f4d
KW
2804
2805 # Call any special handler for before the file.
2806 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2807
2808 # Then the main handler
2809 &{$handler{$addr}}($self);
2810
2811 # Then any special post-file handler.
2812 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2813
2814 # If any errors have been accumulated, output the counts (as the first
2815 # error message in each class was output when it was encountered).
2816 if ($errors{$addr}) {
2817 my $total = 0;
2818 my $types = 0;
2819 foreach my $error (keys %{$errors{$addr}}) {
2820 $total += $errors{$addr}->{$error};
2821 delete $errors{$addr}->{$error};
2822 $types++;
2823 }
2824 if ($total > 1) {
2825 my $message
2826 = "A total of $total lines had errors in $file. ";
2827
2828 $message .= ($types == 1)
2829 ? '(Only the first one was displayed.)'
2830 : '(Only the first of each type was displayed.)';
2831 Carp::my_carp($message);
2832 }
2833 }
2834
2835 if (@{$missings{$addr}}) {
2836 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2837 }
2838
2839 # If a real file handle, close it.
2840 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2841 ref $handle{$addr};
2842 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2843 # the file, as opposed to undef
2844 return;
2845 }
2846
74cd47d0 2847 sub _next_line {
99870f4d
KW
2848 # Sets $_ to be the next logical input line, if any. Returns non-zero
2849 # if such a line exists. 'logical' means that any lines that have
2850 # been added via insert_lines() will be returned in $_ before the file
2851 # is read again.
2852
2853 my $self = shift;
2854 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2855
ffe43484 2856 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2857
2858 # Here the file is open (or if the handle is not a ref, is an open
2859 # 'virtual' file). Get the next line; any inserted lines get priority
2860 # over the file itself.
2861 my $adjusted;
2862
2863 LINE:
2864 while (1) { # Loop until find non-comment, non-empty line
2865 #local $to_trace = 1 if main::DEBUG;
2866 my $inserted_ref = shift @{$added_lines{$addr}};
2867 if (defined $inserted_ref) {
2868 ($adjusted, $_) = @{$inserted_ref};
2869 trace $adjusted, $_ if main::DEBUG && $to_trace;
2870 return 1 if $adjusted;
2871 }
2872 else {
2873 last if ! ref $handle{$addr}; # Don't read unless is real file
2874 last if ! defined ($_ = readline $handle{$addr});
2875 }
2876 chomp;
2877 trace $_ if main::DEBUG && $to_trace;
2878
2879 # See if this line is the comment line that defines what property
2880 # value that code points that are not listed in the file should
2881 # have. The format or existence of these lines is not guaranteed
2882 # by Unicode since they are comments, but the documentation says
2883 # that this was added for machine-readability, so probably won't
2884 # change. This works starting in Unicode Version 5.0. They look
2885 # like:
2886 #
2887 # @missing: 0000..10FFFF; Not_Reordered
2888 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2889 # @missing: 0000..10FFFF; ; NaN
2890 #
2891 # Save the line for a later get_missings() call.
2892 if (/$missing_defaults_prefix/) {
2893 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2894 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2895 }
2896 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2897 my @defaults = split /\s* ; \s*/x, $_;
2898
2899 # The first field is the @missing, which ends in a
2900 # semi-colon, so can safely shift.
2901 shift @defaults;
2902
2903 # Some of these lines may have empty field placeholders
2904 # which get in the way. An example is:
2905 # @missing: 0000..10FFFF; ; NaN
2906 # Remove them. Process starting from the top so the
2907 # splice doesn't affect things still to be looked at.
2908 for (my $i = @defaults - 1; $i >= 0; $i--) {
2909 next if $defaults[$i] ne "";
2910 splice @defaults, $i, 1;
2911 }
2912
2913 # What's left should be just the property (maybe) and the
2914 # default. Having only one element means it doesn't have
2915 # the property.
2916 my $default;
2917 my $property;
2918 if (@defaults >= 1) {
2919 if (@defaults == 1) {
2920 $default = $defaults[0];
2921 }
2922 else {
2923 $property = $defaults[0];
2924 $default = $defaults[1];
2925 }
2926 }
2927
2928 if (@defaults < 1
2929 || @defaults > 2
2930 || ($default =~ /^</
2931 && $default !~ /^<code *point>$/i
09f8d0ac
KW
2932 && $default !~ /^<none>$/i
2933 && $default !~ /^<script>$/i))
99870f4d
KW
2934 {
2935 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2936 }
2937 else {
2938
2939 # If the property is missing from the line, it should
2940 # be the one for the whole file
2941 $property = $property{$addr} if ! defined $property;
2942
2943 # Change <none> to the null string, which is what it
2944 # really means. If the default is the code point
2945 # itself, set it to <code point>, which is what
2946 # Unicode uses (but sometimes they've forgotten the
2947 # space)
2948 if ($default =~ /^<none>$/i) {
2949 $default = "";
2950 }
2951 elsif ($default =~ /^<code *point>$/i) {
2952 $default = $CODE_POINT;
2953 }
09f8d0ac
KW
2954 elsif ($default =~ /^<script>$/i) {
2955
2956 # Special case this one. Currently is from
2957 # ScriptExtensions.txt, and means for all unlisted
2958 # code points, use their Script property values.
2959 # For the code points not listed in that file, the
2960 # default value is 'Unknown'.
2961 $default = "Unknown";
2962 }
99870f4d
KW
2963
2964 # Store them as a sub-arrays with both components.
2965 push @{$missings{$addr}}, [ $default, $property ];
2966 }
2967 }
2968
2969 # There is nothing for the caller to process on this comment
2970 # line.
2971 next;
2972 }
2973
2974 # Remove comments and trailing space, and skip this line if the
2975 # result is empty
2976 s/#.*//;
2977 s/\s+$//;
2978 next if /^$/;
2979
2980 # Call any handlers for this line, and skip further processing of
2981 # the line if the handler sets the line to null.
2982 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2983 &{$sub_ref}($self);
2984 next LINE if /^$/;
2985 }
2986
2987 # Here the line is ok. return success.
2988 return 1;
2989 } # End of looping through lines.
2990
3848e5b7 2991 # If there are EOF handlers, call each (only once) and if it generates
99870f4d 2992 # more lines to process go back in the loop to handle them.
3848e5b7
KW
2993 while ($eof_handler{$addr}->@*) {
2994 &{$eof_handler{$addr}[0]}($self);
2995 shift $eof_handler{$addr}->@*; # Currently only get one shot at it.
99870f4d
KW
2996 goto LINE if $added_lines{$addr};
2997 }
2998
2999 # Return failure -- no more lines.
3000 return 0;
3001
3002 }
3003
74cd47d0
KW
3004 sub _next_line_with_remapped_range {
3005 my $self = shift;
3006 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3007
3008 # like _next_line(), but for use on non-ASCII platforms. It sets $_
3009 # to be the next logical input line, if any. Returns non-zero if such
3010 # a line exists. 'logical' means that any lines that have been added
3011 # via insert_lines() will be returned in $_ before the file is read
3012 # again.
3013 #
3014 # The difference from _next_line() is that this remaps the Unicode
3015 # code points in the input to those of the native platform. Each
3016 # input line contains a single code point, or a single contiguous
3017 # range of them This routine splits each range into its individual
3018 # code points and caches them. It returns the cached values,
3019 # translated into their native equivalents, one at a time, for each
3020 # call, before reading the next line. Since native values can only be
3021 # a single byte wide, no translation is needed for code points above
3022 # 0xFF, and ranges that are entirely above that number are not split.
3023 # If an input line contains the range 254-1000, it would be split into
3024 # three elements: 254, 255, and 256-1000. (The downstream table
3025 # insertion code will sort and coalesce the individual code points
3026 # into appropriate ranges.)
3027
3028 my $addr = do { no overloading; pack 'J', $self; };
3029
3030 while (1) {
3031
3032 # Look in cache before reading the next line. Return any cached
3033 # value, translated
3034 my $inserted = shift @{$remapped_lines{$addr}};
3035 if (defined $inserted) {
3036 trace $inserted if main::DEBUG && $to_trace;
3037 $_ = $inserted =~ s/^ ( \d+ ) /sprintf("%04X", utf8::unicode_to_native($1))/xer;
3038 trace $_ if main::DEBUG && $to_trace;
3039 return 1;
3040 }
3041
3042 # Get the next line.
3043 return 0 unless _next_line($self);
3044
3045 # If there is a special handler for it, return the line,
3046 # untranslated. This should happen only for files that are
3047 # special, not being code-point related, such as property names.
3048 return 1 if $handler{$addr}
3049 != \&main::process_generic_property_file;
3050
3051 my ($range, $property_name, $map, @remainder)
3052 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3053
3054 if (@remainder
3055 || ! defined $property_name
3056 || $range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3057 {
3058 Carp::my_carp_bug("Unrecognized input line '$_'. Ignored");
3059 }
3060
3061 my $low = hex $1;
3062 my $high = (defined $2) ? hex $2 : $low;
3063
3064 # If the input maps the range to another code point, remap the
3065 # target if it is between 0 and 255.
3066 my $tail;
3067 if (defined $map) {
3068 $map =~ s/\b 00 ( [0-9A-F]{2} ) \b/sprintf("%04X", utf8::unicode_to_native(hex $1))/gxe;
3069 $tail = "$property_name; $map";
3070 $_ = "$range; $tail";
3071 }
3072 else {
3073 $tail = $property_name;
3074 }
3075
3076 # If entire range is above 255, just return it, unchanged (except
3077 # any mapped-to code point, already changed above)
3078 return 1 if $low > 255;
3079
3080 # Cache an entry for every code point < 255. For those in the
3081 # range above 255, return a dummy entry for just that portion of
3082 # the range. Note that this will be out-of-order, but that is not
3083 # a problem.
3084 foreach my $code_point ($low .. $high) {
3085 if ($code_point > 255) {
3086 $_ = sprintf "%04X..%04X; $tail", $code_point, $high;
3087 return 1;
3088 }
3089 push @{$remapped_lines{$addr}}, "$code_point; $tail";
3090 }
3091 } # End of looping through lines.
3092
3093 # NOTREACHED
3094 }
3095
99870f4d
KW
3096# Not currently used, not fully tested.
3097# sub peek {
3098# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
3099# # record. Not callable from an each_line_handler(), nor does it call
3100# # an each_line_handler() on the line.
3101#
3102# my $self = shift;
ffe43484 3103# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3104#
3105# foreach my $inserted_ref (@{$added_lines{$addr}}) {
3106# my ($adjusted, $line) = @{$inserted_ref};
3107# next if $adjusted;
3108#
3109# # Remove comments and trailing space, and return a non-empty
3110# # resulting line
3111# $line =~ s/#.*//;
3112# $line =~ s/\s+$//;
3113# return $line if $line ne "";
3114# }
3115#
3116# return if ! ref $handle{$addr}; # Don't read unless is real file
3117# while (1) { # Loop until find non-comment, non-empty line
3118# local $to_trace = 1 if main::DEBUG;
3119# trace $_ if main::DEBUG && $to_trace;
3120# return if ! defined (my $line = readline $handle{$addr});
3121# chomp $line;
3122# push @{$added_lines{$addr}}, [ 0, $line ];
3123#
3124# $line =~ s/#.*//;
3125# $line =~ s/\s+$//;
3126# return $line if $line ne "";
3127# }
3128#
3129# return;
3130# }
3131
3132
3133 sub insert_lines {
3134 # Lines can be inserted so that it looks like they were in the input
3135 # file at the place it was when this routine is called. See also
3136 # insert_adjusted_lines(). Lines inserted via this routine go through
3137 # any each_line_handler()
3138
3139 my $self = shift;
3140
3141 # Each inserted line is an array, with the first element being 0 to
3142 # indicate that this line hasn't been adjusted, and needs to be
3143 # processed.
f998e60c 3144 no overloading;
051df77b 3145 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
3146 return;
3147 }
3148
3149 sub insert_adjusted_lines {
3150 # Lines can be inserted so that it looks like they were in the input
3151 # file at the place it was when this routine is called. See also
3152 # insert_lines(). Lines inserted via this routine are already fully
3153 # adjusted, ready to be processed; each_line_handler()s handlers will
3154 # not be called. This means this is not a completely general
3155 # facility, as only the last each_line_handler on the stack should
3156 # call this. It could be made more general, by passing to each of the
3157 # line_handlers their position on the stack, which they would pass on
3158 # to this routine, and that would replace the boolean first element in
3159 # the anonymous array pushed here, so that the next_line routine could
3160 # use that to call only those handlers whose index is after it on the
3161 # stack. But this is overkill for what is needed now.
3162
3163 my $self = shift;
3164 trace $_[0] if main::DEBUG && $to_trace;
3165
3166 # Each inserted line is an array, with the first element being 1 to
3167 # indicate that this line has been adjusted
f998e60c 3168 no overloading;
051df77b 3169 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
3170 return;
3171 }
3172
3173 sub get_missings {
3174 # Returns the stored up @missings lines' values, and clears the list.
3175 # The values are in an array, consisting of the default in the first
3176 # element, and the property in the 2nd. However, since these lines
3177 # can be stacked up, the return is an array of all these arrays.
3178
3179 my $self = shift;
3180 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3181
ffe43484 3182 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3183
3184 # If not accepting a list return, just return the first one.
3185 return shift @{$missings{$addr}} unless wantarray;
3186
3187 my @return = @{$missings{$addr}};
3188 undef @{$missings{$addr}};
3189 return @return;
3190 }
3191
a0204bb1
KW
3192 sub _exclude_unassigned {
3193
3194 # Takes the range in $_ and excludes code points that aren't assigned
3195 # in this release
3196
3197 state $skip_inserted_count = 0;
3198
3199 # Ignore recursive calls.
3200 if ($skip_inserted_count) {
3201 $skip_inserted_count--;
3202 return;
3203 }
3204
3205 # Find what code points are assigned in this release
3206 main::calculate_Assigned() if ! defined $Assigned;
3207
3208 my $self = shift;
3209 my $addr = do { no overloading; pack 'J', $self; };
3210 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3211
3212 my ($range, @remainder)
3213 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3214
3215 # Examine the range.
3216 if ($range =~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3217 {
3218 my $low = hex $1;
3219 my $high = (defined $2) ? hex $2 : $low;
3220
3221 # Split the range into subranges of just those code points in it
3222 # that are assigned.
3223 my @ranges = (Range_List->new(Initialize
3224 => Range->new($low, $high)) & $Assigned)->ranges;
3225
3226 # Do nothing if nothing in the original range is assigned in this
3227 # release; handle normally if everything is in this release.
3228 if (! @ranges) {
3229 $_ = "";
3230 }
3231 elsif (@ranges != 1) {
3232
3233 # Here, some code points in the original range aren't in this
3234 # release; @ranges gives the ones that are. Create fake input
3235 # lines for each of the ranges, and set things up so that when
3236 # this routine is called on that fake input, it will do
3237 # nothing.
3238 $skip_inserted_count = @ranges;
3239 my $remainder = join ";", @remainder;
3240 for my $range (@ranges) {
3241 $self->insert_lines(sprintf("%04X..%04X;%s",
3242 $range->start, $range->end, $remainder));
3243 }
3244 $_ = ""; # The original range is now defunct.
3245 }
3246 }
3247
3248 return;
3249 }
3250
3251 sub _fixup_obsolete_hanguls {
3252
3253 # This is called only when compiling Unicode version 1. All Unicode
3254 # data for subsequent releases assumes that the code points that were
3255 # Hangul syllables in this release only are something else, so if
3256 # using such data, we have to override it
3257
3258 my $self = shift;
3259 my $addr = do { no overloading; pack 'J', $self; };
3260 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3261
3262 my $object = main::property_ref($property{$addr});
53adf6a2
KW
3263 $object->add_map($FIRST_REMOVED_HANGUL_SYLLABLE,
3264 $FINAL_REMOVED_HANGUL_SYLLABLE,
a0204bb1
KW
3265 $early{$addr}[3], # Passed-in value for these
3266 Replace => $UNCONDITIONALLY);
3267 }
3268
99870f4d
KW
3269 sub _insert_property_into_line {
3270 # Add a property field to $_, if this file requires it.
3271
f998e60c 3272 my $self = shift;
ffe43484 3273 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 3274 my $property = $property{$addr};
99870f4d
KW
3275 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3276
3277 $_ =~ s/(;|$)/; $property$1/;
3278 return;
3279 }
3280
3281 sub carp_bad_line {
3282 # Output consistent error messages, using either a generic one, or the
3283 # one given by the optional parameter. To avoid gazillions of the
3284 # same message in case the syntax of a file is way off, this routine
3285 # only outputs the first instance of each message, incrementing a
3286 # count so the totals can be output at the end of the file.
3287
3288 my $self = shift;
3289 my $message = shift;
3290 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3291
ffe43484 3292 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3293
3294 $message = 'Unexpected line' unless $message;
3295
3296 # No trailing punctuation so as to fit with our addenda.
3297 $message =~ s/[.:;,]$//;
3298
3299 # If haven't seen this exact message before, output it now. Otherwise
3300 # increment the count of how many times it has occurred
3301 unless ($errors{$addr}->{$message}) {
3302 Carp::my_carp("$message in '$_' in "
f998e60c 3303 . $file{$addr}
99870f4d
KW
3304 . " at line $.. Skipping this line;");
3305 $errors{$addr}->{$message} = 1;
3306 }
3307 else {
3308 $errors{$addr}->{$message}++;
3309 }
3310
3311 # Clear the line to prevent any further (meaningful) processing of it.
3312 $_ = "";
3313
3314 return;
3315 }
3316} # End closure
3317
3318package Multi_Default;
3319
3320# Certain properties in early versions of Unicode had more than one possible
3321# default for code points missing from the files. In these cases, one
3322# default applies to everything left over after all the others are applied,
3323# and for each of the others, there is a description of which class of code
3324# points applies to it. This object helps implement this by storing the
3325# defaults, and for all but that final default, an eval string that generates
3326# the class that it applies to.
3327
3328
3329{ # Closure
3330
3331 main::setup_package();
3332
3333 my %class_defaults;
3334 # The defaults structure for the classes
3335 main::set_access('class_defaults', \%class_defaults);
3336
3337 my %other_default;
3338 # The default that applies to everything left over.
3339 main::set_access('other_default', \%other_default, 'r');
3340
3341
3342 sub new {
3343 # The constructor is called with default => eval pairs, terminated by
3344 # the left-over default. e.g.
3345 # Multi_Default->new(
3346 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
3347 # - 0x200D',
3348 # 'R' => 'some other expression that evaluates to code points',
3349 # .
3350 # .
3351 # .
3352 # 'U'));
68c3923f
KW
3353 # It is best to leave the final value be the one that matches the
3354 # above-Unicode code points.
99870f4d
KW
3355
3356 my $class = shift;
3357
3358 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 3359 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3360
3361 while (@_ > 1) {
3362 my $default = shift;
3363 my $eval = shift;
3364 $class_defaults{$addr}->{$default} = $eval;
3365 }
3366
3367 $other_default{$addr} = shift;
3368
3369 return $self;
3370 }
3371
3372 sub get_next_defaults {
3373 # Iterates and returns the next class of defaults.
3374 my $self = shift;
3375 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3376
ffe43484 3377 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3378
3379 return each %{$class_defaults{$addr}};
3380 }
3381}
3382
3383package Alias;
3384
3385# An alias is one of the names that a table goes by. This class defines them
3386# including some attributes. Everything is currently setup in the
3387# constructor.
3388
3389
3390{ # Closure
3391
3392 main::setup_package();
3393
3394 my %name;
3395 main::set_access('name', \%name, 'r');
3396
3397 my %loose_match;
c12f2655 3398 # Should this name match loosely or not.
99870f4d
KW
3399 main::set_access('loose_match', \%loose_match, 'r');
3400
33e96e72
KW
3401 my %make_re_pod_entry;
3402 # Some aliases should not get their own entries in the re section of the
3403 # pod, because they are covered by a wild-card, and some we want to
3404 # discourage use of. Binary
f82fe4ba 3405 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r', 's');
99870f4d 3406
fd1e3e84
KW
3407 my %ucd;
3408 # Is this documented to be accessible via Unicode::UCD
3409 main::set_access('ucd', \%ucd, 'r', 's');
3410
99870f4d
KW
3411 my %status;
3412 # Aliases have a status, like deprecated, or even suppressed (which means
3413 # they don't appear in documentation). Enum
3414 main::set_access('status', \%status, 'r');
3415
0eac1e20 3416 my %ok_as_filename;
99870f4d
KW
3417 # Similarly, some aliases should not be considered as usable ones for
3418 # external use, such as file names, or we don't want documentation to
3419 # recommend them. Boolean
0eac1e20 3420 main::set_access('ok_as_filename', \%ok_as_filename, 'r');
99870f4d
KW
3421
3422 sub new {
3423 my $class = shift;
3424
3425 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3426 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3427
3428 $name{$addr} = shift;
3429 $loose_match{$addr} = shift;
33e96e72 3430 $make_re_pod_entry{$addr} = shift;
0eac1e20 3431 $ok_as_filename{$addr} = shift;
99870f4d 3432 $status{$addr} = shift;
fd1e3e84 3433 $ucd{$addr} = shift;
99870f4d
KW
3434
3435 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3436
3437 # Null names are never ok externally
0eac1e20 3438 $ok_as_filename{$addr} = 0 if $name{$addr} eq "";
99870f4d
KW
3439
3440 return $self;
3441 }
3442}
3443
3444package Range;
3445
3446# A range is the basic unit for storing code points, and is described in the
3447# comments at the beginning of the program. Each range has a starting code
3448# point; an ending code point (not less than the starting one); a value
3449# that applies to every code point in between the two end-points, inclusive;
3450# and an enum type that applies to the value. The type is for the user's
3451# convenience, and has no meaning here, except that a non-zero type is
3452# considered to not obey the normal Unicode rules for having standard forms.
3453#
3454# The same structure is used for both map and match tables, even though in the
3455# latter, the value (and hence type) is irrelevant and could be used as a
3456# comment. In map tables, the value is what all the code points in the range
3457# map to. Type 0 values have the standardized version of the value stored as
3458# well, so as to not have to recalculate it a lot.
3459
3460sub trace { return main::trace(@_); }
3461
3462{ # Closure
3463
3464 main::setup_package();
3465
3466 my %start;
3467 main::set_access('start', \%start, 'r', 's');
3468
3469 my %end;
3470 main::set_access('end', \%end, 'r', 's');
3471
3472 my %value;
3473 main::set_access('value', \%value, 'r');
3474
3475 my %type;
3476 main::set_access('type', \%type, 'r');
3477
3478 my %standard_form;
3479 # The value in internal standard form. Defined only if the type is 0.
3480 main::set_access('standard_form', \%standard_form);
3481
3482 # Note that if these fields change, the dump() method should as well
3483
3484 sub new {
3485 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
3486 my $class = shift;
3487
3488 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3489 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3490
3491 $start{$addr} = shift;
3492 $end{$addr} = shift;
3493
3494 my %args = @_;
3495
3496 my $value = delete $args{'Value'}; # Can be 0
3497 $value = "" unless defined $value;
3498 $value{$addr} = $value;
3499
3500 $type{$addr} = delete $args{'Type'} || 0;
3501
3502 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3503
99870f4d
KW
3504 return $self;
3505 }
3506
3507 use overload
3508 fallback => 0,
3509 qw("") => "_operator_stringify",
3510 "." => \&main::_operator_dot,
1285127e 3511 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3512 ;
3513
3514 sub _operator_stringify {
3515 my $self = shift;
ffe43484 3516 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3517
3518 # Output it like '0041..0065 (value)'
3519 my $return = sprintf("%04X", $start{$addr})
3520 . '..'
3521 . sprintf("%04X", $end{$addr});
3522 my $value = $value{$addr};
3523 my $type = $type{$addr};
3524 $return .= ' (';
3525 $return .= "$value";
3526 $return .= ", Type=$type" if $type != 0;
3527 $return .= ')';
3528
3529 return $return;
3530 }
3531
3532 sub standard_form {
c292d35a
NC
3533 # Calculate the standard form only if needed, and cache the result.
3534 # The standard form is the value itself if the type is special.
3535 # This represents a considerable CPU and memory saving - at the time
3536 # of writing there are 368676 non-special objects, but the standard
3537 # form is only requested for 22047 of them - ie about 6%.
99870f4d
KW
3538
3539 my $self = shift;
3540 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3541
ffe43484 3542 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3543
3544 return $standard_form{$addr} if defined $standard_form{$addr};
c292d35a
NC
3545
3546 my $value = $value{$addr};
3547 return $value if $type{$addr};
3548 return $standard_form{$addr} = main::standardize($value);
99870f4d
KW
3549 }
3550
3551 sub dump {
3552 # Human, not machine readable. For machine readable, comment out this
3553 # entire routine and let the standard one take effect.
3554 my $self = shift;
3555 my $indent = shift;
3556 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3557
ffe43484 3558 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3559
3560 my $return = $indent
3561 . sprintf("%04X", $start{$addr})
3562 . '..'
3563 . sprintf("%04X", $end{$addr})
3564 . " '$value{$addr}';";
3565 if (! defined $standard_form{$addr}) {
3566 $return .= "(type=$type{$addr})";
3567 }
3568 elsif ($standard_form{$addr} ne $value{$addr}) {
3569 $return .= "(standard '$standard_form{$addr}')";
3570 }
3571 return $return;
3572 }
3573} # End closure
3574
3575package _Range_List_Base;
3576
3577# Base class for range lists. A range list is simply an ordered list of
3578# ranges, so that the ranges with the lowest starting numbers are first in it.
3579#
3580# When a new range is added that is adjacent to an existing range that has the
3581# same value and type, it merges with it to form a larger range.
3582#
3583# Ranges generally do not overlap, except that there can be multiple entries
3584# of single code point ranges. This is because of NameAliases.txt.
3585#
3586# In this program, there is a standard value such that if two different
3587# values, have the same standard value, they are considered equivalent. This
3588# value was chosen so that it gives correct results on Unicode data
3589
3590# There are a number of methods to manipulate range lists, and some operators
3591# are overloaded to handle them.
3592
99870f4d
KW
3593sub trace { return main::trace(@_); }
3594
3595{ # Closure
3596
3597 our $addr;
3598
5b348b71
KW
3599 # Max is initialized to a negative value that isn't adjacent to 0, for
3600 # simpler tests
3601 my $max_init = -2;
3602
99870f4d
KW
3603 main::setup_package();
3604
3605 my %ranges;
3606 # The list of ranges
3607 main::set_access('ranges', \%ranges, 'readable_array');
3608
3609 my %max;
3610 # The highest code point in the list. This was originally a method, but
3611 # actual measurements said it was used a lot.
3612 main::set_access('max', \%max, 'r');
3613
3614 my %each_range_iterator;
3615 # Iterator position for each_range()
3616 main::set_access('each_range_iterator', \%each_range_iterator);
3617
3618 my %owner_name_of;
3619 # Name of parent this is attached to, if any. Solely for better error
3620 # messages.
3621 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
3622
3623 my %_search_ranges_cache;
3624 # A cache of the previous result from _search_ranges(), for better
3625 # performance
3626 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
3627
3628 sub new {
3629 my $class = shift;
3630 my %args = @_;
3631
3632 # Optional initialization data for the range list.
3633 my $initialize = delete $args{'Initialize'};
3634
3635 my $self;
3636
3637 # Use _union() to initialize. _union() returns an object of this
3638 # class, which means that it will call this constructor recursively.
3639 # But it won't have this $initialize parameter so that it won't
3640 # infinitely loop on this.
3641 return _union($class, $initialize, %args) if defined $initialize;
3642
3643 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3644 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3645
3646 # Optional parent object, only for debug info.
3647 $owner_name_of{$addr} = delete $args{'Owner'};
3648 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
3649
3650 # Stringify, in case it is an object.
3651 $owner_name_of{$addr} = "$owner_name_of{$addr}";
3652
3653 # This is used only for error messages, and so a colon is added
3654 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
3655
3656 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3657
5b348b71 3658 $max{$addr} = $max_init;
99870f4d
KW
3659
3660 $_search_ranges_cache{$addr} = 0;
3661 $ranges{$addr} = [];
3662
3663 return $self;
3664 }
3665
3666 use overload
3667 fallback => 0,
3668 qw("") => "_operator_stringify",
3669 "." => \&main::_operator_dot,
1285127e 3670 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3671 ;
3672
3673 sub _operator_stringify {
3674 my $self = shift;
ffe43484 3675 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3676
3677 return "Range_List attached to '$owner_name_of{$addr}'"
3678 if $owner_name_of{$addr};
3679 return "anonymous Range_List " . \$self;
3680 }
3681
3682 sub _union {
3683 # Returns the union of the input code points. It can be called as
3684 # either a constructor or a method. If called as a method, the result
3685 # will be a new() instance of the calling object, containing the union
3686 # of that object with the other parameter's code points; if called as
d59563d0 3687 # a constructor, the first parameter gives the class that the new object
99870f4d
KW
3688 # should be, and the second parameter gives the code points to go into
3689 # it.
3690 # In either case, there are two parameters looked at by this routine;
3691 # any additional parameters are passed to the new() constructor.
3692 #
3693 # The code points can come in the form of some object that contains
3694 # ranges, and has a conventionally named method to access them; or
3695 # they can be an array of individual code points (as integers); or
3696 # just a single code point.
3697 #
3698 # If they are ranges, this routine doesn't make any effort to preserve
3198cc57
KW
3699 # the range values and types of one input over the other. Therefore
3700 # this base class should not allow _union to be called from other than
99870f4d
KW
3701 # initialization code, so as to prevent two tables from being added
3702 # together where the range values matter. The general form of this
3703 # routine therefore belongs in a derived class, but it was moved here
3704 # to avoid duplication of code. The failure to overload this in this
3705 # class keeps it safe.
3198cc57
KW
3706 #
3707 # It does make the effort during initialization to accept tables with
3708 # multiple values for the same code point, and to preserve the order
3709 # of these. If there is only one input range or range set, it doesn't