This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
mktables: Don't output anything above IV_MAX
[perl5.git] / lib / unicore / mktables
CommitLineData
d73e5302 1#!/usr/bin/perl -w
99870f4d
KW
2
3# !!!!!!!!!!!!!! IF YOU MODIFY THIS FILE !!!!!!!!!!!!!!!!!!!!!!!!!
4# Any files created or read by this program should be listed in 'mktables.lst'
5# Use -makelist to regenerate it.
6
9be22ca5
KW
7# There was an attempt when this was first rewritten to make it 5.8
8# compatible, but that has now been abandoned, and newer constructs are used
9# as convenient.
6c68572b 10
80f41172
JH
11# NOTE: this script can run quite slowly in older/slower systems.
12# It can also consume a lot of memory (128 MB or more), you may need
13# to raise your process resource limits (e.g. in bash, "ulimit -a"
14# to inspect, and "ulimit -d ..." or "ulimit -m ..." to set)
15
cdcef19a 16my $start_time;
98dc9551 17BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
18 # get it as close as possible
19 $start_time= time;
20}
21
23e33b60 22require 5.010_001;
d73e5302 23use strict;
99870f4d 24use warnings;
cf25bb62 25use Carp;
bd9ebcfd 26use Config;
99870f4d
KW
27use File::Find;
28use File::Path;
d07a55ed 29use File::Spec;
99870f4d 30use Text::Tabs;
6b64c11c 31use re "/aa";
0dee21d6 32use feature 'state';
99870f4d
KW
33
34sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 35my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d 36
74cd47d0
KW
37sub NON_ASCII_PLATFORM { ord("A") != 65 }
38
6295dc14
KW
39# When a new version of Unicode is published, unfortunately the algorithms for
40# dealing with various bounds, like \b{gcb}, \b{lb} may have to be updated
41# manually. The changes may or may not be backward compatible with older
42# releases. The code is in regen/mk_invlist.pl and regexec.c. Make the
43# changes, then come back here and set the variable below to what version the
44# code is expecting. If a newer version of Unicode is being compiled than
45# expected, a warning will be generated. If an older version is being
46# compiled, any bounds tests that fail in the generated test file (-maketest
47# option) will be marked as TODO.
c492f156 48my $version_of_mk_invlist_bounds = v10.0.0;
6295dc14 49
99870f4d
KW
50##########################################################################
51#
52# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
53# from the Unicode database files (lib/unicore/.../*.txt), It also generates
232ed87f 54# a pod file and .t files, depending on option parameters.
99870f4d
KW
55#
56# The structure of this file is:
57# First these introductory comments; then
58# code needed for everywhere, such as debugging stuff; then
59# code to handle input parameters; then
60# data structures likely to be of external interest (some of which depend on
61# the input parameters, so follows them; then
62# more data structures and subroutine and package (class) definitions; then
63# the small actual loop to process the input files and finish up; then
64# a __DATA__ section, for the .t tests
65#
232ed87f
KW
66# This program works on all releases of Unicode so far. The outputs have been
67# scrutinized most intently for release 5.1. The others have been checked for
68# somewhat more than just sanity. It can handle all non-provisional Unicode
69# character properties in those releases.
99870f4d 70#
99870f4d
KW
71# This program is mostly about Unicode character (or code point) properties.
72# A property describes some attribute or quality of a code point, like if it
73# is lowercase or not, its name, what version of Unicode it was first defined
74# in, or what its uppercase equivalent is. Unicode deals with these disparate
75# possibilities by making all properties into mappings from each code point
76# into some corresponding value. In the case of it being lowercase or not,
77# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
78# property maps each Unicode code point to a single value, called a "property
232ed87f
KW
79# value". (Some more recently defined properties, map a code point to a set
80# of values.)
99870f4d
KW
81#
82# When using a property in a regular expression, what is desired isn't the
83# mapping of the code point to its property's value, but the reverse (or the
84# mathematical "inverse relation"): starting with the property value, "Does a
85# code point map to it?" These are written in a "compound" form:
86# \p{property=value}, e.g., \p{category=punctuation}. This program generates
87# files containing the lists of code points that map to each such regular
88# expression property value, one file per list
89#
90# There is also a single form shortcut that Perl adds for many of the commonly
91# used properties. This happens for all binary properties, plus script,
92# general_category, and block properties.
93#
94# Thus the outputs of this program are files. There are map files, mostly in
95# the 'To' directory; and there are list files for use in regular expression
96# matching, all in subdirectories of the 'lib' directory, with each
97# subdirectory being named for the property that the lists in it are for.
98# Bookkeeping, test, and documentation files are also generated.
99
100my $matches_directory = 'lib'; # Where match (\p{}) files go.
101my $map_directory = 'To'; # Where map files go.
102
103# DATA STRUCTURES
104#
105# The major data structures of this program are Property, of course, but also
106# Table. There are two kinds of tables, very similar to each other.
107# "Match_Table" is the data structure giving the list of code points that have
108# a particular property value, mentioned above. There is also a "Map_Table"
109# data structure which gives the property's mapping from code point to value.
110# There are two structures because the match tables need to be combined in
111# various ways, such as constructing unions, intersections, complements, etc.,
112# and the map ones don't. And there would be problems, perhaps subtle, if
113# a map table were inadvertently operated on in some of those ways.
114# The use of separate classes with operations defined on one but not the other
115# prevents accidentally confusing the two.
116#
117# At the heart of each table's data structure is a "Range_List", which is just
118# an ordered list of "Ranges", plus ancillary information, and methods to
119# operate on them. A Range is a compact way to store property information.
120# Each range has a starting code point, an ending code point, and a value that
121# is meant to apply to all the code points between the two end points,
122# inclusive. For a map table, this value is the property value for those
123# code points. Two such ranges could be written like this:
124# 0x41 .. 0x5A, 'Upper',
125# 0x61 .. 0x7A, 'Lower'
126#
127# Each range also has a type used as a convenience to classify the values.
128# Most ranges in this program will be Type 0, or normal, but there are some
129# ranges that have a non-zero type. These are used only in map tables, and
130# are for mappings that don't fit into the normal scheme of things. Mappings
131# that require a hash entry to communicate with utf8.c are one example;
132# another example is mappings for charnames.pm to use which indicate a name
232ed87f 133# that is algorithmically determinable from its code point (and the reverse).
99870f4d
KW
134# These are used to significantly compact these tables, instead of listing
135# each one of the tens of thousands individually.
136#
137# In a match table, the value of a range is irrelevant (and hence the type as
138# well, which will always be 0), and arbitrarily set to the null string.
139# Using the example above, there would be two match tables for those two
140# entries, one named Upper would contain the 0x41..0x5A range, and the other
141# named Lower would contain 0x61..0x7A.
142#
143# Actually, there are two types of range lists, "Range_Map" is the one
144# associated with map tables, and "Range_List" with match tables.
232ed87f
KW
145# Again, this is so that methods can be defined on one and not the others so
146# as to prevent operating on them in incorrect ways.
99870f4d
KW
147#
148# Eventually, most tables are written out to files to be read by utf8_heavy.pl
149# in the perl core. All tables could in theory be written, but some are
150# suppressed because there is no current practical use for them. It is easy
151# to change which get written by changing various lists that are near the top
152# of the actual code in this file. The table data structures contain enough
153# ancillary information to allow them to be treated as separate entities for
154# writing, such as the path to each one's file. There is a heading in each
155# map table that gives the format of its entries, and what the map is for all
156# the code points missing from it. (This allows tables to be more compact.)
678f13d5 157#
99870f4d
KW
158# The Property data structure contains one or more tables. All properties
159# contain a map table (except the $perl property which is a
160# pseudo-property containing only match tables), and any properties that
161# are usable in regular expression matches also contain various matching
162# tables, one for each value the property can have. A binary property can
163# have two values, True and False (or Y and N, which are preferred by Unicode
164# terminology). Thus each of these properties will have a map table that
165# takes every code point and maps it to Y or N (but having ranges cuts the
166# number of entries in that table way down), and two match tables, one
167# which has a list of all the code points that map to Y, and one for all the
232ed87f 168# code points that map to N. (For each binary property, a third table is also
99870f4d 169# generated for the pseudo Perl property. It contains the identical code
232ed87f
KW
170# points as the Y table, but can be written in regular expressions, not in the
171# compound form, but in a "single" form like \p{IsUppercase}.) Many
172# properties are binary, but some properties have several possible values,
173# some have many, and properties like Name have a different value for every
174# named code point. Those will not, unless the controlling lists are changed,
175# have their match tables written out. But all the ones which can be used in
176# regular expression \p{} and \P{} constructs will. Prior to 5.14, generally
177# a property would have either its map table or its match tables written but
178# not both. Again, what gets written is controlled by lists which can easily
179# be changed. Starting in 5.14, advantage was taken of this, and all the map
180# tables needed to reconstruct the Unicode db are now written out, while
181# suppressing the Unicode .txt files that contain the data. Our tables are
182# much more compact than the .txt files, so a significant space savings was
183# achieved. Also, tables are not written out that are trivially derivable
184# from tables that do get written. So, there typically is no file containing
185# the code points not matched by a binary property (the table for \P{} versus
186# lowercase \p{}), since you just need to invert the True table to get the
187# False table.
188
189# Properties have a 'Type', like 'binary', or 'string', or 'enum' depending on
190# how many match tables there are and the content of the maps. This 'Type' is
c12f2655
KW
191# different than a range 'Type', so don't get confused by the two concepts
192# having the same name.
678f13d5 193#
99870f4d
KW
194# For information about the Unicode properties, see Unicode's UAX44 document:
195
196my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
197
198# As stated earlier, this program will work on any release of Unicode so far.
199# Most obvious problems in earlier data have NOT been corrected except when
be864b6c 200# necessary to make Perl or this program work reasonably, and to keep out
232ed87f
KW
201# potential security issues. For example, no folding information was given in
202# early releases, so this program substitutes lower case instead, just so that
203# a regular expression with the /i option will do something that actually
204# gives the right results in many cases. There are also a couple other
205# corrections for version 1.1.5, commented at the point they are made. As an
206# example of corrections that weren't made (but could be) is this statement
207# from DerivedAge.txt: "The supplementary private use code points and the
208# non-character code points were assigned in version 2.0, but not specifically
209# listed in the UCD until versions 3.0 and 3.1 respectively." (To be precise
210# it was 3.0.1 not 3.0.0) More information on Unicode version glitches is
211# further down in these introductory comments.
99870f4d 212#
232ed87f
KW
213# This program works on all non-provisional properties as of the current
214# Unicode release, though the files for some are suppressed for various
215# reasons. You can change which are output by changing lists in this program.
678f13d5 216#
dc85bd38 217# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
218# loose matchings rules (from Unicode TR18):
219#
220# The recommended names for UCD properties and property values are in
221# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
222# [PropValue]. There are both abbreviated names and longer, more
223# descriptive names. It is strongly recommended that both names be
224# recognized, and that loose matching of property names be used,
225# whereby the case distinctions, whitespace, hyphens, and underbar
226# are ignored.
232ed87f 227#
99870f4d
KW
228# The program still allows Fuzzy to override its determination of if loose
229# matching should be used, but it isn't currently used, as it is no longer
230# needed; the calculations it makes are good enough.
678f13d5 231#
99870f4d
KW
232# SUMMARY OF HOW IT WORKS:
233#
234# Process arguments
235#
236# A list is constructed containing each input file that is to be processed
237#
238# Each file on the list is processed in a loop, using the associated handler
239# code for each:
240# The PropertyAliases.txt and PropValueAliases.txt files are processed
241# first. These files name the properties and property values.
242# Objects are created of all the property and property value names
243# that the rest of the input should expect, including all synonyms.
244# The other input files give mappings from properties to property
245# values. That is, they list code points and say what the mapping
246# is under the given property. Some files give the mappings for
247# just one property; and some for many. This program goes through
232ed87f
KW
248# each file and populates the properties and their map tables from
249# them. Some properties are listed in more than one file, and
250# Unicode has set up a precedence as to which has priority if there
251# is a conflict. Thus the order of processing matters, and this
252# program handles the conflict possibility by processing the
253# overriding input files last, so that if necessary they replace
254# earlier values.
99870f4d
KW
255# After this is all done, the program creates the property mappings not
256# furnished by Unicode, but derivable from what it does give.
257# The tables of code points that match each property value in each
258# property that is accessible by regular expressions are created.
259# The Perl-defined properties are created and populated. Many of these
260# require data determined from the earlier steps
261# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 262# and Unicode are reconciled and warned about.
99870f4d
KW
263# All the properties are written to files
264# Any other files are written, and final warnings issued.
678f13d5 265#
99870f4d
KW
266# For clarity, a number of operators have been overloaded to work on tables:
267# ~ means invert (take all characters not in the set). The more
268# conventional '!' is not used because of the possibility of confusing
269# it with the actual boolean operation.
270# + means union
271# - means subtraction
272# & means intersection
273# The precedence of these is the order listed. Parentheses should be
274# copiously used. These are not a general scheme. The operations aren't
275# defined for a number of things, deliberately, to avoid getting into trouble.
276# Operations are done on references and affect the underlying structures, so
277# that the copy constructors for them have been overloaded to not return a new
278# clone, but the input object itself.
678f13d5 279#
99870f4d
KW
280# The bool operator is deliberately not overloaded to avoid confusion with
281# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
282#
283# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
284#
285# This program needs to be able to run under miniperl. Therefore, it uses a
286# minimum of other modules, and hence implements some things itself that could
287# be gotten from CPAN
288#
289# This program uses inputs published by the Unicode Consortium. These can
290# change incompatibly between releases without the Perl maintainers realizing
291# it. Therefore this program is now designed to try to flag these. It looks
292# at the directories where the inputs are, and flags any unrecognized files.
293# It keeps track of all the properties in the files it handles, and flags any
294# that it doesn't know how to handle. It also flags any input lines that
295# don't match the expected syntax, among other checks.
296#
297# It is also designed so if a new input file matches one of the known
298# templates, one hopefully just needs to add it to a list to have it
299# processed.
300#
301# As mentioned earlier, some properties are given in more than one file. In
302# particular, the files in the extracted directory are supposedly just
303# reformattings of the others. But they contain information not easily
9be22ca5
KW
304# derivable from the other files, including results for Unihan (which isn't
305# usually available to this program) and for unassigned code points. They
678f13d5
KW
306# also have historically had errors or been incomplete. In an attempt to
307# create the best possible data, this program thus processes them first to
308# glean information missing from the other files; then processes those other
309# files to override any errors in the extracted ones. Much of the design was
310# driven by this need to store things and then possibly override them.
311#
312# It tries to keep fatal errors to a minimum, to generate something usable for
313# testing purposes. It always looks for files that could be inputs, and will
314# warn about any that it doesn't know how to handle (the -q option suppresses
315# the warning).
99870f4d 316#
678f13d5
KW
317# Why is there more than one type of range?
318# This simplified things. There are some very specialized code points that
319# have to be handled specially for output, such as Hangul syllable names.
320# By creating a range type (done late in the development process), it
321# allowed this to be stored with the range, and overridden by other input.
322# Originally these were stored in another data structure, and it became a
323# mess trying to decide if a second file that was for the same property was
324# overriding the earlier one or not.
325#
326# Why are there two kinds of tables, match and map?
327# (And there is a base class shared by the two as well.) As stated above,
328# they actually are for different things. Development proceeded much more
329# smoothly when I (khw) realized the distinction. Map tables are used to
330# give the property value for every code point (actually every code point
331# that doesn't map to a default value). Match tables are used for regular
332# expression matches, and are essentially the inverse mapping. Separating
333# the two allows more specialized methods, and error checks so that one
334# can't just take the intersection of two map tables, for example, as that
335# is nonsensical.
99870f4d 336#
232ed87f
KW
337# What about 'fate' and 'status'. The concept of a table's fate was created
338# late when it became clear that something more was needed. The difference
339# between this and 'status' is unclean, and could be improved if someone
340# wanted to spend the effort.
341#
23e33b60
KW
342# DEBUGGING
343#
678f13d5
KW
344# This program is written so it will run under miniperl. Occasionally changes
345# will cause an error where the backtrace doesn't work well under miniperl.
346# To diagnose the problem, you can instead run it under regular perl, if you
347# have one compiled.
348#
349# There is a good trace facility. To enable it, first sub DEBUG must be set
350# to return true. Then a line like
351#
352# local $to_trace = 1 if main::DEBUG;
353#
232ed87f
KW
354# can be added to enable tracing in its lexical scope (plus dynamic) or until
355# you insert another line:
678f13d5
KW
356#
357# local $to_trace = 0 if main::DEBUG;
358#
232ed87f 359# To actually trace, use a line like "trace $a, @b, %c, ...;
678f13d5
KW
360#
361# Some of the more complex subroutines already have trace statements in them.
362# Permanent trace statements should be like:
363#
364# trace ... if main::DEBUG && $to_trace;
365#
fd2f7781
KW
366# main::stack_trace() will display what its name implies
367#
678f13d5
KW
368# If there is just one or a few files that you're debugging, you can easily
369# cause most everything else to be skipped. Change the line
370#
371# my $debug_skip = 0;
372#
373# to 1, and every file whose object is in @input_file_objects and doesn't have
232ed87f
KW
374# a, 'non_skip => 1,' in its constructor will be skipped. However, skipping
375# Jamo.txt or UnicodeData.txt will likely cause fatal errors.
678f13d5 376#
b4a0206c 377# To compare the output tables, it may be useful to specify the -annotate
97a8a595
KW
378# flag. (As of this writing, this can't be done on a clean workspace, due to
379# requirements in Text::Tabs used in this option; so first run mktables
380# without this option.) This option adds comment lines to each table, one for
381# each non-algorithmically named character giving, currently its code point,
382# name, and graphic representation if printable (and you have a font that
383# knows about it). This makes it easier to see what the particular code
384# points are in each output table. Non-named code points are annotated with a
385# description of their status, and contiguous ones with the same description
386# will be output as a range rather than individually. Algorithmically named
387# characters are also output as ranges, except when there are just a few
388# contiguous ones.
c4019d52 389#
99870f4d
KW
390# FUTURE ISSUES
391#
392# The program would break if Unicode were to change its names so that
393# interior white space, underscores, or dashes differences were significant
394# within property and property value names.
395#
396# It might be easier to use the xml versions of the UCD if this program ever
397# would need heavy revision, and the ability to handle old versions was not
398# required.
399#
400# There is the potential for name collisions, in that Perl has chosen names
401# that Unicode could decide it also likes. There have been such collisions in
402# the past, with mostly Perl deciding to adopt the Unicode definition of the
403# name. However in the 5.2 Unicode beta testing, there were a number of such
404# collisions, which were withdrawn before the final release, because of Perl's
405# and other's protests. These all involved new properties which began with
406# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
407# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
408# Unicode document, so they are unlikely to be used by Unicode for another
409# purpose. However, they might try something beginning with 'In', or use any
410# of the other Perl-defined properties. This program will warn you of name
411# collisions, and refuse to generate tables with them, but manual intervention
412# will be required in this event. One scheme that could be implemented, if
413# necessary, would be to have this program generate another file, or add a
414# field to mktables.lst that gives the date of first definition of a property.
415# Each new release of Unicode would use that file as a basis for the next
416# iteration. And the Perl synonym addition code could sort based on the age
417# of the property, so older properties get priority, and newer ones that clash
418# would be refused; hence existing code would not be impacted, and some other
419# synonym would have to be used for the new property. This is ugly, and
420# manual intervention would certainly be easier to do in the short run; lets
421# hope it never comes to this.
678f13d5 422#
99870f4d
KW
423# A NOTE ON UNIHAN
424#
44b0f8c1 425# This program can generate tables from the Unihan database. But that DB
b9e57660
KW
426# isn't normally available, so it is marked as optional. Prior to version
427# 5.2, this database was in a single file, Unihan.txt. In 5.2 the database
428# was split into 8 different files, all beginning with the letters 'Unihan'.
429# If you plunk those files down into the directory mktables ($0) is in, this
430# program will read them and automatically create tables for the properties
431# from it that are listed in PropertyAliases.txt and PropValueAliases.txt,
432# plus any you add to the @cjk_properties array and the @cjk_property_values
433# array, being sure to add necessary '# @missings' lines to the latter. For
434# Unicode versions earlier than 5.2, most of the Unihan properties are not
435# listed at all in PropertyAliases nor PropValueAliases. This program assumes
436# for these early releases that you want the properties that are specified in
437# the 5.2 release.
99870f4d
KW
438#
439# You may need to adjust the entries to suit your purposes. setup_unihan(),
440# and filter_unihan_line() are the functions where this is done. This program
441# already does some adjusting to make the lines look more like the rest of the
442# Unicode DB; You can see what that is in filter_unihan_line()
443#
444# There is a bug in the 3.2 data file in which some values for the
445# kPrimaryNumeric property have commas and an unexpected comment. A filter
9be22ca5
KW
446# could be added to correct these; or for a particular installation, the
447# Unihan.txt file could be edited to fix them.
99870f4d 448#
678f13d5
KW
449# HOW TO ADD A FILE TO BE PROCESSED
450#
451# A new file from Unicode needs to have an object constructed for it in
452# @input_file_objects, probably at the end or at the end of the extracted
453# ones. The program should warn you if its name will clash with others on
454# restrictive file systems, like DOS. If so, figure out a better name, and
455# add lines to the README.perl file giving that. If the file is a character
232ed87f 456# property, it should be in the format that Unicode has implicitly
678f13d5
KW
457# standardized for such files for the more recently introduced ones.
458# If so, the Input_file constructor for @input_file_objects can just be the
459# file name and release it first appeared in. If not, then it should be
460# possible to construct an each_line_handler() to massage the line into the
461# standardized form.
462#
463# For non-character properties, more code will be needed. You can look at
464# the existing entries for clues.
465#
466# UNICODE VERSIONS NOTES
467#
468# The Unicode UCD has had a number of errors in it over the versions. And
469# these remain, by policy, in the standard for that version. Therefore it is
470# risky to correct them, because code may be expecting the error. So this
471# program doesn't generally make changes, unless the error breaks the Perl
472# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
473# for U+1105, which causes real problems for the algorithms for Jamo
474# calculations, so it is changed here.
475#
476# But it isn't so clear cut as to what to do about concepts that are
477# introduced in a later release; should they extend back to earlier releases
478# where the concept just didn't exist? It was easier to do this than to not,
479# so that's what was done. For example, the default value for code points not
480# in the files for various properties was probably undefined until changed by
481# some version. No_Block for blocks is such an example. This program will
482# assign No_Block even in Unicode versions that didn't have it. This has the
483# benefit that code being written doesn't have to special case earlier
484# versions; and the detriment that it doesn't match the Standard precisely for
485# the affected versions.
486#
487# Here are some observations about some of the issues in early versions:
488#
232ed87f
KW
489# Prior to version 3.0, there were 3 character decompositions. These are not
490# handled by Unicode::Normalize, nor will it compile when presented a version
491# that has them. However, you can trivially get it to compile by simply
492# ignoring those decompositions, by changing the croak to a carp. At the time
3baae3fa
KW
493# of this writing, the line (in dist/Unicode-Normalize/Normalize.pm or
494# dist/Unicode-Normalize/mkheader) reads
232ed87f
KW
495#
496# croak("Weird Canonical Decomposition of U+$h");
497#
28807e1d 498# Simply comment it out. It will compile, but will not know about any three
9be22ca5 499# character decompositions.
232ed87f
KW
500
501# The number of code points in \p{alpha=True} halved in 2.1.9. It turns out
502# that the reason is that the CJK block starting at 4E00 was removed from
503# PropList, and was not put back in until 3.1.0. The Perl extension (the
504# single property name \p{alpha}) has the correct values. But the compound
505# form is simply not generated until 3.1, as it can be argued that prior to
506# this release, this was not an official property. The comments for
507# filter_old_style_proplist() give more details.
678f13d5
KW
508#
509# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
510# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
511# reason is that 3.2 introduced U+205F=medium math space, which was not
512# classed as white space, but Perl figured out that it should have been. 4.0
513# reclassified it correctly.
514#
515# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
232ed87f
KW
516# this was erroneously a synonym for 202 (it should be 200). In 4.0, ATB
517# became 202, and ATBL was left with no code points, as all the ones that
518# mapped to 202 stayed mapped to 202. Thus if your program used the numeric
519# name for the class, it would not have been affected, but if it used the
520# mnemonic, it would have been.
678f13d5 521#
9be22ca5 522# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that, code
678f13d5
KW
523# points which eventually came to have this script property value, instead
524# mapped to "Unknown". But in the next release all these code points were
525# moved to \p{sc=common} instead.
9be22ca5
KW
526
527# The tests furnished by Unicode for testing WordBreak and SentenceBreak
528# generate errors in 5.0 and earlier.
99870f4d
KW
529#
530# The default for missing code points for BidiClass is complicated. Starting
531# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
532# tries to do the best it can for earlier releases. It is done in
533# process_PropertyAliases()
534#
232ed87f
KW
535# In version 2.1.2, the entry in UnicodeData.txt:
536# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;;019F;
537# should instead be
538# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;019F;;019F
539# Without this change, there are casing problems for this character.
540#
7803ad2d
KW
541# Search for $string_compare_versions to see how to compare changes to
542# properties between Unicode versions
543#
99870f4d
KW
544##############################################################################
545
546my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
547 # and errors
548my $MAX_LINE_WIDTH = 78;
549
550# Debugging aid to skip most files so as to not be distracted by them when
551# concentrating on the ones being debugged. Add
552# non_skip => 1,
553# to the constructor for those files you want processed when you set this.
554# Files with a first version number of 0 are special: they are always
c12f2655
KW
555# processed regardless of the state of this flag. Generally, Jamo.txt and
556# UnicodeData.txt must not be skipped if you want this program to not die
557# before normal completion.
99870f4d
KW
558my $debug_skip = 0;
559
e9c4b4f8
KW
560
561# Normally these are suppressed.
562my $write_Unicode_deprecated_tables = 0;
563
99870f4d
KW
564# Set to 1 to enable tracing.
565our $to_trace = 0;
566
567{ # Closure for trace: debugging aid
568 my $print_caller = 1; # ? Include calling subroutine name
569 my $main_with_colon = 'main::';
570 my $main_colon_length = length($main_with_colon);
571
572 sub trace {
573 return unless $to_trace; # Do nothing if global flag not set
574
575 my @input = @_;
576
577 local $DB::trace = 0;
578 $DB::trace = 0; # Quiet 'used only once' message
579
580 my $line_number;
581
582 # Loop looking up the stack to get the first non-trace caller
583 my $caller_line;
584 my $caller_name;
585 my $i = 0;
586 do {
587 $line_number = $caller_line;
588 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
589 $caller = $main_with_colon unless defined $caller;
590
591 $caller_name = $caller;
592
593 # get rid of pkg
594 $caller_name =~ s/.*:://;
595 if (substr($caller_name, 0, $main_colon_length)
596 eq $main_with_colon)
597 {
598 $caller_name = substr($caller_name, $main_colon_length);
599 }
600
601 } until ($caller_name ne 'trace');
602
603 # If the stack was empty, we were called from the top level
604 $caller_name = 'main' if ($caller_name eq ""
605 || $caller_name eq 'trace');
606
607 my $output = "";
9346f59a 608 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
99870f4d 609 foreach my $string (@input) {
99870f4d
KW
610 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
611 $output .= simple_dumper($string);
612 }
613 else {
614 $string = "$string" if ref $string;
615 $string = $UNDEF unless defined $string;
616 chomp $string;
617 $string = '""' if $string eq "";
618 $output .= " " if $output ne ""
619 && $string ne ""
620 && substr($output, -1, 1) ne " "
621 && substr($string, 0, 1) ne " ";
622 $output .= $string;
623 }
624 }
625
99f78760
KW
626 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
627 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
628 print STDERR $output, "\n";
629 return;
630 }
631}
632
fd2f7781
KW
633sub stack_trace() {
634 local $to_trace = 1 if main::DEBUG;
635 my $line = (caller(0))[2];
636 my $i = 1;
637
638 # Accumulate the stack trace
639 while (1) {
640 my ($pkg, $file, $caller_line, $caller) = caller $i++;
641
642 last unless defined $caller;
643
644 trace "called from $caller() at line $line";
645 $line = $caller_line;
646 }
647}
648
99870f4d
KW
649# This is for a rarely used development feature that allows you to compare two
650# versions of the Unicode standard without having to deal with changes caused
9be22ca5 651# by the code points introduced in the later version. You probably also want
8f21889c
KW
652# to use the -annotate option when using this. Run this program on a unicore
653# containing the starting release you want to compare. Save that output
44b0f8c1 654# structure. Then, switching to a unicore with the ending release, change the
8f21889c
KW
655# 0 in the $string_compare_versions definition just below to a string
656# containing a SINGLE dotted Unicode release number (e.g. "2.1") corresponding
657# to the starting release. This program will then compile, but throw away all
658# code points introduced after the starting release. Finally use a diff tool
659# to compare the two directory structures. They include only the code points
660# common to both releases, and you can see the changes caused just by the
661# underlying release semantic changes. For versions earlier than 3.2, you
662# must copy a version of DAge.txt into the directory.
272af058 663my $string_compare_versions = DEBUG && ""; # e.g., "2.1";
99870f4d
KW
664my $compare_versions = DEBUG
665 && $string_compare_versions
666 && pack "C*", split /\./, $string_compare_versions;
667
668sub uniques {
669 # Returns non-duplicated input values. From "Perl Best Practices:
670 # Encapsulated Cleverness". p. 455 in first edition.
671
672 my %seen;
0e407844
NC
673 # Arguably this breaks encapsulation, if the goal is to permit multiple
674 # distinct objects to stringify to the same value, and be interchangeable.
675 # However, for this program, no two objects stringify identically, and all
676 # lists passed to this function are either objects or strings. So this
677 # doesn't affect correctness, but it does give a couple of percent speedup.
678 no overloading;
99870f4d
KW
679 return grep { ! $seen{$_}++ } @_;
680}
681
682$0 = File::Spec->canonpath($0);
683
684my $make_test_script = 0; # ? Should we output a test script
6b5ab373 685my $make_norm_test_script = 0; # ? Should we output a normalization test script
99870f4d
KW
686my $write_unchanged_files = 0; # ? Should we update the output files even if
687 # we don't think they have changed
688my $use_directory = ""; # ? Should we chdir somewhere.
689my $pod_directory; # input directory to store the pod file.
690my $pod_file = 'perluniprops';
691my $t_path; # Path to the .t test file
692my $file_list = 'mktables.lst'; # File to store input and output file names.
693 # This is used to speed up the build, by not
694 # executing the main body of the program if
695 # nothing on the list has changed since the
696 # previous build
697my $make_list = 1; # ? Should we write $file_list. Set to always
698 # make a list so that when the pumpking is
699 # preparing a release, s/he won't have to do
700 # special things
701my $glob_list = 0; # ? Should we try to include unknown .txt files
702 # in the input.
bd9ebcfd
KW
703my $output_range_counts = $debugging_build; # ? Should we include the number
704 # of code points in ranges in
705 # the output
558712cf 706my $annotate = 0; # ? Should character names be in the output
9ef2b94f 707
99870f4d
KW
708# Verbosity levels; 0 is quiet
709my $NORMAL_VERBOSITY = 1;
710my $PROGRESS = 2;
711my $VERBOSE = 3;
712
713my $verbosity = $NORMAL_VERBOSITY;
714
0458fbc1
KW
715# Stored in mktables.lst so that if this program is called with different
716# options, will regenerate even if the files otherwise look like they're
717# up-to-date.
718my $command_line_arguments = join " ", @ARGV;
719
99870f4d
KW
720# Process arguments
721while (@ARGV) {
cf25bb62
JH
722 my $arg = shift @ARGV;
723 if ($arg eq '-v') {
99870f4d
KW
724 $verbosity = $VERBOSE;
725 }
726 elsif ($arg eq '-p') {
727 $verbosity = $PROGRESS;
728 $| = 1; # Flush buffers as we go.
729 }
730 elsif ($arg eq '-q') {
731 $verbosity = 0;
732 }
733 elsif ($arg eq '-w') {
44b0f8c1
SF
734 # update the files even if they haven't changed
735 $write_unchanged_files = 1;
99870f4d
KW
736 }
737 elsif ($arg eq '-check') {
6ae7e459
YO
738 my $this = shift @ARGV;
739 my $ok = shift @ARGV;
740 if ($this ne $ok) {
741 print "Skipping as check params are not the same.\n";
742 exit(0);
743 }
00a8df5c 744 }
99870f4d
KW
745 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
746 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
747 }
3df51b85
KW
748 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
749 {
99870f4d 750 $make_test_script = 1;
99870f4d 751 }
6b5ab373
KW
752 elsif ($arg eq '-makenormtest')
753 {
754 $make_norm_test_script = 1;
755 }
99870f4d
KW
756 elsif ($arg eq '-makelist') {
757 $make_list = 1;
758 }
759 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
760 -d $use_directory or croak "Unknown directory '$use_directory'";
761 }
762 elsif ($arg eq '-L') {
763
764 # Existence not tested until have chdir'd
765 $file_list = shift;
766 }
767 elsif ($arg eq '-globlist') {
768 $glob_list = 1;
769 }
770 elsif ($arg eq '-c') {
771 $output_range_counts = ! $output_range_counts
772 }
b4a0206c 773 elsif ($arg eq '-annotate') {
558712cf 774 $annotate = 1;
bd9ebcfd
KW
775 $debugging_build = 1;
776 $output_range_counts = 1;
9ef2b94f 777 }
99870f4d
KW
778 else {
779 my $with_c = 'with';
780 $with_c .= 'out' if $output_range_counts; # Complements the state
781 croak <<END;
782usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
783 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
784 [-check A B ]
785 -c : Output comments $with_c number of code points in ranges
786 -q : Quiet Mode: Only output serious warnings.
787 -p : Set verbosity level to normal plus show progress.
788 -v : Set Verbosity level high: Show progress and non-serious
789 warnings
790 -w : Write files regardless
791 -C dir : Change to this directory before proceeding. All relative paths
792 except those specified by the -P and -T options will be done
793 with respect to this directory.
794 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 795 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
796 -L filelist : Use alternate 'filelist' instead of standard one
797 -globlist : Take as input all non-Test *.txt files in current and sub
798 directories
3df51b85
KW
799 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
800 overrides -T
99870f4d 801 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 802 -annotate : Output an annotation for each character in the table files;
97a8a595
KW
803 useful for debugging mktables, looking at diffs; but is slow
804 and memory intensive
99870f4d
KW
805 -check A B : Executes $0 only if A and B are the same
806END
807 }
808}
809
810# Stores the most-recently changed file. If none have changed, can skip the
811# build
aeab6150 812my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
813
814# Change directories now, because need to read 'version' early.
815if ($use_directory) {
3df51b85 816 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
817 $pod_directory = File::Spec->rel2abs($pod_directory);
818 }
3df51b85 819 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 820 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 821 }
99870f4d 822 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 823 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 824 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 825 }
3df51b85 826 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 827 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 828 }
00a8df5c
YO
829}
830
99870f4d
KW
831# Get Unicode version into regular and v-string. This is done now because
832# various tables below get populated based on it. These tables are populated
833# here to be near the top of the file, and so easily seeable by those needing
834# to modify things.
835open my $VERSION, "<", "version"
836 or croak "$0: can't open required file 'version': $!\n";
837my $string_version = <$VERSION>;
838close $VERSION;
839chomp $string_version;
840my $v_version = pack "C*", split /\./, $string_version; # v string
841
8f21889c
KW
842my $unicode_version = ($compare_versions)
843 ? ( "$string_compare_versions (using "
844 . "$string_version rules)")
845 : $string_version;
846
99870f4d
KW
847# The following are the complete names of properties with property values that
848# are known to not match any code points in some versions of Unicode, but that
849# may change in the future so they should be matchable, hence an empty file is
850# generated for them.
a9c9e371
KW
851my @tables_that_may_be_empty;
852push @tables_that_may_be_empty, 'Joining_Type=Left_Joining'
853 if $v_version lt v6.3.0;
99870f4d
KW
854push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
855push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
856push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
857 if $v_version ge v4.1.0;
82aed44a
KW
858push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
859 if $v_version ge v6.0.0;
f583b44c
KW
860push @tables_that_may_be_empty, 'Grapheme_Cluster_Break=Prepend'
861 if $v_version ge v6.1.0;
1e958ea9
KW
862push @tables_that_may_be_empty, 'Canonical_Combining_Class=CCC133'
863 if $v_version ge v6.2.0;
99870f4d
KW
864
865# The lists below are hashes, so the key is the item in the list, and the
866# value is the reason why it is in the list. This makes generation of
867# documentation easier.
868
869my %why_suppressed; # No file generated for these.
870
871# Files aren't generated for empty extraneous properties. This is arguable.
872# Extraneous properties generally come about because a property is no longer
873# used in a newer version of Unicode. If we generated a file without code
874# points, programs that used to work on that property will still execute
875# without errors. It just won't ever match (or will always match, with \P{}).
876# This means that the logic is now likely wrong. I (khw) think its better to
877# find this out by getting an error message. Just move them to the table
878# above to change this behavior
879my %why_suppress_if_empty_warn_if_not = (
880
881 # It is the only property that has ever officially been removed from the
882 # Standard. The database never contained any code points for it.
883 'Special_Case_Condition' => 'Obsolete',
884
885 # Apparently never official, but there were code points in some versions of
886 # old-style PropList.txt
887 'Non_Break' => 'Obsolete',
888);
889
890# These would normally go in the warn table just above, but they were changed
891# a long time before this program was written, so warnings about them are
892# moot.
893if ($v_version gt v3.2.0) {
894 push @tables_that_may_be_empty,
895 'Canonical_Combining_Class=Attached_Below_Left'
896}
897
9be22ca5
KW
898# Enum values for to_output_map() method in the Map_Table package. (0 is don't
899# output)
272501f6
KW
900my $EXTERNAL_MAP = 1;
901my $INTERNAL_MAP = 2;
ce712c88 902my $OUTPUT_ADJUSTED = 3;
272501f6 903
fcf1973c
KW
904# To override computed values for writing the map tables for these properties.
905# The default for enum map tables is to write them out, so that the Unicode
906# .txt files can be removed, but all the data to compute any property value
907# for any code point is available in a more compact form.
908my %global_to_output_map = (
909 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
910 # get stuck supporting it if things change. Since it is a STRING
911 # property, it normally would be listed in the pod, but INTERNAL_MAP
912 # suppresses that.
fcf1973c
KW
913 Unicode_1_Name => $INTERNAL_MAP,
914
915 Present_In => 0, # Suppress, as easily computed from Age
74cd47d0
KW
916 Block => (NON_ASCII_PLATFORM) ? 1 : 0, # Suppress, as Blocks.txt is
917 # retained, but needed for
918 # non-ASCII
53d34b6c
KW
919
920 # Suppress, as mapping can be found instead from the
921 # Perl_Decomposition_Mapping file
922 Decomposition_Type => 0,
fcf1973c
KW
923);
924
99870f4d
KW
925# There are several types of obsolete properties defined by Unicode. These
926# must be hand-edited for every new Unicode release.
927my %why_deprecated; # Generates a deprecated warning message if used.
928my %why_stabilized; # Documentation only
929my %why_obsolete; # Documentation only
930
931{ # Closure
8364f9eb 932 my $simple = 'Perl uses the more complete version';
99870f4d
KW
933 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
934
935 my $other_properties = 'other properties';
936 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 937 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
938
939 %why_deprecated = (
5f7264c7 940 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
941 'Jamo_Short_Name' => $contributory,
942 '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',
943 'Other_Alphabetic' => $contributory,
944 'Other_Default_Ignorable_Code_Point' => $contributory,
945 'Other_Grapheme_Extend' => $contributory,
946 'Other_ID_Continue' => $contributory,
947 'Other_ID_Start' => $contributory,
948 'Other_Lowercase' => $contributory,
949 'Other_Math' => $contributory,
950 'Other_Uppercase' => $contributory,
e22aaf5c
KW
951 'Expands_On_NFC' => $why_no_expand,
952 'Expands_On_NFD' => $why_no_expand,
953 'Expands_On_NFKC' => $why_no_expand,
954 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
955 );
956
957 %why_suppressed = (
5f7264c7 958 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
959 # contains the same information, but without the algorithmically
960 # determinable Hangul syllables'. This file is not published, so it's
961 # existence is not noted in the comment.
12fee290 962 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize or prop_invmap() or charprop() in Unicode::UCD::',
99870f4d 963
5f8d1a89
KW
964 # Don't suppress ISO_Comment, as otherwise special handling is needed
965 # to differentiate between it and gc=c, which can be written as 'isc',
966 # which is the same characters as ISO_Comment's short name.
99870f4d 967
12fee290 968 'Name' => "Accessible via \\N{...} or 'use charnames;' or charprop() or prop_invmap() in Unicode::UCD::",
e0b29447 969
12fee290
KW
970 'Simple_Case_Folding' => "$simple. Can access this through casefold(), charprop(), or prop_invmap() in Unicode::UCD",
971 'Simple_Lowercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
972 'Simple_Titlecase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
973 'Simple_Uppercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
99870f4d 974
dac6f618 975 FC_NFKC_Closure => 'Deprecated by Unicode, and supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
976 );
977
1704a0ea
KW
978 foreach my $property (
979
980 # The following are suppressed because they were made contributory
981 # or deprecated by Unicode before Perl ever thought about
982 # supporting them.
983 'Jamo_Short_Name',
984 'Grapheme_Link',
985 'Expands_On_NFC',
986 'Expands_On_NFD',
987 'Expands_On_NFKC',
988 'Expands_On_NFKD',
989
990 # The following are suppressed because they have been marked
991 # as deprecated for a sufficient amount of time
992 'Other_Alphabetic',
993 'Other_Default_Ignorable_Code_Point',
994 'Other_Grapheme_Extend',
995 'Other_ID_Continue',
996 'Other_ID_Start',
997 'Other_Lowercase',
998 'Other_Math',
999 'Other_Uppercase',
e22aaf5c 1000 ) {
99870f4d
KW
1001 $why_suppressed{$property} = $why_deprecated{$property};
1002 }
cf25bb62 1003
99870f4d
KW
1004 # Customize the message for all the 'Other_' properties
1005 foreach my $property (keys %why_deprecated) {
1006 next if (my $main_property = $property) !~ s/^Other_//;
1007 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
1008 }
1009}
1010
e9c4b4f8
KW
1011if ($write_Unicode_deprecated_tables) {
1012 foreach my $property (keys %why_suppressed) {
1013 delete $why_suppressed{$property} if $property =~
1014 / ^ Other | Grapheme /x;
1015 }
1016}
1017
99870f4d
KW
1018if ($v_version ge 4.0.0) {
1019 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
1020 if ($v_version ge 6.0.0) {
1021 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
1022 }
99870f4d 1023}
5f7264c7 1024if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 1025 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 1026 if ($v_version ge 6.0.0) {
63f74647 1027 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 1028 }
99870f4d
KW
1029}
1030
1031# Probably obsolete forever
1032if ($v_version ge v4.1.0) {
82aed44a
KW
1033 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
1034}
1035if ($v_version ge v6.0.0) {
caa75395 1036 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana" (or both)';
2b352efd 1037 $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
1038}
1039
1040# This program can create files for enumerated-like properties, such as
1041# 'Numeric_Type'. This file would be the same format as for a string
1042# property, with a mapping from code point to its value, so you could look up,
1043# for example, the script a code point is in. But no one so far wants this
1044# mapping, or they have found another way to get it since this is a new
1045# feature. So no file is generated except if it is in this list.
1046my @output_mapped_properties = split "\n", <<END;
1047END
1048
b9e57660
KW
1049# If you want more Unihan properties than the default, you need to add them to
1050# these arrays. Depending on the property type, @missing lines might have to
1051# be added to the second array. A sample entry would be (including the '#'):
1052# @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
99870f4d 1053my @cjk_properties = split "\n", <<'END';
99870f4d 1054END
99870f4d 1055my @cjk_property_values = split "\n", <<'END';
99870f4d
KW
1056END
1057
1058# The input files don't list every code point. Those not listed are to be
1059# defaulted to some value. Below are hard-coded what those values are for
1060# non-binary properties as of 5.1. Starting in 5.0, there are
caa75395 1061# machine-parsable comment lines in the files that give the defaults; so this
99870f4d
KW
1062# list shouldn't have to be extended. The claim is that all missing entries
1063# for binary properties will default to 'N'. Unicode tried to change that in
1064# 5.2, but the beta period produced enough protest that they backed off.
1065#
1066# The defaults for the fields that appear in UnicodeData.txt in this hash must
1067# be in the form that it expects. The others may be synonyms.
1068my $CODE_POINT = '<code point>';
1069my %default_mapping = (
1070 Age => "Unassigned",
1071 # Bidi_Class => Complicated; set in code
1072 Bidi_Mirroring_Glyph => "",
1073 Block => 'No_Block',
1074 Canonical_Combining_Class => 0,
1075 Case_Folding => $CODE_POINT,
1076 Decomposition_Mapping => $CODE_POINT,
1077 Decomposition_Type => 'None',
1078 East_Asian_Width => "Neutral",
1079 FC_NFKC_Closure => $CODE_POINT,
ec2cd995 1080 General_Category => ($v_version le 6.3.0) ? 'Cn' : 'Unassigned',
99870f4d
KW
1081 Grapheme_Cluster_Break => 'Other',
1082 Hangul_Syllable_Type => 'NA',
1083 ISO_Comment => "",
1084 Jamo_Short_Name => "",
1085 Joining_Group => "No_Joining_Group",
1086 # Joining_Type => Complicated; set in code
1087 kIICore => 'N', # Is converted to binary
1088 #Line_Break => Complicated; set in code
1089 Lowercase_Mapping => $CODE_POINT,
1090 Name => "",
1091 Name_Alias => "",
1092 NFC_QC => 'Yes',
1093 NFD_QC => 'Yes',
1094 NFKC_QC => 'Yes',
1095 NFKD_QC => 'Yes',
1096 Numeric_Type => 'None',
1097 Numeric_Value => 'NaN',
1098 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1099 Sentence_Break => 'Other',
1100 Simple_Case_Folding => $CODE_POINT,
1101 Simple_Lowercase_Mapping => $CODE_POINT,
1102 Simple_Titlecase_Mapping => $CODE_POINT,
1103 Simple_Uppercase_Mapping => $CODE_POINT,
1104 Titlecase_Mapping => $CODE_POINT,
1105 Unicode_1_Name => "",
1106 Unicode_Radical_Stroke => "",
1107 Uppercase_Mapping => $CODE_POINT,
1108 Word_Break => 'Other',
1109);
1110
678f13d5 1111### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1112
1113my $HEADER=<<"EOF";
1114# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85 1115# This file is machine-generated by $0 from the Unicode
8f21889c 1116# database, Version $unicode_version. Any changes made here will be lost!
cf25bb62
JH
1117EOF
1118
126c3d4e 1119my $INTERNAL_ONLY_HEADER = <<"EOF";
99870f4d
KW
1120
1121# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1122# This file is for internal use by core Perl only. The format and even the
1123# name or existence of this file are subject to change without notice. Don't
d9ae3878
KW
1124# use it directly. Use Unicode::UCD to access the Unicode character data
1125# base.
99870f4d
KW
1126EOF
1127
1128my $DEVELOPMENT_ONLY=<<"EOF";
1129# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1130# This file contains information artificially constrained to code points
1131# present in Unicode release $string_compare_versions.
1132# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1133# not be used for production.
b6922eda
KW
1134
1135EOF
1136
346a20cf
KW
1137my $MAX_UNICODE_CODEPOINT_STRING = ($v_version ge v2.0.0)
1138 ? "10FFFF"
1139 : "FFFF";
6189eadc
KW
1140my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1141my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d 1142
026d3e0c
KW
1143# We work with above-Unicode code points, up to IV_MAX, but we may want to use
1144# sentinels above that number. Therefore for internal use, we use a much
1145# smaller number, translating it to IV_MAX only for output. The exact number
1146# is immaterial (all Unicode code points are treated exactly the same), but
1147# the algorithm requires it to be at least 2 * $MAX_UNICODE_CODEPOINTS + 1;
2d88a86a
KW
1148my $MAX_WORKING_CODEPOINTS= $MAX_UNICODE_CODEPOINT * 8;
1149my $MAX_WORKING_CODEPOINT = $MAX_WORKING_CODEPOINTS - 1;
1150my $MAX_WORKING_CODEPOINT_STRING = sprintf("%X", $MAX_WORKING_CODEPOINT);
1151
026d3e0c 1152my $MAX_PLATFORM_CODEPOINT = ~0 >> 1;
2d88a86a 1153
99870f4d
KW
1154# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1155# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1156# to decrease backtracking. The first regex allows the code point to be at
1157# the end of a word, but to work properly, the word shouldn't end with a valid
1158# hex character. The second one won't match a code point at the end of a
1159# word, and doesn't have the run-on issue
8c32d378
KW
1160my $run_on_code_point_re =
1161 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1162my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d 1163
44b0f8c1 1164# This matches the beginning of the line in the Unicode DB files that give the
99870f4d
KW
1165# defaults for code points not listed (i.e., missing) in the file. The code
1166# depends on this ending with a semi-colon, so it can assume it is a valid
1167# field when the line is split() by semi-colons
346a20cf 1168my $missing_defaults_prefix = qr/^#\s+\@missing:\s+0000\.\.10FFFF\s*;/;
99870f4d
KW
1169
1170# Property types. Unicode has more types, but these are sufficient for our
1171# purposes.
1172my $UNKNOWN = -1; # initialized to illegal value
1173my $NON_STRING = 1; # Either binary or enum
1174my $BINARY = 2;
06f26c45
KW
1175my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1176 # tables, additional true and false tables are
1177 # generated so that false is anything matching the
1178 # default value, and true is everything else.
1179my $ENUM = 4; # Include catalog
1180my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1181
1182# Some input files have lines that give default values for code points not
1183# contained in the file. Sometimes these should be ignored.
1184my $NO_DEFAULTS = 0; # Must evaluate to false
1185my $NOT_IGNORED = 1;
1186my $IGNORED = 2;
1187
1188# Range types. Each range has a type. Most ranges are type 0, for normal,
1189# and will appear in the main body of the tables in the output files, but
1190# there are other types of ranges as well, listed below, that are specially
1191# handled. There are pseudo-types as well that will never be stored as a
1192# type, but will affect the calculation of the type.
1193
1194# 0 is for normal, non-specials
1195my $MULTI_CP = 1; # Sequence of more than code point
1196my $HANGUL_SYLLABLE = 2;
1197my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1198my $NULL = 4; # The map is to the null string; utf8.c can't
1199 # handle these, nor is there an accepted syntax
1200 # for them in \p{} constructs
f86864ac 1201my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1202 # otherwise be $MULTI_CP type are instead type 0
1203
1204# process_generic_property_file() can accept certain overrides in its input.
1205# Each of these must begin AND end with $CMD_DELIM.
1206my $CMD_DELIM = "\a";
1207my $REPLACE_CMD = 'replace'; # Override the Replace
1208my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1209
1210my $NO = 0;
1211my $YES = 1;
1212
1213# Values for the Replace argument to add_range.
1214# $NO # Don't replace; add only the code points not
1215 # already present.
1216my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1217 # the comments at the subroutine definition.
1218my $UNCONDITIONALLY = 2; # Replace without conditions.
9470941f 1219my $MULTIPLE_BEFORE = 4; # Don't replace, but add a duplicate record if
99870f4d 1220 # already there
7f4b1e25
KW
1221my $MULTIPLE_AFTER = 5; # Don't replace, but add a duplicate record if
1222 # already there
1223my $CROAK = 6; # Die with an error if is already there
99870f4d
KW
1224
1225# Flags to give property statuses. The phrases are to remind maintainers that
1226# if the flag is changed, the indefinite article referring to it in the
1227# documentation may need to be as well.
1228my $NORMAL = "";
99870f4d
KW
1229my $DEPRECATED = 'D';
1230my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1231my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1232my $DISCOURAGED = 'X';
1233my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1234my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1235my $STRICTER = 'T';
1236my $a_bold_stricter = "a 'B<$STRICTER>'";
1237my $A_bold_stricter = "A 'B<$STRICTER>'";
1238my $STABILIZED = 'S';
1239my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1240my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1241my $OBSOLETE = 'O';
1242my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1243my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1244
5c6d9947
KW
1245# Aliases can also have an extra status:
1246my $INTERNAL_ALIAS = 'P';
1247
99870f4d
KW
1248my %status_past_participles = (
1249 $DISCOURAGED => 'discouraged',
99870f4d
KW
1250 $STABILIZED => 'stabilized',
1251 $OBSOLETE => 'obsolete',
37e2e78e 1252 $DEPRECATED => 'deprecated',
5c6d9947 1253 $INTERNAL_ALIAS => 'reserved for Perl core internal use only',
99870f4d
KW
1254);
1255
395dfc19
KW
1256# Table fates. These are somewhat ordered, so that fates < $MAP_PROXIED should be
1257# externally documented.
301ba948 1258my $ORDINARY = 0; # The normal fate.
395dfc19
KW
1259my $MAP_PROXIED = 1; # The map table for the property isn't written out,
1260 # but there is a file written that can be used to
1261 # reconstruct this table
3cdaf629 1262my $INTERNAL_ONLY = 2; # The file for this table is written out, but it is
301ba948 1263 # for Perl's internal use only
277b7b16
KW
1264my $LEGACY_ONLY = 3; # Like $INTERNAL_ONLY, but not actually used by Perl.
1265 # Is for backwards compatibility for applications that
1266 # read the file directly, so it's format is
1267 # unchangeable.
1268my $SUPPRESSED = 4; # The file for this table is not written out, and as a
3cdaf629
KW
1269 # result, we don't bother to do many computations on
1270 # it.
277b7b16 1271my $PLACEHOLDER = 5; # Like $SUPPRESSED, but we go through all the
3cdaf629
KW
1272 # computations anyway, as the values are needed for
1273 # things to work. This happens when we have Perl
1274 # extensions that depend on Unicode tables that
1275 # wouldn't normally be in a given Unicode version.
301ba948 1276
f5817e0a
KW
1277# The format of the values of the tables:
1278my $EMPTY_FORMAT = "";
99870f4d
KW
1279my $BINARY_FORMAT = 'b';
1280my $DECIMAL_FORMAT = 'd';
1281my $FLOAT_FORMAT = 'f';
1282my $INTEGER_FORMAT = 'i';
1283my $HEX_FORMAT = 'x';
1284my $RATIONAL_FORMAT = 'r';
1285my $STRING_FORMAT = 's';
d11155ec 1286my $ADJUST_FORMAT = 'a';
24303724 1287my $HEX_ADJUST_FORMAT = 'ax';
a14f3cb1 1288my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1289my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1290
1291my %map_table_formats = (
1292 $BINARY_FORMAT => 'binary',
1293 $DECIMAL_FORMAT => 'single decimal digit',
1294 $FLOAT_FORMAT => 'floating point number',
1295 $INTEGER_FORMAT => 'integer',
add63c13 1296 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1297 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1298 $STRING_FORMAT => 'string',
d11155ec 1299 $ADJUST_FORMAT => 'some entries need adjustment',
24303724 1300 $HEX_ADJUST_FORMAT => 'mapped value in hex; some entries need adjustment',
92f9d56c 1301 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1302 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1303);
1304
1305# Unicode didn't put such derived files in a separate directory at first.
1306my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1307my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1308my $AUXILIARY = 'auxiliary';
1309
3854b4b8
KW
1310# Hashes and arrays that will eventually go into Heavy.pl for the use of
1311# utf8_heavy.pl and into UCD.pl for the use of UCD.pm
99870f4d
KW
1312my %loose_to_file_of; # loosely maps table names to their respective
1313 # files
1314my %stricter_to_file_of; # same; but for stricter mapping.
315bfd4e 1315my %loose_property_to_file_of; # Maps a loose property name to its map file
e1dc048d 1316my %strict_property_to_file_of; # Same, but strict
3854b4b8
KW
1317my @inline_definitions = "V0"; # Each element gives a definition of a unique
1318 # inversion list. When a definition is inlined,
1319 # its value in the hash it's in (one of the two
1320 # defined just above) will include an index into
1321 # this array. The 0th element is initialized to
9be22ca5 1322 # the definition for a zero length inversion list
89cf10cc
KW
1323my %file_to_swash_name; # Maps the file name to its corresponding key name
1324 # in the hash %utf8::SwashInfo
99870f4d
KW
1325my %nv_floating_to_rational; # maps numeric values floating point numbers to
1326 # their rational equivalent
c12f2655
KW
1327my %loose_property_name_of; # Loosely maps (non_string) property names to
1328 # standard form
e1dc048d
KW
1329my %strict_property_name_of; # Strictly maps (non_string) property names to
1330 # standard form
86a52d1e 1331my %string_property_loose_to_name; # Same, for string properties.
c15fda25
KW
1332my %loose_defaults; # keys are of form "prop=value", where 'prop' is
1333 # the property name in standard loose form, and
1334 # 'value' is the default value for that property,
1335 # also in standard loose form.
9e4a1e86
KW
1336my %loose_to_standard_value; # loosely maps table names to the canonical
1337 # alias for them
2df7880f
KW
1338my %ambiguous_names; # keys are alias names (in standard form) that
1339 # have more than one possible meaning.
d963b40d
KW
1340my %combination_property; # keys are alias names (in standard form) that
1341 # have both a map table, and a binary one that
1342 # yields true for all non-null maps.
5d1df013
KW
1343my %prop_aliases; # Keys are standard property name; values are each
1344 # one's aliases
1e863613
KW
1345my %prop_value_aliases; # Keys of top level are standard property name;
1346 # values are keys to another hash, Each one is
1347 # one of the property's values, in standard form.
1348 # The values are that prop-val's aliases.
b9e57660 1349my %skipped_files; # List of files that we skip
2df7880f 1350my %ucd_pod; # Holds entries that will go into the UCD section of the pod
99870f4d 1351
d867ccfb
KW
1352# Most properties are immune to caseless matching, otherwise you would get
1353# nonsensical results, as properties are a function of a code point, not
1354# everything that is caselessly equivalent to that code point. For example,
1355# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1356# be true because 's' and 'S' are equivalent caselessly. However,
1357# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1358# extend that concept to those very few properties that are like this. Each
1359# such property will match the full range caselessly. They are hard-coded in
1360# the program; it's not worth trying to make it general as it's extremely
1361# unlikely that they will ever change.
1362my %caseless_equivalent_to;
1363
53adf6a2
KW
1364# This is the range of characters that were in Release 1 of Unicode, and
1365# removed in Release 2 (replaced with the current Hangul syllables starting at
1366# U+AC00). The range was reused starting in Release 3 for other purposes.
1367my $FIRST_REMOVED_HANGUL_SYLLABLE = 0x3400;
1368my $FINAL_REMOVED_HANGUL_SYLLABLE = 0x4DFF;
1369
99870f4d
KW
1370# These constants names and values were taken from the Unicode standard,
1371# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1372# syllables. The '_string' versions are so generated tables can retain the
1373# hex format, which is the more familiar value
1374my $SBase_string = "0xAC00";
1375my $SBase = CORE::hex $SBase_string;
1376my $LBase_string = "0x1100";
1377my $LBase = CORE::hex $LBase_string;
1378my $VBase_string = "0x1161";
1379my $VBase = CORE::hex $VBase_string;
1380my $TBase_string = "0x11A7";
1381my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1382my $SCount = 11172;
1383my $LCount = 19;
1384my $VCount = 21;
1385my $TCount = 28;
1386my $NCount = $VCount * $TCount;
1387
1388# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1389# with the above published constants.
1390my %Jamo;
1391my %Jamo_L; # Leading consonants
1392my %Jamo_V; # Vowels
1393my %Jamo_T; # Trailing consonants
1394
bb1dd3da
KW
1395# For code points whose name contains its ordinal as a '-ABCD' suffix.
1396# The key is the base name of the code point, and the value is an
1397# array giving all the ranges that use this base name. Each range
1398# is actually a hash giving the 'low' and 'high' values of it.
1399my %names_ending_in_code_point;
1400my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1401 # removed from the names
1402# Inverse mapping. The list of ranges that have these kinds of
1403# names. Each element contains the low, high, and base names in an
1404# anonymous hash.
1405my @code_points_ending_in_code_point;
1406
6b5ab373
KW
1407# To hold Unicode's normalization test suite
1408my @normalization_tests;
1409
bb1dd3da
KW
1410# Boolean: does this Unicode version have the hangul syllables, and are we
1411# writing out a table for them?
1412my $has_hangul_syllables = 0;
1413
1414# Does this Unicode version have code points whose names end in their
1415# respective code points, and are we writing out a table for them? 0 for no;
1416# otherwise points to first property that a table is needed for them, so that
1417# if multiple tables are needed, we don't create duplicates
1418my $needing_code_points_ending_in_code_point = 0;
1419
37e2e78e 1420my @backslash_X_tests; # List of tests read in for testing \X
6b659339 1421my @LB_tests; # List of tests read in for testing \b{lb}
06ae2722 1422my @SB_tests; # List of tests read in for testing \b{sb}
ae3bb8ea 1423my @WB_tests; # List of tests read in for testing \b{wb}
99870f4d
KW
1424my @unhandled_properties; # Will contain a list of properties found in
1425 # the input that we didn't process.
f86864ac 1426my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1427 # listed in the pod
1428my @map_properties; # Properties that get map files written
1429my @named_sequences; # NamedSequences.txt contents.
1430my %potential_files; # Generated list of all .txt files in the directory
1431 # structure so we can warn if something is being
1432 # ignored.
a0204bb1
KW
1433my @missing_early_files; # Generated list of absent files that we need to
1434 # proceed in compiling this early Unicode version
99870f4d
KW
1435my @files_actually_output; # List of files we generated.
1436my @more_Names; # Some code point names are compound; this is used
1437 # to store the extra components of them.
1438my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1439 # the minimum before we consider it equivalent to a
1440 # candidate rational
1441my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1442
1443# These store references to certain commonly used property objects
1254636b 1444my $age;
3c88a801 1445my $ccc;
99870f4d
KW
1446my $gc;
1447my $perl;
1448my $block;
3e20195b
KW
1449my $perl_charname;
1450my $print;
ac7dbdb6 1451my $All;
bc0c431d 1452my $Assigned; # All assigned characters in this Unicode release
e47e66b9
KW
1453my $DI; # Default_Ignorable_Code_Point property
1454my $NChar; # Noncharacter_Code_Point property
359523e2 1455my $script;
99870f4d
KW
1456
1457# Are there conflicting names because of beginning with 'In_', or 'Is_'
1458my $has_In_conflicts = 0;
1459my $has_Is_conflicts = 0;
1460
1461sub internal_file_to_platform ($) {
1462 # Convert our file paths which have '/' separators to those of the
1463 # platform.
1464
1465 my $file = shift;
1466 return undef unless defined $file;
1467
1468 return File::Spec->join(split '/', $file);
d07a55ed 1469}
5beb625e 1470
99870f4d
KW
1471sub file_exists ($) { # platform independent '-e'. This program internally
1472 # uses slash as a path separator.
1473 my $file = shift;
1474 return 0 if ! defined $file;
1475 return -e internal_file_to_platform($file);
1476}
5beb625e 1477
99870f4d 1478sub objaddr($) {
23e33b60
KW
1479 # Returns the address of the blessed input object.
1480 # It doesn't check for blessedness because that would do a string eval
1481 # every call, and the program is structured so that this is never called
1482 # for a non-blessed object.
99870f4d 1483
23e33b60 1484 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1485
1486 # Numifying a ref gives its address.
051df77b 1487 return pack 'J', $_[0];
99870f4d
KW
1488}
1489
558712cf 1490# These are used only if $annotate is true.
c4019d52
KW
1491# The entire range of Unicode characters is examined to populate these
1492# after all the input has been processed. But most can be skipped, as they
1493# have the same descriptive phrases, such as being unassigned
1494my @viacode; # Contains the 1 million character names
4b06b8de 1495my @age; # And their ages ("" if none)
c4019d52
KW
1496my @printable; # boolean: And are those characters printable?
1497my @annotate_char_type; # Contains a type of those characters, specifically
1498 # for the purposes of annotation.
1499my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1500 # name for the purposes of annotation. They map to the
c4019d52
KW
1501 # upper edge of the range, so that the end point can
1502 # be immediately found. This is used to skip ahead to
1503 # the end of a range, and avoid processing each
1504 # individual code point in it.
1505my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1506 # characters, but excluding those which are
1507 # also noncharacter code points
1508
1509# The annotation types are an extension of the regular range types, though
1510# some of the latter are folded into one. Make the new types negative to
1511# avoid conflicting with the regular types
1512my $SURROGATE_TYPE = -1;
1513my $UNASSIGNED_TYPE = -2;
1514my $PRIVATE_USE_TYPE = -3;
1515my $NONCHARACTER_TYPE = -4;
1516my $CONTROL_TYPE = -5;
2d88a86a
KW
1517my $ABOVE_UNICODE_TYPE = -6;
1518my $UNKNOWN_TYPE = -7; # Used only if there is a bug in this program
c4019d52
KW
1519
1520sub populate_char_info ($) {
558712cf 1521 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1522 # input code point's info that are needed for outputting more detailed
1523 # comments. If calling context wants a return, it is the end point of
1524 # any contiguous range of characters that share essentially the same info
1525
1526 my $i = shift;
1527 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1528
1529 $viacode[$i] = $perl_charname->value_of($i) || "";
1254636b 1530 $age[$i] = (defined $age)
272af058 1531 ? (($age->value_of($i) =~ / ^ \d+ \. \d+ $ /x)
1254636b
KW
1532 ? $age->value_of($i)
1533 : "")
1534 : "";
c4019d52
KW
1535
1536 # A character is generally printable if Unicode says it is,
1537 # but below we make sure that most Unicode general category 'C' types
1538 # aren't.
1539 $printable[$i] = $print->contains($i);
1540
1254636b
KW
1541 # But the characters in this range were removed in v2.0 and replaced by
1542 # different ones later. Modern fonts will be for the replacement
1543 # characters, so suppress printing them.
1544 if (($v_version lt v2.0
1545 || ($compare_versions && $compare_versions lt v2.0))
1546 && ( $i >= $FIRST_REMOVED_HANGUL_SYLLABLE
1547 && $i <= $FINAL_REMOVED_HANGUL_SYLLABLE))
1548 {
1549 $printable[$i] = 0;
1550 }
1551
c4019d52
KW
1552 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1553
1554 # Only these two regular types are treated specially for annotations
1555 # purposes
1556 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1557 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1558
1559 # Give a generic name to all code points that don't have a real name.
1560 # We output ranges, if applicable, for these. Also calculate the end
1561 # point of the range.
1562 my $end;
1563 if (! $viacode[$i]) {
2d88a86a
KW
1564 if ($i > $MAX_UNICODE_CODEPOINT) {
1565 $viacode[$i] = 'Above-Unicode';
1566 $annotate_char_type[$i] = $ABOVE_UNICODE_TYPE;
1567 $printable[$i] = 0;
1568 $end = $MAX_WORKING_CODEPOINT;
1569 }
1570 elsif ($gc-> table('Private_use')->contains($i)) {
c4019d52
KW
1571 $viacode[$i] = 'Private Use';
1572 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1573 $printable[$i] = 0;
1574 $end = $gc->table('Private_Use')->containing_range($i)->end;
1575 }
e47e66b9 1576 elsif ($NChar->contains($i)) {
c4019d52
KW
1577 $viacode[$i] = 'Noncharacter';
1578 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1579 $printable[$i] = 0;
e47e66b9 1580 $end = $NChar->containing_range($i)->end;
c4019d52
KW
1581 }
1582 elsif ($gc-> table('Control')->contains($i)) {
1254636b
KW
1583 my $name_ref = property_ref('Name_Alias');
1584 $name_ref = property_ref('Unicode_1_Name') if ! defined $name_ref;
1585 $viacode[$i] = (defined $name_ref)
1586 ? $name_ref->value_of($i)
1587 : 'Control';
c4019d52
KW
1588 $annotate_char_type[$i] = $CONTROL_TYPE;
1589 $printable[$i] = 0;
c4019d52
KW
1590 }
1591 elsif ($gc-> table('Unassigned')->contains($i)) {
c4019d52
KW
1592 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1593 $printable[$i] = 0;
1254636b
KW
1594 $viacode[$i] = 'Unassigned';
1595
349edad4 1596 if (defined $block) { # No blocks in earliest releases
1254636b 1597 $viacode[$i] .= ', block=' . $block-> value_of($i);
1d025d66 1598 $end = $gc-> table('Unassigned')->containing_range($i)->end;
c4019d52 1599
bf06c733
KW
1600 # Because we name the unassigned by the blocks they are in, it
1601 # can't go past the end of that block, and it also can't go
1602 # past the unassigned range it is in. The special table makes
1603 # sure that the non-characters, which are unassigned, are
1604 # separated out.
1605 $end = min($block->containing_range($i)->end,
1606 $unassigned_sans_noncharacters->
1607 containing_range($i)->end);
1d025d66 1608 }
1254636b
KW
1609 else {
1610 $end = $i + 1;
1611 while ($unassigned_sans_noncharacters->contains($end)) {
1612 $end++;
1613 }
1614 $end--;
1615 }
1d025d66 1616 }
35e2b416 1617 elsif ($perl->table('_Perl_Surrogate')->contains($i)) {
1d025d66
KW
1618 $viacode[$i] = 'Surrogate';
1619 $annotate_char_type[$i] = $SURROGATE_TYPE;
1620 $printable[$i] = 0;
1621 $end = $gc->table('Surrogate')->containing_range($i)->end;
13ca76ff
KW
1622 }
1623 else {
1624 Carp::my_carp_bug("Can't figure out how to annotate "
1625 . sprintf("U+%04X", $i)
1626 . ". Proceeding anyway.");
c4019d52
KW
1627 $viacode[$i] = 'UNKNOWN';
1628 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1629 $printable[$i] = 0;
1630 }
1631 }
1632
1633 # Here, has a name, but if it's one in which the code point number is
1634 # appended to the name, do that.
1635 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1636 $viacode[$i] .= sprintf("-%04X", $i);
4b06b8de 1637
4b06b8de 1638 my $limit = $perl_charname->containing_range($i)->end;
1254636b
KW
1639 if (defined $age) {
1640 # Do all these as groups of the same age, instead of individually,
1641 # because their names are so meaningless, and there are typically
1642 # large quantities of them.
1643 $end = $i + 1;
1644 while ($end <= $limit && $age->value_of($end) == $age[$i]) {
1645 $end++;
1646 }
1647 $end--;
1648 }
1649 else {
1650 $end = $limit;
4b06b8de 1651 }
c4019d52
KW
1652 }
1653
1654 # And here, has a name, but if it's a hangul syllable one, replace it with
1655 # the correct name from the Unicode algorithm
1656 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1657 use integer;
1658 my $SIndex = $i - $SBase;
1659 my $L = $LBase + $SIndex / $NCount;
1660 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1661 my $T = $TBase + $SIndex % $TCount;
1662 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1663 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1664 $end = $perl_charname->containing_range($i)->end;
1665 }
1666
1667 return if ! defined wantarray;
1668 return $i if ! defined $end; # If not a range, return the input
1669
1670 # Save this whole range so can find the end point quickly
1671 $annotate_ranges->add_map($i, $end, $end);
1672
1673 return $end;
1674}
1675
23e33b60
KW
1676# Commented code below should work on Perl 5.8.
1677## This 'require' doesn't necessarily work in miniperl, and even if it does,
1678## the native perl version of it (which is what would operate under miniperl)
1679## is extremely slow, as it does a string eval every call.
7e017d6d 1680#my $has_fast_scalar_util = $^X !~ /miniperl/
23e33b60
KW
1681# && defined eval "require Scalar::Util";
1682#
1683#sub objaddr($) {
1684# # Returns the address of the blessed input object. Uses the XS version if
1685# # available. It doesn't check for blessedness because that would do a
1686# # string eval every call, and the program is structured so that this is
1687# # never called for a non-blessed object.
1688#
1689# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1690#
1691# # Check at least that is a ref.
1692# my $pkg = ref($_[0]) or return undef;
1693#
1694# # Change to a fake package to defeat any overloaded stringify
1695# bless $_[0], 'main::Fake';
1696#
1697# # Numifying a ref gives its address.
051df77b 1698# my $addr = pack 'J', $_[0];
23e33b60
KW
1699#
1700# # Return to original class
1701# bless $_[0], $pkg;
1702# return $addr;
1703#}
1704
99870f4d
KW
1705sub max ($$) {
1706 my $a = shift;
1707 my $b = shift;
1708 return $a if $a >= $b;
1709 return $b;
1710}
1711
1712sub min ($$) {
1713 my $a = shift;
1714 my $b = shift;
1715 return $a if $a <= $b;
1716 return $b;
1717}
1718
1719sub clarify_number ($) {
1720 # This returns the input number with underscores inserted every 3 digits
1721 # in large (5 digits or more) numbers. Input must be entirely digits, not
1722 # checked.
1723
1724 my $number = shift;
1725 my $pos = length($number) - 3;
1726 return $number if $pos <= 1;
1727 while ($pos > 0) {
1728 substr($number, $pos, 0) = '_';
1729 $pos -= 3;
5beb625e 1730 }
99870f4d 1731 return $number;
99598c8c
JH
1732}
1733
731cb813
KW
1734sub clarify_code_point_count ($) {
1735 # This is like clarify_number(), but the input is assumed to be a count of
1736 # code points, rather than a generic number.
1737
2d88a86a
KW
1738 my $append = "";
1739
1740 my $number = shift;
1741 if ($number > $MAX_UNICODE_CODEPOINTS) {
1742 $number -= ($MAX_WORKING_CODEPOINTS - $MAX_UNICODE_CODEPOINTS);
1743 return "All above-Unicode code points" if $number == 0;
1744 $append = " + all above-Unicode code points";
1745 }
1746 return clarify_number($number) . $append;
731cb813 1747}
12ac2576 1748
99870f4d 1749package Carp;
7ebf06b3 1750
99870f4d
KW
1751# These routines give a uniform treatment of messages in this program. They
1752# are placed in the Carp package to cause the stack trace to not include them,
1753# although an alternative would be to use another package and set @CARP_NOT
1754# for it.
12ac2576 1755
99870f4d 1756our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1757
99f78760
KW
1758# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1759# and overload trying to load Scalar:Util under miniperl. See
1760# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1761undef $overload::VERSION;
1762
99870f4d
KW
1763sub my_carp {
1764 my $message = shift || "";
1765 my $nofold = shift || 0;
7ebf06b3 1766
99870f4d
KW
1767 if ($message) {
1768 $message = main::join_lines($message);
1769 $message =~ s/^$0: *//; # Remove initial program name
1770 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1771 $message = "\n$0: $message;";
12ac2576 1772
99870f4d
KW
1773 # Fold the message with program name, semi-colon end punctuation
1774 # (which looks good with the message that carp appends to it), and a
1775 # hanging indent for continuation lines.
1776 $message = main::simple_fold($message, "", 4) unless $nofold;
1777 $message =~ s/\n$//; # Remove the trailing nl so what carp
1778 # appends is to the same line
1779 }
12ac2576 1780
99870f4d 1781 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1782
99870f4d
KW
1783 carp $message;
1784 return;
1785}
7ebf06b3 1786
99870f4d
KW
1787sub my_carp_bug {
1788 # This is called when it is clear that the problem is caused by a bug in
1789 # this program.
7ebf06b3 1790
99870f4d
KW
1791 my $message = shift;
1792 $message =~ s/^$0: *//;
1793 $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");
1794 carp $message;
1795 return;
1796}
7ebf06b3 1797
99870f4d
KW
1798sub carp_too_few_args {
1799 if (@_ != 2) {
1800 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1801 return;
12ac2576 1802 }
7ebf06b3 1803
99870f4d
KW
1804 my $args_ref = shift;
1805 my $count = shift;
7ebf06b3 1806
99870f4d
KW
1807 my_carp_bug("Need at least $count arguments to "
1808 . (caller 1)[3]
1809 . ". Instead got: '"
1810 . join ', ', @$args_ref
1811 . "'. No action taken.");
1812 return;
12ac2576
JP
1813}
1814
99870f4d
KW
1815sub carp_extra_args {
1816 my $args_ref = shift;
1817 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1818
99870f4d
KW
1819 unless (ref $args_ref) {
1820 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1821 return;
1822 }
1823 my ($package, $file, $line) = caller;
1824 my $subroutine = (caller 1)[3];
cf25bb62 1825
99870f4d
KW
1826 my $list;
1827 if (ref $args_ref eq 'HASH') {
1828 foreach my $key (keys %$args_ref) {
1829 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1830 }
99870f4d 1831 $list = join ', ', each %{$args_ref};
cf25bb62 1832 }
99870f4d
KW
1833 elsif (ref $args_ref eq 'ARRAY') {
1834 foreach my $arg (@$args_ref) {
1835 $arg = $UNDEF unless defined $arg;
1836 }
1837 $list = join ', ', @$args_ref;
1838 }
1839 else {
1840 my_carp_bug("Can't cope with ref "
1841 . ref($args_ref)
1842 . " . argument to 'carp_extra_args'. Not checking arguments.");
1843 return;
1844 }
1845
1846 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1847 return;
d73e5302
JH
1848}
1849
99870f4d
KW
1850package main;
1851
1852{ # Closure
1853
1854 # This program uses the inside-out method for objects, as recommended in
1b6b3fa9
KW
1855 # "Perl Best Practices". (This is the best solution still, since this has
1856 # to run under miniperl.) This closure aids in generating those. There
99870f4d
KW
1857 # are two routines. setup_package() is called once per package to set
1858 # things up, and then set_access() is called for each hash representing a
1859 # field in the object. These routines arrange for the object to be
1860 # properly destroyed when no longer used, and for standard accessor
1861 # functions to be generated. If you need more complex accessors, just
1862 # write your own and leave those accesses out of the call to set_access().
1863 # More details below.
1864
1865 my %constructor_fields; # fields that are to be used in constructors; see
1866 # below
1867
1868 # The values of this hash will be the package names as keys to other
1869 # hashes containing the name of each field in the package as keys, and
1870 # references to their respective hashes as values.
1871 my %package_fields;
1872
1873 sub setup_package {
1874 # Sets up the package, creating standard DESTROY and dump methods
1875 # (unless already defined). The dump method is used in debugging by
1876 # simple_dumper().
1877 # The optional parameters are:
1878 # a) a reference to a hash, that gets populated by later
1879 # set_access() calls with one of the accesses being
1880 # 'constructor'. The caller can then refer to this, but it is
1881 # not otherwise used by these two routines.
1882 # b) a reference to a callback routine to call during destruction
1883 # of the object, before any fields are actually destroyed
1884
1885 my %args = @_;
1886 my $constructor_ref = delete $args{'Constructor_Fields'};
1887 my $destroy_callback = delete $args{'Destroy_Callback'};
1888 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1889
1890 my %fields;
1891 my $package = (caller)[0];
1892
1893 $package_fields{$package} = \%fields;
1894 $constructor_fields{$package} = $constructor_ref;
1895
1896 unless ($package->can('DESTROY')) {
1897 my $destroy_name = "${package}::DESTROY";
1898 no strict "refs";
1899
1900 # Use typeglob to give the anonymous subroutine the name we want
1901 *$destroy_name = sub {
1902 my $self = shift;
ffe43484 1903 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1904
1905 $self->$destroy_callback if $destroy_callback;
1906 foreach my $field (keys %{$package_fields{$package}}) {
1907 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1908 delete $package_fields{$package}{$field}{$addr};
1909 }
1910 return;
1911 }
1912 }
1913
1914 unless ($package->can('dump')) {
1915 my $dump_name = "${package}::dump";
1916 no strict "refs";
1917 *$dump_name = sub {
1918 my $self = shift;
1919 return dump_inside_out($self, $package_fields{$package}, @_);
1920 }
1921 }
1922 return;
1923 }
1924
1925 sub set_access {
1926 # Arrange for the input field to be garbage collected when no longer
1927 # needed. Also, creates standard accessor functions for the field
1928 # based on the optional parameters-- none if none of these parameters:
1929 # 'addable' creates an 'add_NAME()' accessor function.
1930 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1931 # function.
1932 # 'settable' creates a 'set_NAME()' accessor function.
1933 # 'constructor' doesn't create an accessor function, but adds the
1934 # field to the hash that was previously passed to
1935 # setup_package();
1936 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1937 # 'add' etc. all mean 'addable'.
1938 # The read accessor function will work on both array and scalar
1939 # values. If another accessor in the parameter list is 'a', the read
1940 # access assumes an array. You can also force it to be array access
1941 # by specifying 'readable_array' instead of 'readable'
1942 #
1943 # A sort-of 'protected' access can be set-up by preceding the addable,
1944 # readable or settable with some initial portion of 'protected_' (but,
1945 # the underscore is required), like 'p_a', 'pro_set', etc. The
1946 # "protection" is only by convention. All that happens is that the
1947 # accessor functions' names begin with an underscore. So instead of
1948 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1949 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1950 # end of each package, and then storing the __LINE__ ranges and
1951 # checking them on every accessor. But that is way overkill.)
1952
1953 # We create anonymous subroutines as the accessors and then use
1954 # typeglobs to assign them to the proper package and name
1955
1956 my $name = shift; # Name of the field
1957 my $field = shift; # Reference to the inside-out hash containing the
1958 # field
1959
1960 my $package = (caller)[0];
1961
1962 if (! exists $package_fields{$package}) {
1963 croak "$0: Must call 'setup_package' before 'set_access'";
1964 }
d73e5302 1965
99870f4d
KW
1966 # Stash the field so DESTROY can get it.
1967 $package_fields{$package}{$name} = $field;
cf25bb62 1968
99870f4d
KW
1969 # Remaining arguments are the accessors. For each...
1970 foreach my $access (@_) {
1971 my $access = lc $access;
cf25bb62 1972
99870f4d 1973 my $protected = "";
cf25bb62 1974
99870f4d
KW
1975 # Match the input as far as it goes.
1976 if ($access =~ /^(p[^_]*)_/) {
1977 $protected = $1;
1978 if (substr('protected_', 0, length $protected)
1979 eq $protected)
1980 {
1981
1982 # Add 1 for the underscore not included in $protected
1983 $access = substr($access, length($protected) + 1);
1984 $protected = '_';
1985 }
1986 else {
1987 $protected = "";
1988 }
1989 }
1990
1991 if (substr('addable', 0, length $access) eq $access) {
1992 my $subname = "${package}::${protected}add_$name";
1993 no strict "refs";
1994
1995 # add_ accessor. Don't add if already there, which we
1996 # determine using 'eq' for scalars and '==' otherwise.
1997 *$subname = sub {
1998 use strict "refs";
1999 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
2000 my $self = shift;
2001 my $value = shift;
ffe43484 2002 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2003 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2004 if (ref $value) {
f998e60c 2005 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
2006 }
2007 else {
f998e60c 2008 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 2009 }
f998e60c 2010 push @{$field->{$addr}}, $value;
99870f4d
KW
2011 return;
2012 }
2013 }
2014 elsif (substr('constructor', 0, length $access) eq $access) {
2015 if ($protected) {
2016 Carp::my_carp_bug("Can't set-up 'protected' constructors")
2017 }
2018 else {
2019 $constructor_fields{$package}{$name} = $field;
2020 }
2021 }
2022 elsif (substr('readable_array', 0, length $access) eq $access) {
2023
2024 # Here has read access. If one of the other parameters for
2025 # access is array, or this one specifies array (by being more
2026 # than just 'readable_'), then create a subroutine that
2027 # assumes the data is an array. Otherwise just a scalar
2028 my $subname = "${package}::${protected}$name";
2029 if (grep { /^a/i } @_
2030 or length($access) > length('readable_'))
2031 {
2032 no strict "refs";
2033 *$subname = sub {
2034 use strict "refs";
23e33b60 2035 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 2036 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
2037 if (ref $field->{$addr} ne 'ARRAY') {
2038 my $type = ref $field->{$addr};
2039 $type = 'scalar' unless $type;
2040 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
2041 return;
2042 }
2043 return scalar @{$field->{$addr}} unless wantarray;
2044
2045 # Make a copy; had problems with caller modifying the
2046 # original otherwise
2047 my @return = @{$field->{$addr}};
2048 return @return;
2049 }
2050 }
2051 else {
2052
2053 # Here not an array value, a simpler function.
2054 no strict "refs";
2055 *$subname = sub {
2056 use strict "refs";
23e33b60 2057 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 2058 no overloading;
051df77b 2059 return $field->{pack 'J', $_[0]};
99870f4d
KW
2060 }
2061 }
2062 }
2063 elsif (substr('settable', 0, length $access) eq $access) {
2064 my $subname = "${package}::${protected}set_$name";
2065 no strict "refs";
2066 *$subname = sub {
2067 use strict "refs";
23e33b60
KW
2068 if (main::DEBUG) {
2069 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
2070 Carp::carp_extra_args(\@_) if @_ > 2;
2071 }
2072 # $self is $_[0]; $value is $_[1]
f998e60c 2073 no overloading;
051df77b 2074 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
2075 return;
2076 }
2077 }
2078 else {
2079 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
2080 }
cf25bb62 2081 }
99870f4d 2082 return;
cf25bb62 2083 }
99870f4d
KW
2084}
2085
2086package Input_file;
2087
2088# All input files use this object, which stores various attributes about them,
2089# and provides for convenient, uniform handling. The run method wraps the
2090# processing. It handles all the bookkeeping of opening, reading, and closing
2091# the file, returning only significant input lines.
2092#
2093# Each object gets a handler which processes the body of the file, and is
74cd47d0
KW
2094# called by run(). All character property files must use the generic,
2095# default handler, which has code scrubbed to handle things you might not
2096# expect, including automatic EBCDIC handling. For files that don't deal with
2097# mapping code points to a property value, such as test files,
2098# PropertyAliases, PropValueAliases, and named sequences, you can override the
2099# handler to be a custom one. Such a handler should basically be a
2100# while(next_line()) {...} loop.
99870f4d
KW
2101#
2102# You can also set up handlers to
40045c2d 2103# 0) call during object construction time, after everything else is done
537124e4 2104# 1) call before the first line is read, for pre processing
83b68635
KW
2105# 2) call to adjust each line of the input before the main handler gets
2106# them. This can be automatically generated, if appropriately simple
44b0f8c1 2107# enough, by specifying a Properties parameter in the constructor.
99870f4d 2108# 3) call upon EOF before the main handler exits its loop
537124e4 2109# 4) call at the end, for post processing
99870f4d
KW
2110#
2111# $_ is used to store the input line, and is to be filtered by the
2112# each_line_handler()s. So, if the format of the line is not in the desired
2113# format for the main handler, these are used to do that adjusting. They can
2114# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
44b0f8c1 2115# so the $_ output of one is used as the input to the next. The EOF handler
3848e5b7
KW
2116# is also stackable, but none of the others are, but could easily be changed
2117# to be so.
99870f4d 2118#
a0204bb1
KW
2119# Some properties are used by the Perl core but aren't defined until later
2120# Unicode releases. The perl interpreter would have problems working when
2121# compiled with an earlier Unicode version that doesn't have them, so we need
2122# to define them somehow for those releases. The 'Early' constructor
2123# parameter can be used to automatically handle this. It is essentially
2124# ignored if the Unicode version being compiled has a data file for this
2125# property. Either code to execute or a file to read can be specified.
2126# Details are at the %early definition.
2127#
99870f4d
KW
2128# Most of the handlers can call insert_lines() or insert_adjusted_lines()
2129# which insert the parameters as lines to be processed before the next input
3848e5b7 2130# file line is read. This allows the EOF handler(s) to flush buffers, for
99870f4d
KW
2131# example. The difference between the two routines is that the lines inserted
2132# by insert_lines() are subjected to the each_line_handler()s. (So if you
9be22ca5
KW
2133# called it from such a handler, you would get infinite recursion without some
2134# mechanism to prevent that.) Lines inserted by insert_adjusted_lines() go
2135# directly to the main handler without any adjustments. If the
2136# post-processing handler calls any of these, there will be no effect. Some
2137# error checking for these conditions could be added, but it hasn't been done.
99870f4d
KW
2138#
2139# carp_bad_line() should be called to warn of bad input lines, which clears $_
2140# to prevent further processing of the line. This routine will output the
2141# message as a warning once, and then keep a count of the lines that have the
2142# same message, and output that count at the end of the file's processing.
2143# This keeps the number of messages down to a manageable amount.
2144#
2145# get_missings() should be called to retrieve any @missing input lines.
2146# Messages will be raised if this isn't done if the options aren't to ignore
2147# missings.
2148
2149sub trace { return main::trace(@_); }
2150
99870f4d
KW
2151{ # Closure
2152 # Keep track of fields that are to be put into the constructor.
2153 my %constructor_fields;
2154
2155 main::setup_package(Constructor_Fields => \%constructor_fields);
2156
2157 my %file; # Input file name, required
2158 main::set_access('file', \%file, qw{ c r });
2159
2160 my %first_released; # Unicode version file was first released in, required
2161 main::set_access('first_released', \%first_released, qw{ c r });
2162
2163 my %handler; # Subroutine to process the input file, defaults to
2164 # 'process_generic_property_file'
2165 main::set_access('handler', \%handler, qw{ c });
2166
2167 my %property;
2168 # name of property this file is for. defaults to none, meaning not
2169 # applicable, or is otherwise determinable, for example, from each line.
696609bf 2170 main::set_access('property', \%property, qw{ c r });
99870f4d
KW
2171
2172 my %optional;
3274d896
KW
2173 # This is either an unsigned number, or a list of property names. In the
2174 # former case, if it is non-zero, it means the file is optional, so if the
2175 # file is absent, no warning about that is output. In the latter case, it
2176 # is a list of properties that the file (exclusively) defines. If the
2177 # file is present, tables for those properties will be produced; if
2178 # absent, none will, even if they are listed elsewhere (namely
2179 # PropertyAliases.txt and PropValueAliases.txt) as being in this release,
2180 # and no warnings will be raised about them not being available. (And no
2181 # warning about the file itself will be raised.)
2182 main::set_access('optional', \%optional, qw{ c readable_array } );
99870f4d
KW
2183
2184 my %non_skip;
2185 # This is used for debugging, to skip processing of all but a few input
2186 # files. Add 'non_skip => 1' to the constructor for those files you want
2187 # processed when you set the $debug_skip global.
2188 main::set_access('non_skip', \%non_skip, 'c');
2189
37e2e78e 2190 my %skip;
0dee21d6
KW
2191 # This is used to skip processing of this input file (semi-) permanently.
2192 # The value should be the reason the file is being skipped. It is used
2193 # for files that we aren't planning to process anytime soon, but want to
2194 # allow to be in the directory and be checked for their names not
2195 # conflicting with any other files on a DOS 8.3 name filesystem, but to
2196 # not otherwise be processed, and to not raise a warning about not being
2197 # handled. In the constructor call, any value that evaluates to a numeric
2198 # 0 or undef means don't skip. Any other value is a string giving the
44b0f8c1 2199 # reason it is being skipped, and this will appear in generated pod.
0dee21d6
KW
2200 # However, an empty string reason will suppress the pod entry.
2201 # Internally, calls that evaluate to numeric 0 are changed into undef to
2202 # distinguish them from an empty string call.
2203 main::set_access('skip', \%skip, 'c', 'r');
37e2e78e 2204
99870f4d
KW
2205 my %each_line_handler;
2206 # list of subroutines to look at and filter each non-comment line in the
2207 # file. defaults to none. The subroutines are called in order, each is
2208 # to adjust $_ for the next one, and the final one adjusts it for
2209 # 'handler'
2210 main::set_access('each_line_handler', \%each_line_handler, 'c');
2211
4e058f44
KW
2212 my %retain_trailing_comments;
2213 # This is used to not discard the comments that end data lines. This
2214 # would be used only for files with non-typical syntax, and most code here
2215 # assumes that comments have been stripped, so special handlers would have
2216 # to be written. It is assumed that the code will use these in
2217 # single-quoted contexts, and so any "'" marks in the comment will be
2218 # prefixed by a backslash.
2219 main::set_access('retain_trailing_comments', \%retain_trailing_comments, 'c');
2220
83b68635
KW
2221 my %properties; # Optional ordered list of the properties that occur in each
2222 # meaningful line of the input file. If present, an appropriate
2223 # each_line_handler() is automatically generated and pushed onto the stack
2224 # of such handlers. This is useful when a file contains multiple
44b0f8c1 2225 # properties per line, but no other special considerations are necessary.
83b68635
KW
2226 # The special value "<ignored>" means to discard the corresponding input
2227 # field.
2228 # Any @missing lines in the file should also match this syntax; no such
2229 # files exist as of 6.3. But if it happens in a future release, the code
2230 # could be expanded to properly parse them.
2231 main::set_access('properties', \%properties, qw{ c r });
2232
99870f4d
KW
2233 my %has_missings_defaults;
2234 # ? Are there lines in the file giving default values for code points
2235 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
2236 # the norm, but IGNORED means it has such lines, but the handler doesn't
2237 # use them. Having these three states allows us to catch changes to the
83b68635
KW
2238 # UCD that this program should track. XXX This could be expanded to
2239 # specify the syntax for such lines, like %properties above.
99870f4d
KW
2240 main::set_access('has_missings_defaults',
2241 \%has_missings_defaults, qw{ c r });
2242
40045c2d
KW
2243 my %construction_time_handler;
2244 # Subroutine to call at the end of the new method. If undef, no such
2245 # handler is called.
2246 main::set_access('construction_time_handler',
2247 \%construction_time_handler, qw{ c });
2248
99870f4d
KW
2249 my %pre_handler;
2250 # Subroutine to call before doing anything else in the file. If undef, no
2251 # such handler is called.
2252 main::set_access('pre_handler', \%pre_handler, qw{ c });
2253
2254 my %eof_handler;
3848e5b7 2255 # Subroutines to call upon getting an EOF on the input file, but before
99870f4d
KW
2256 # that is returned to the main handler. This is to allow buffers to be
2257 # flushed. The handler is expected to call insert_lines() or
2258 # insert_adjusted() with the buffered material
3848e5b7 2259 main::set_access('eof_handler', \%eof_handler, qw{ c });
99870f4d
KW
2260
2261 my %post_handler;
2262 # Subroutine to call after all the lines of the file are read in and
9be22ca5
KW
2263 # processed. If undef, no such handler is called. Note that this cannot
2264 # add lines to be processed; instead use eof_handler
99870f4d
KW
2265 main::set_access('post_handler', \%post_handler, qw{ c });
2266
2267 my %progress_message;
2268 # Message to print to display progress in lieu of the standard one
2269 main::set_access('progress_message', \%progress_message, qw{ c });
2270
2271 my %handle;
2272 # cache open file handle, internal. Is undef if file hasn't been
2273 # processed at all, empty if has;
2274 main::set_access('handle', \%handle);
2275
2276 my %added_lines;
2277 # cache of lines added virtually to the file, internal
2278 main::set_access('added_lines', \%added_lines);
2279
74cd47d0
KW
2280 my %remapped_lines;
2281 # cache of lines added virtually to the file, internal
2282 main::set_access('remapped_lines', \%remapped_lines);
2283
99870f4d
KW
2284 my %errors;
2285 # cache of errors found, internal
2286 main::set_access('errors', \%errors);
2287
2288 my %missings;
2289 # storage of '@missing' defaults lines
2290 main::set_access('missings', \%missings);
2291
a0204bb1
KW
2292 my %early;
2293 # Used for properties that must be defined (for Perl's purposes) on
2294 # versions of Unicode earlier than Unicode itself defines them. The
2295 # parameter is an array (it would be better to be a hash, but not worth
2296 # bothering about due to its rare use).
2297 #
2298 # The first element is either a code reference to call when in a release
2299 # earlier than the Unicode file is available in, or it is an alternate
2300 # file to use instead of the non-existent one. This file must have been
2301 # plunked down in the same directory as mktables. Should you be compiling
2302 # on a release that needs such a file, mktables will abort the
2303 # compilation, and tell you where to get the necessary file(s), and what
2304 # name(s) to use to store them as.
2305 # In the case of specifying an alternate file, the array must contain two
2306 # further elements:
2307 #
2308 # [1] is the name of the property that will be generated by this file.
2309 # The class automatically takes the input file and excludes any code
2310 # points in it that were not assigned in the Unicode version being
2311 # compiled. It then uses this result to define the property in the given
2312 # version. Since the property doesn't actually exist in the Unicode
2313 # version being compiled, this should be a name accessible only by core
2314 # perl. If it is the same name as the regular property, the constructor
2315 # will mark the output table as a $PLACEHOLDER so that it doesn't actually
2316 # get output, and so will be unusable by non-core code. Otherwise it gets
2317 # marked as $INTERNAL_ONLY.
2318 #
2319 # [2] is a property value to assign (only when compiling Unicode 1.1.5) to
2320 # the Hangul syllables in that release (which were ripped out in version
2321 # 2) for the given property . (Hence it is ignored except when compiling
2322 # version 1. You only get one value that applies to all of them, which
2323 # may not be the actual reality, but probably nobody cares anyway for
2324 # these obsolete characters.)
2325 #
3d36fe12
KW
2326 # [3] if present is the default value for the property to assign for code
2327 # points not given in the input. If not present, the default from the
2328 # normal property is used
2329 #
4ed2b786
KW
2330 # [-1] If there is an extra final element that is the string 'ONLY_EARLY'.
2331 # it means to not add the name in [1] as an alias to the property name
2332 # used for these. Normally, when compiling Unicode versions that don't
2333 # invoke the early handling, the name is added as a synonym.
cbdbe9d4 2334 #
a0204bb1
KW
2335 # Not all files can be handled in the above way, and so the code ref
2336 # alternative is available. It can do whatever it needs to. The other
2337 # array elements are optional in this case, and the code is free to use or
2338 # ignore them if they are present.
2339 #
2340 # Internally, the constructor unshifts a 0 or 1 onto this array to
2341 # indicate if an early alternative is actually being used or not. This
2342 # makes for easier testing later on.
2343 main::set_access('early', \%early, 'c');
2344
4ed2b786
KW
2345 my %only_early;
2346 main::set_access('only_early', \%only_early, 'c');
2347
0dee21d6
KW
2348 my %required_even_in_debug_skip;
2349 # debug_skip is used to speed up compilation during debugging by skipping
2350 # processing files that are not needed for the task at hand. However,
2351 # some files pretty much can never be skipped, and this is used to specify
2352 # that this is one of them. In order to skip this file, the call to the
2353 # constructor must be edited to comment out this parameter.
2354 main::set_access('required_even_in_debug_skip',
2355 \%required_even_in_debug_skip, 'c');
2356
d8062573
KW
2357 my %withdrawn;
2358 # Some files get removed from the Unicode DB. This is a version object
2359 # giving the first release without this file.
2360 main::set_access('withdrawn', \%withdrawn, 'c');
2361
0dee21d6
KW
2362 my %in_this_release;
2363 # Calculated value from %first_released and %withdrawn. Are we compiling
2364 # a Unicode release which includes this file?
2365 main::set_access('in_this_release', \%in_this_release);
2366
74cd47d0
KW
2367 sub _next_line;
2368 sub _next_line_with_remapped_range;
2369
99870f4d
KW
2370 sub new {
2371 my $class = shift;
2372
2373 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2374 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2375
2376 # Set defaults
2377 $handler{$addr} = \&main::process_generic_property_file;
4e058f44 2378 $retain_trailing_comments{$addr} = 0;
99870f4d 2379 $non_skip{$addr} = 0;
0dee21d6 2380 $skip{$addr} = undef;
99870f4d
KW
2381 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2382 $handle{$addr} = undef;
2383 $added_lines{$addr} = [ ];
74cd47d0 2384 $remapped_lines{$addr} = [ ];
99870f4d 2385 $each_line_handler{$addr} = [ ];
3848e5b7 2386 $eof_handler{$addr} = [ ];
99870f4d
KW
2387 $errors{$addr} = { };
2388 $missings{$addr} = [ ];
a0204bb1 2389 $early{$addr} = [ ];
3274d896 2390 $optional{$addr} = [ ];
99870f4d
KW
2391
2392 # Two positional parameters.
99f78760 2393 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2394 $file{$addr} = main::internal_file_to_platform(shift);
2395 $first_released{$addr} = shift;
2396
2397 # The rest of the arguments are key => value pairs
2398 # %constructor_fields has been set up earlier to list all possible
2399 # ones. Either set or push, depending on how the default has been set
2400 # up just above.
2401 my %args = @_;
2402 foreach my $key (keys %args) {
2403 my $argument = $args{$key};
2404
2405 # Note that the fields are the lower case of the constructor keys
2406 my $hash = $constructor_fields{lc $key};
2407 if (! defined $hash) {
2408 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2409 next;
2410 }
2411 if (ref $hash->{$addr} eq 'ARRAY') {
2412 if (ref $argument eq 'ARRAY') {
2413 foreach my $argument (@{$argument}) {
2414 next if ! defined $argument;
2415 push @{$hash->{$addr}}, $argument;
2416 }
2417 }
2418 else {
2419 push @{$hash->{$addr}}, $argument if defined $argument;
2420 }
2421 }
2422 else {
2423 $hash->{$addr} = $argument;
2424 }
2425 delete $args{$key};
2426 };
2427
0dee21d6
KW
2428 $non_skip{$addr} = 1 if $required_even_in_debug_skip{$addr};
2429
2430 # Convert 0 (meaning don't skip) to undef
2431 undef $skip{$addr} unless $skip{$addr};
2432
3274d896
KW
2433 # Handle the case where this file is optional
2434 my $pod_message_for_non_existent_optional = "";
2435 if ($optional{$addr}->@*) {
2436
2437 # First element is the pod message
2438 $pod_message_for_non_existent_optional
2439 = shift $optional{$addr}->@*;
2440 # Convert a 0 'Optional' argument to an empty list to make later
2441 # code more concise.
2442 if ( $optional{$addr}->@*
2443 && $optional{$addr}->@* == 1
2444 && $optional{$addr}[0] ne ""
2445 && $optional{$addr}[0] !~ /\D/
2446 && $optional{$addr}[0] == 0)
2447 {
2448 $optional{$addr} = [ ];
2449 }
2450 else { # But if the only element doesn't evaluate to 0, make sure
2451 # that this file is indeed considered optional below.
2452 unshift $optional{$addr}->@*, 1;
2453 }
2454 }
2455
0dee21d6 2456 my $progress;
a0204bb1
KW
2457 my $function_instead_of_file = 0;
2458
4ed2b786
KW
2459 if ($early{$addr}->@* && $early{$addr}[-1] eq 'ONLY_EARLY') {
2460 $only_early{$addr} = 1;
2461 pop $early{$addr}->@*;
cbdbe9d4
KW
2462 }
2463
a0204bb1
KW
2464 # If we are compiling a Unicode release earlier than the file became
2465 # available, the constructor may have supplied a substitute
2466 if ($first_released{$addr} gt $v_version && $early{$addr}->@*) {
2467
2468 # Yes, we have a substitute, that we will use; mark it so
2469 unshift $early{$addr}->@*, 1;
2470
2471 # See the definition of %early for what the array elements mean.
4ed2b786
KW
2472 # Note that we have just unshifted onto the array, so the numbers
2473 # below are +1 of those in the %early description.
a0204bb1
KW
2474 # If we have a property this defines, create a table and default
2475 # map for it now (at essentially compile time), so that it will be
2476 # available for the whole of run time. (We will want to add this
2477 # name as an alias when we are using the official property name;
2478 # but this must be deferred until run(), because at construction
2479 # time the official names have yet to be defined.)
2480 if ($early{$addr}[2]) {
2481 my $fate = ($property{$addr}
2482 && $property{$addr} eq $early{$addr}[2])
2483 ? $PLACEHOLDER
2484 : $INTERNAL_ONLY;
2485 my $prop_object = Property->new($early{$addr}[2],
2486 Fate => $fate,
2487 Perl_Extension => 1,
2488 );
2489
3d36fe12
KW
2490 # If not specified by the constructor, use the default mapping
2491 # for the regular property for this substitute one.
4ed2b786
KW
2492 if ($early{$addr}[4]) {
2493 $prop_object->set_default_map($early{$addr}[4]);
3d36fe12
KW
2494 }
2495 elsif ( defined $property{$addr}
2496 && defined $default_mapping{$property{$addr}})
a0204bb1
KW
2497 {
2498 $prop_object
2499 ->set_default_map($default_mapping{$property{$addr}});
2500 }
2501 }
2502
2503 if (ref $early{$addr}[1] eq 'CODE') {
2504 $function_instead_of_file = 1;
2505
2506 # If the first element of the array is a code ref, the others
2507 # are optional.
2508 $handler{$addr} = $early{$addr}[1];
2509 $property{$addr} = $early{$addr}[2]
2510 if defined $early{$addr}[2];
2511 $progress = "substitute $file{$addr}";
2512
2513 undef $file{$addr};
2514 }
2515 else { # Specifying a substitute file
2516
2517 if (! main::file_exists($early{$addr}[1])) {
2518
2519 # If we don't see the substitute file, generate an error
2520 # message giving the needed things, and add it to the list
2521 # of such to output before actual processing happens
2522 # (hence the user finds out all of them in one run).
ce8d64d9
KW
2523 # Instead of creating a general method for NameAliases,
2524 # hard-code it here, as there is unlikely to ever be a
2525 # second one which needs special handling.
2526 my $string_version = ($file{$addr} eq "NameAliases.txt")
2527 ? 'at least 6.1 (the later, the better)'
2528 : sprintf "%vd", $first_released{$addr};
a0204bb1
KW
2529 push @missing_early_files, <<END;
2530'$file{$addr}' version $string_version should be copied to '$early{$addr}[1]'.
2531END
2532 ;
2533 return;
2534 }
2535 $progress = $early{$addr}[1];
2536 $progress .= ", substituting for $file{$addr}" if $file{$addr};
2537 $file{$addr} = $early{$addr}[1];
2538 $property{$addr} = $early{$addr}[2];
0dee21d6 2539
a0204bb1
KW
2540 # Ignore code points not in the version being compiled
2541 push $each_line_handler{$addr}->@*, \&_exclude_unassigned;
2542
2543 if ( $v_version lt v2.0 # Hanguls in this release ...
4ed2b786 2544 && defined $early{$addr}[3]) # ... need special treatment
a0204bb1
KW
2545 {
2546 push $eof_handler{$addr}->@*, \&_fixup_obsolete_hanguls;
2547 }
2548 }
2549
2550 # And this substitute is valid for all releases.
2551 $first_released{$addr} = v0;
2552 }
2553 else { # Normal behavior
0dee21d6 2554 $progress = $file{$addr};
a0204bb1 2555 unshift $early{$addr}->@*, 0; # No substitute
99870f4d
KW
2556 }
2557
0dee21d6
KW
2558 my $file = $file{$addr};
2559 $progress_message{$addr} = "Processing $progress"
2560 unless $progress_message{$addr};
2561
d8062573
KW
2562 # A file should be there if it is within the window of versions for
2563 # which Unicode supplies it
2564 if ($withdrawn{$addr} && $withdrawn{$addr} le $v_version) {
2565 $in_this_release{$addr} = 0;
2566 $skip{$addr} = "";
2567 }
2568 else {
0dee21d6
KW
2569 $in_this_release{$addr} = $first_released{$addr} le $v_version;
2570
a0204bb1
KW
2571 # Check that the file for this object (possibly using a substitute
2572 # for early releases) exists or we have a function alternative
2573 if ( ! $function_instead_of_file
2574 && ! main::file_exists($file))
0dee21d6
KW
2575 {
2576 # Here there is nothing available for this release. This is
2577 # fine if we aren't expecting anything in this release.
2578 if (! $in_this_release{$addr}) {
2579 $skip{$addr} = ""; # Don't remark since we expected
2580 # nothing and got nothing
2581 }
3274d896 2582 elsif ($optional{$addr}->@*) {
0dee21d6 2583
3274d896
KW
2584 # Here the file is optional in this release; Use the
2585 # passed in text to document this case in the pod.
2586 $skip{$addr} = $pod_message_for_non_existent_optional;
0dee21d6
KW
2587 }
2588 elsif ( $in_this_release{$addr}
2589 && ! defined $skip{$addr}
2590 && defined $file)
2591 { # Doesn't exist but should.
2592 $skip{$addr} = "'$file' not found. Possibly Big problems";
2593 Carp::my_carp($skip{$addr});
2594 }
2595 }
2596 elsif ($debug_skip && ! defined $skip{$addr} && ! $non_skip{$addr})
2597 {
2598
2599 # The file exists; if not skipped for another reason, and we are
2600 # skipping most everything during debugging builds, use that as
2601 # the skip reason.
2602 $skip{$addr} = '$debug_skip is on'
2603 }
d8062573 2604 }
0dee21d6
KW
2605
2606 if ( ! $debug_skip
2607 && $non_skip{$addr}
2608 && ! $required_even_in_debug_skip{$addr}
2609 && $verbosity)
2610 {
2611 print "Warning: " . __PACKAGE__ . " constructor for $file has useless 'non_skip' in it\n";
a3a8c5f0 2612 }
99870f4d 2613
0dee21d6 2614 # Here, we have figured out if we will be skipping this file or not.
3274d896
KW
2615 # If so, we add any single property it defines to any passed in
2616 # optional property list. These will be dealt with at run time.
0dee21d6 2617 if (defined $skip{$addr}) {
3274d896
KW
2618 if ($property{$addr}) {
2619 push $optional{$addr}->@*, $property{$addr};
2620 }
2621 } # Otherwise, are going to process the file.
0dee21d6
KW
2622 elsif ($property{$addr}) {
2623
2624 # If the file has a property defined in the constructor for it, it
2625 # means that the property is not listed in the file's entries. So
2626 # add a handler (to the list of line handlers) to insert the
2627 # property name into the lines, to provide a uniform interface to
2628 # the final processing subroutine.
2629 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
09ca89ce 2630 }
83b68635
KW
2631 elsif ($properties{$addr}) {
2632
9be22ca5
KW
2633 # Similarly, there may be more than one property represented on
2634 # each line, with no clue but the constructor input what those
2635 # might be. Add a handler for each line in the input so that it
2636 # creates a separate input line for each property in those input
2637 # lines, thus making them suitable to handle generically.
83b68635
KW
2638
2639 push @{$each_line_handler{$addr}},
2640 sub {
2641 my $file = shift;
2642 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2643
2644 my @fields = split /\s*;\s*/, $_, -1;
2645
2646 if (@fields - 1 > @{$properties{$addr}}) {
2647 $file->carp_bad_line('Extra fields');
2648 $_ = "";
2649 return;
2650 }
2651 my $range = shift @fields; # 0th element is always the
2652 # range
2653
2654 # The next fields in the input line correspond
2655 # respectively to the stored properties.
2656 for my $i (0 .. @{$properties{$addr}} - 1) {
2657 my $property_name = $properties{$addr}[$i];
2658 next if $property_name eq '<ignored>';
2659 $file->insert_adjusted_lines(
2660 "$range; $property_name; $fields[$i]");
2661 }
2662 $_ = "";
2663
2664 return;
2665 };
2666 }
37e2e78e 2667
9be22ca5 2668 { # On non-ascii platforms, we use a special pre-handler
74cd47d0
KW
2669 no strict;
2670 no warnings 'once';
2671 *next_line = (main::NON_ASCII_PLATFORM)
2672 ? *_next_line_with_remapped_range
2673 : *_next_line;
2674 }
2675
40045c2d
KW
2676 &{$construction_time_handler{$addr}}($self)
2677 if $construction_time_handler{$addr};
2678
99870f4d 2679 return $self;
d73e5302
JH
2680 }
2681
cf25bb62 2682
99870f4d
KW
2683 use overload
2684 fallback => 0,
2685 qw("") => "_operator_stringify",
2686 "." => \&main::_operator_dot,
1285127e 2687 ".=" => \&main::_operator_dot_equal,
99870f4d 2688 ;
cf25bb62 2689
99870f4d
KW
2690 sub _operator_stringify {
2691 my $self = shift;
cf25bb62 2692
99870f4d 2693 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2694 }
d73e5302 2695
99870f4d
KW
2696 sub run {
2697 # Process the input object $self. This opens and closes the file and
2698 # calls all the handlers for it. Currently, this can only be called
3848e5b7 2699 # once per file, as it destroy's the EOF handlers
d73e5302 2700
0dee21d6 2701 # flag to make sure extracted files are processed early
f8b69004 2702 state $seen_non_extracted = 0;
0dee21d6 2703
99870f4d
KW
2704 my $self = shift;
2705 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2706
ffe43484 2707 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2708
99870f4d 2709 my $file = $file{$addr};
d73e5302 2710
0dee21d6
KW
2711 if (! $file) {
2712 $handle{$addr} = 'pretend_is_open';
99870f4d
KW
2713 }
2714 else {
f8b69004 2715 if ($seen_non_extracted) {
9be22ca5
KW
2716 if ($file =~ /$EXTRACTED/i) # Some platforms may change the
2717 # case of the file's name
2718 {
1675ea0d 2719 Carp::my_carp_bug(main::join_lines(<<END
99f78760 2720$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2721anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2722have subtle problems
2723END
2724 ));
2725 }
2726 }
2727 elsif ($EXTRACTED_DIR
0dee21d6
KW
2728
2729 # We only do this check for generic property files
2730 && $handler{$addr} == \&main::process_generic_property_file
2731
f8b69004 2732 && $file !~ /$EXTRACTED/i)
99870f4d
KW
2733 {
2734 # We don't set this (by the 'if' above) if we have no
2735 # extracted directory, so if running on an early version,
2736 # this test won't work. Not worth worrying about.
f8b69004 2737 $seen_non_extracted = 1;
99870f4d
KW
2738 }
2739
9be22ca5 2740 # Mark the file as having being processed, and warn if it
99870f4d
KW
2741 # isn't a file we are expecting. As we process the files,
2742 # they are deleted from the hash, so any that remain at the
2743 # end of the program are files that we didn't process.
517956bf 2744 my $fkey = File::Spec->rel2abs($file);
0dee21d6 2745 my $exists = delete $potential_files{lc($fkey)};
faf3cf6b 2746
0dee21d6
KW
2747 Carp::my_carp("Was not expecting '$file'.")
2748 if $exists && ! $in_this_release{$addr};
2749
a0204bb1
KW
2750 # If there is special handling for compiling Unicode releases
2751 # earlier than the first one in which Unicode defines this
2752 # property ...
2753 if ($early{$addr}->@* > 1) {
2754
2755 # Mark as processed any substitute file that would be used in
2756 # such a release
2757 $fkey = File::Spec->rel2abs($early{$addr}[1]);
2758 delete $potential_files{lc($fkey)};
2759
2760 # As commented in the constructor code, when using the
2761 # official property, we still have to allow the publicly
2762 # inaccessible early name so that the core code which uses it
2763 # will work regardless.
4ed2b786
KW
2764 if ( ! $only_early{$addr}
2765 && ! $early{$addr}[0]
2766 && $early{$addr}->@* > 2)
2767 {
a0204bb1 2768 my $early_property_name = $early{$addr}[2];
4ed2b786 2769 if ($property{$addr} ne $early_property_name) {
a0204bb1
KW
2770 main::property_ref($property{$addr})
2771 ->add_alias($early_property_name);
2772 }
2773 }
2774 }
2775
0dee21d6
KW
2776 # We may be skipping this file ...
2777 if (defined $skip{$addr}) {
2778
2779 # If the file isn't supposed to be in this release, there is
2780 # nothing to do
2781 if ($in_this_release{$addr}) {
2782
2783 # But otherwise, we may print a message
2784 if ($debug_skip) {
2785 print STDERR "Skipping input file '$file'",
2786 " because '$skip{$addr}'\n";
2787 }
2788
2789 # And add it to the list of skipped files, which is later
2790 # used to make the pod
2791 $skipped_files{$file} = $skip{$addr};
3274d896
KW
2792
2793 # The 'optional' list contains properties that are also to
2794 # be skipped along with the file. (There may also be
2795 # digits which are just placeholders to make sure it isn't
2796 # an empty list
2797 foreach my $property ($optional{$addr}->@*) {
2798 next unless $property =~ /\D/;
2799 my $prop_object = main::property_ref($property);
2800 next unless defined $prop_object;
2801 $prop_object->set_fate($SUPPRESSED, $skip{$addr});
2802 }
0dee21d6 2803 }
99870f4d 2804
37e2e78e
KW
2805 return;
2806 }
2807
9be22ca5
KW
2808 # Here, we are going to process the file. Open it, converting the
2809 # slashes used in this program into the proper form for the OS
99870f4d
KW
2810 my $file_handle;
2811 if (not open $file_handle, "<", $file) {
2812 Carp::my_carp("Can't open $file. Skipping: $!");
d3664eb9 2813 return;
99870f4d
KW
2814 }
2815 $handle{$addr} = $file_handle; # Cache the open file handle
9e65c3f4 2816
0dee21d6
KW
2817 # If possible, make sure that the file is the correct version.
2818 # (This data isn't available on early Unicode releases or in
2819 # UnicodeData.txt.) We don't do this check if we are using a
2820 # substitute file instead of the official one (though the code
2821 # could be extended to do so).
2822 if ($in_this_release{$addr}
a0204bb1 2823 && ! $early{$addr}[0]
0dee21d6
KW
2824 && lc($file) ne 'unicodedata.txt')
2825 {
96f226dc 2826 if ($file !~ /^Unihan/i) {
0dee21d6
KW
2827
2828 # The non-Unihan files started getting version numbers in
2829 # 3.2, but some files in 4.0 are unchanged from 3.2, and
2830 # marked as 3.2. 4.0.1 is the first version where there
2831 # are no files marked as being from less than 4.0, though
2832 # some are marked as 4.0. In versions after that, the
2833 # numbers are correct.
2834 if ($v_version ge v4.0.1) {
2835 $_ = <$file_handle>; # The version number is in the
2836 # very first line
2837 if ($_ !~ / - $string_version \. /x) {
2838 chomp;
2839 $_ =~ s/^#\s*//;
2840
2841 # 4.0.1 had some valid files that weren't updated.
2842 if (! ($v_version eq v4.0.1 && $_ =~ /4\.0\.0/)) {
2843 die Carp::my_carp("File '$file' is version "
2844 . "'$_'. It should be "
2845 . "version $string_version");
2846 }
2847 }
cafe9cf0 2848 }
96f226dc 2849 }
0dee21d6
KW
2850 elsif ($v_version ge v6.0.0) { # Unihan
2851
2852 # Unihan files didn't get accurate version numbers until
2853 # 6.0. The version is somewhere in the first comment
2854 # block
96f226dc
KW
2855 while (<$file_handle>) {
2856 if ($_ !~ /^#/) {
0dee21d6
KW
2857 Carp::my_carp_bug("Could not find the expected "
2858 . "version info in file '$file'");
96f226dc
KW
2859 last;
2860 }
2861 chomp;
2862 $_ =~ s/^#\s*//;
2863 next if $_ !~ / version: /x;
2864 last if $_ =~ /$string_version/;
0dee21d6
KW
2865 die Carp::my_carp("File '$file' is version "
2866 . "'$_'. It should be "
2867 . "version $string_version");
96f226dc
KW
2868 }
2869 }
9e65c3f4 2870 }
99870f4d
KW
2871 }
2872
0dee21d6 2873 print "$progress_message{$addr}\n" if $verbosity >= $PROGRESS;
99870f4d
KW
2874
2875 # Call any special handler for before the file.
2876 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2877
2878 # Then the main handler
2879 &{$handler{$addr}}($self);
2880
2881 # Then any special post-file handler.
2882 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2883
2884 # If any errors have been accumulated, output the counts (as the first
2885 # error message in each class was output when it was encountered).
2886 if ($errors{$addr}) {
2887 my $total = 0;
2888 my $types = 0;
2889 foreach my $error (keys %{$errors{$addr}}) {
2890 $total += $errors{$addr}->{$error};
2891 delete $errors{$addr}->{$error};
2892 $types++;
2893 }
2894 if ($total > 1) {
2895 my $message
2896 = "A total of $total lines had errors in $file. ";
2897
2898 $message .= ($types == 1)
2899 ? '(Only the first one was displayed.)'
2900 : '(Only the first of each type was displayed.)';
2901 Carp::my_carp($message);
2902 }
2903 }
2904
2905 if (@{$missings{$addr}}) {
2906 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2907 }
2908
2909 # If a real file handle, close it.
2910 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2911 ref $handle{$addr};
2912 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2913 # the file, as opposed to undef
2914 return;
2915 }
2916
74cd47d0 2917 sub _next_line {
99870f4d
KW
2918 # Sets $_ to be the next logical input line, if any. Returns non-zero
2919 # if such a line exists. 'logical' means that any lines that have
2920 # been added via insert_lines() will be returned in $_ before the file
2921 # is read again.
2922
2923 my $self = shift;
2924 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2925
ffe43484 2926 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2927
2928 # Here the file is open (or if the handle is not a ref, is an open
2929 # 'virtual' file). Get the next line; any inserted lines get priority
2930 # over the file itself.
2931 my $adjusted;
2932
2933 LINE:
2934 while (1) { # Loop until find non-comment, non-empty line
2935 #local $to_trace = 1 if main::DEBUG;
2936 my $inserted_ref = shift @{$added_lines{$addr}};
2937 if (defined $inserted_ref) {
2938 ($adjusted, $_) = @{$inserted_ref};
2939 trace $adjusted, $_ if main::DEBUG && $to_trace;
2940 return 1 if $adjusted;
2941 }
2942 else {
2943 last if ! ref $handle{$addr}; # Don't read unless is real file
2944 last if ! defined ($_ = readline $handle{$addr});
2945 }
2946 chomp;
2947 trace $_ if main::DEBUG && $to_trace;
2948
2949 # See if this line is the comment line that defines what property
2950 # value that code points that are not listed in the file should
2951 # have. The format or existence of these lines is not guaranteed
2952 # by Unicode since they are comments, but the documentation says
2953 # that this was added for machine-readability, so probably won't
2954 # change. This works starting in Unicode Version 5.0. They look
2955 # like:
2956 #
2957 # @missing: 0000..10FFFF; Not_Reordered
2958 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2959 # @missing: 0000..10FFFF; ; NaN
2960 #
2961 # Save the line for a later get_missings() call.
2962 if (/$missing_defaults_prefix/) {
2963 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2964 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2965 }
2966 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2967 my @defaults = split /\s* ; \s*/x, $_;
2968
2969 # The first field is the @missing, which ends in a
2970 # semi-colon, so can safely shift.
2971 shift @defaults;
2972
2973 # Some of these lines may have empty field placeholders
2974 # which get in the way. An example is:
2975 # @missing: 0000..10FFFF; ; NaN
2976 # Remove them. Process starting from the top so the
2977 # splice doesn't affect things still to be looked at.
2978 for (my $i = @defaults - 1; $i >= 0; $i--) {
2979 next if $defaults[$i] ne "";
2980 splice @defaults, $i, 1;
2981 }
2982
2983 # What's left should be just the property (maybe) and the
2984 # default. Having only one element means it doesn't have
2985 # the property.
2986 my $default;
2987 my $property;
2988 if (@defaults >= 1) {
2989 if (@defaults == 1) {
2990 $default = $defaults[0];
2991 }
2992 else {
2993 $property = $defaults[0];
2994 $default = $defaults[1];
2995 }
2996 }
2997
2998 if (@defaults < 1
2999 || @defaults > 2
3000 || ($default =~ /^</
3001 && $default !~ /^<code *point>$/i
09f8d0ac
KW
3002 && $default !~ /^<none>$/i
3003 && $default !~ /^<script>$/i))
99870f4d
KW
3004 {
3005 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
3006 }
3007 else {
3008
3009 # If the property is missing from the line, it should
3010 # be the one for the whole file
3011 $property = $property{$addr} if ! defined $property;
3012
3013 # Change <none> to the null string, which is what it
3014 # really means. If the default is the code point
3015 # itself, set it to <code point>, which is what
3016 # Unicode uses (but sometimes they've forgotten the
3017 # space)
3018 if ($default =~ /^<none>$/i) {
3019 $default = "";
3020 }
3021 elsif ($default =~ /^<code *point>$/i) {
3022 $default = $CODE_POINT;
3023 }
09f8d0ac
KW
3024 elsif ($default =~ /^<script>$/i) {
3025
3026 # Special case this one. Currently is from
3027 # ScriptExtensions.txt, and means for all unlisted
3028 # code points, use their Script property values.
3029 # For the code points not listed in that file, the
3030 # default value is 'Unknown'.
3031 $default = "Unknown";
3032 }
99870f4d
KW
3033
3034 # Store them as a sub-arrays with both components.
3035 push @{$missings{$addr}}, [ $default, $property ];
3036 }
3037 }
3038
3039 # There is nothing for the caller to process on this comment
3040 # line.
3041 next;
3042 }
3043
4e058f44
KW
3044 # Unless to keep, remove comments. If to keep, ignore
3045 # comment-only lines
3046 if ($retain_trailing_comments{$addr}) {
3047 next if / ^ \s* \# /x;
3048
3049 # But escape any single quotes (done in both the comment and
3050 # non-comment portion; this could be a bug someday, but not
3051 # likely)
3052 s/'/\\'/g;
3053 }
3054 else {
3055 s/#.*//;
3056 }
3057
3058 # Remove trailing space, and skip this line if the result is empty
99870f4d
KW
3059 s/\s+$//;
3060 next if /^$/;
3061
3062 # Call any handlers for this line, and skip further processing of
3063 # the line if the handler sets the line to null.
3064 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
3065 &{$sub_ref}($self);
3066 next LINE if /^$/;
3067 }
3068
3069 # Here the line is ok. return success.
3070 return 1;
3071 } # End of looping through lines.
3072
3848e5b7 3073 # If there are EOF handlers, call each (only once) and if it generates
99870f4d 3074 # more lines to process go back in the loop to handle them.
3848e5b7
KW
3075 while ($eof_handler{$addr}->@*) {
3076 &{$eof_handler{$addr}[0]}($self);
3077 shift $eof_handler{$addr}->@*; # Currently only get one shot at it.
99870f4d
KW
3078 goto LINE if $added_lines{$addr};
3079 }
3080
3081 # Return failure -- no more lines.
3082 return 0;
3083
3084 }
3085
74cd47d0
KW
3086 sub _next_line_with_remapped_range {
3087 my $self = shift;
3088 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3089
3090 # like _next_line(), but for use on non-ASCII platforms. It sets $_
3091 # to be the next logical input line, if any. Returns non-zero if such
3092 # a line exists. 'logical' means that any lines that have been added
3093 # via insert_lines() will be returned in $_ before the file is read
3094 # again.
3095 #
3096 # The difference from _next_line() is that this remaps the Unicode
3097 # code points in the input to those of the native platform. Each
3098 # input line contains a single code point, or a single contiguous
3099 # range of them This routine splits each range into its individual
3100 # code points and caches them. It returns the cached values,
3101 # translated into their native equivalents, one at a time, for each
3102 # call, before reading the next line. Since native values can only be
3103 # a single byte wide, no translation is needed for code points above
3104 # 0xFF, and ranges that are entirely above that number are not split.
3105 # If an input line contains the range 254-1000, it would be split into
3106 # three elements: 254, 255, and 256-1000. (The downstream table
3107 # insertion code will sort and coalesce the individual code points
3108 # into appropriate ranges.)
3109
3110 my $addr = do { no overloading; pack 'J', $self; };
3111
3112 while (1) {
3113
3114 # Look in cache before reading the next line. Return any cached
3115 # value, translated
3116 my $inserted = shift @{$remapped_lines{$addr}};
3117 if (defined $inserted) {
3118 trace $inserted if main::DEBUG && $to_trace;
3119 $_ = $inserted =~ s/^ ( \d+ ) /sprintf("%04X", utf8::unicode_to_native($1))/xer;
3120 trace $_ if main::DEBUG && $to_trace;
3121 return 1;
3122 }
3123
3124 # Get the next line.
3125 return 0 unless _next_line($self);
3126
3127 # If there is a special handler for it, return the line,
3128 # untranslated. This should happen only for files that are
3129 # special, not being code-point related, such as property names.
3130 return 1 if $handler{$addr}
3131 != \&main::process_generic_property_file;
3132
3133 my ($range, $property_name, $map, @remainder)
3134 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3135
3136 if (@remainder
3137 || ! defined $property_name
3138 || $range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3139 {
3140 Carp::my_carp_bug("Unrecognized input line '$_'. Ignored");
3141 }
3142
3143 my $low = hex $1;
3144 my $high = (defined $2) ? hex $2 : $low;
3145
3146 # If the input maps the range to another code point, remap the
3147 # target if it is between 0 and 255.
3148 my $tail;
3149 if (defined $map) {
3150 $map =~ s/\b 00 ( [0-9A-F]{2} ) \b/sprintf("%04X", utf8::unicode_to_native(hex $1))/gxe;
3151 $tail = "$property_name; $map";
3152 $_ = "$range; $tail";
3153 }
3154 else {
3155 $tail = $property_name;
3156 }
3157
3158 # If entire range is above 255, just return it, unchanged (except
3159 # any mapped-to code point, already changed above)
3160 return 1 if $low > 255;
3161
3162 # Cache an entry for every code point < 255. For those in the
3163 # range above 255, return a dummy entry for just that portion of
3164 # the range. Note that this will be out-of-order, but that is not
3165 # a problem.
3166 foreach my $code_point ($low .. $high) {
3167 if ($code_point > 255) {
3168 $_ = sprintf "%04X..%04X; $tail", $code_point, $high;
3169 return 1;
3170 }
3171 push @{$remapped_lines{$addr}}, "$code_point; $tail";
3172 }
3173 } # End of looping through lines.
3174
3175 # NOTREACHED
3176 }
3177
99870f4d
KW
3178# Not currently used, not fully tested.
3179# sub peek {
f67a5002 3180# # Non-destructive lookahead one non-adjusted, non-comment, non-blank
99870f4d
KW
3181# # record. Not callable from an each_line_handler(), nor does it call
3182# # an each_line_handler() on the line.
3183#
3184# my $self = shift;
ffe43484 3185# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3186#
3187# foreach my $inserted_ref (@{$added_lines{$addr}}) {
3188# my ($adjusted, $line) = @{$inserted_ref};
3189# next if $adjusted;
3190#
3191# # Remove comments and trailing space, and return a non-empty
3192# # resulting line
3193# $line =~ s/#.*//;
3194# $line =~ s/\s+$//;
3195# return $line if $line ne "";
3196# }
3197#
3198# return if ! ref $handle{$addr}; # Don't read unless is real file
3199# while (1) { # Loop until find non-comment, non-empty line
3200# local $to_trace = 1 if main::DEBUG;
3201# trace $_ if main::DEBUG && $to_trace;
3202# return if ! defined (my $line = readline $handle{$addr});
3203# chomp $line;
3204# push @{$added_lines{$addr}}, [ 0, $line ];
3205#
3206# $line =~ s/#.*//;
3207# $line =~ s/\s+$//;
3208# return $line if $line ne "";
3209# }
3210#
3211# return;
3212# }
3213
3214
3215 sub insert_lines {
3216 # Lines can be inserted so that it looks like they were in the input
3217 # file at the place it was when this routine is called. See also
3218 # insert_adjusted_lines(). Lines inserted via this routine go through
3219 # any each_line_handler()
3220
3221 my $self = shift;
3222
3223 # Each inserted line is an array, with the first element being 0 to
3224 # indicate that this line hasn't been adjusted, and needs to be
3225 # processed.
f998e60c 3226 no overloading;
051df77b 3227 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
3228 return;
3229 }
3230
3231 sub insert_adjusted_lines {
3232 # Lines can be inserted so that it looks like they were in the input
3233 # file at the place it was when this routine is called. See also
3234 # insert_lines(). Lines inserted via this routine are already fully
3235 # adjusted, ready to be processed; each_line_handler()s handlers will
3236 # not be called. This means this is not a completely general
3237 # facility, as only the last each_line_handler on the stack should
3238 # call this. It could be made more general, by passing to each of the
3239 # line_handlers their position on the stack, which they would pass on
3240 # to this routine, and that would replace the boolean first element in
3241 # the anonymous array pushed here, so that the next_line routine could
3242 # use that to call only those handlers whose index is after it on the
3243 # stack. But this is overkill for what is needed now.
3244
3245 my $self = shift;
3246 trace $_[0] if main::DEBUG && $to_trace;
3247
3248 # Each inserted line is an array, with the first element being 1 to
3249 # indicate that this line has been adjusted
f998e60c 3250 no overloading;
051df77b 3251 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
3252 return;
3253 }
3254
3255 sub get_missings {
3256 # Returns the stored up @missings lines' values, and clears the list.
3257 # The values are in an array, consisting of the default in the first
3258 # element, and the property in the 2nd. However, since these lines
3259 # can be stacked up, the return is an array of all these arrays.
3260
3261 my $self = shift;
3262 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3263
ffe43484 3264 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3265
3266 # If not accepting a list return, just return the first one.
3267 return shift @{$missings{$addr}} unless wantarray;
3268
3269 my @return = @{$missings{$addr}};
3270 undef @{$missings{$addr}};
3271 return @return;
3272 }
3273
a0204bb1
KW
3274 sub _exclude_unassigned {
3275
3276 # Takes the range in $_ and excludes code points that aren't assigned
3277 # in this release
3278
3279 state $skip_inserted_count = 0;
3280
3281 # Ignore recursive calls.
3282 if ($skip_inserted_count) {
3283 $skip_inserted_count--;
3284 return;
3285 }
3286
3287 # Find what code points are assigned in this release
3288 main::calculate_Assigned() if ! defined $Assigned;
3289
3290 my $self = shift;
3291 my $addr = do { no overloading; pack 'J', $self; };
3292 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3293
3294 my ($range, @remainder)
3295 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3296
3297 # Examine the range.
3298 if ($range =~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3299 {
3300 my $low = hex $1;
3301 my $high = (defined $2) ? hex $2 : $low;
3302
3303 # Split the range into subranges of just those code points in it
3304 # that are assigned.
3305 my @ranges = (Range_List->new(Initialize
3306 => Range->new($low, $high)) & $Assigned)->ranges;
3307
3308 # Do nothing if nothing in the original range is assigned in this
3309 # release; handle normally if everything is in this release.
3310 if (! @ranges) {
3311 $_ = "";
3312 }
3313 elsif (@ranges != 1) {
3314
3315 # Here, some code points in the original range aren't in this
3316 # release; @ranges gives the ones that are. Create fake input
3317 # lines for each of the ranges, and set things up so that when
3318 # this routine is called on that fake input, it will do
3319 # nothing.
3320 $skip_inserted_count = @ranges;
3321 my $remainder = join ";", @remainder;
3322 for my $range (@ranges) {
3323 $self->insert_lines(sprintf("%04X..%04X;%s",
3324 $range->start, $range->end, $remainder));
3325 }
3326 $_ = ""; # The original range is now defunct.
3327 }
3328 }
3329
3330 return;
3331 }
3332
3333 sub _fixup_obsolete_hanguls {
3334
3335 # This is called only when compiling Unicode version 1. All Unicode
3336 # data for subsequent releases assumes that the code points that were
3337 # Hangul syllables in this release only are something else, so if
3338 # using such data, we have to override it
3339
3340 my $self = shift;
3341 my $addr = do { no overloading; pack 'J', $self; };
3342 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3343
3344 my $object = main::property_ref($property{$addr});
53adf6a2
KW
3345 $object->add_map($FIRST_REMOVED_HANGUL_SYLLABLE,
3346 $FINAL_REMOVED_HANGUL_SYLLABLE,
a0204bb1
KW
3347 $early{$addr}[3], # Passed-in value for these
3348 Replace => $UNCONDITIONALLY);
3349 }
3350
99870f4d
KW
3351 sub _insert_property_into_line {
3352 # Add a property field to $_, if this file requires it.
3353
f998e60c 3354 my $self = shift;
ffe43484 3355 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 3356 my $property = $property{$addr};
99870f4d
KW
3357 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3358
3359 $_ =~ s/(;|$)/; $property$1/;
3360 return;
3361 }
3362
3363 sub carp_bad_line {
3364 # Output consistent error messages, using either a generic one, or the
3365 # one given by the optional parameter. To avoid gazillions of the
3366 # same message in case the syntax of a file is way off, this routine
3367 # only outputs the first instance of each message, incrementing a
3368 # count so the totals can be output at the end of the file.
3369
3370 my $self = shift;
3371 my $message = shift;
3372 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3373
ffe43484 3374 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3375
3376 $message = 'Unexpected line' unless $message;
3377
3378 # No trailing punctuation so as to fit with our addenda.
3379 $message =~ s/[.:;,]$//;
3380
3381 # If haven't seen this exact message before, output it now. Otherwise
3382 # increment the count of how many times it has occurred
3383 unless ($errors{$addr}->{$message}) {
3384 Carp::my_carp("$message in '$_' in "
f998e60c 3385 . $file{$addr}
99870f4d
KW
3386 . " at line $.. Skipping this line;");
3387 $errors{$addr}->{$message} = 1;
3388 }
3389 else {
3390 $errors{$addr}->{$message}++;
3391 }
3392
3393 # Clear the line to prevent any further (meaningful) processing of it.
3394 $_ = "";
3395
3396 return;
3397 }
3398} # End closure
3399
3400package Multi_Default;
3401
3402# Certain properties in early versions of Unicode had more than one possible
3403# default for code points missing from the files. In these cases, one
3404# default applies to everything left over after all the others are applied,
3405# and for each of the others, there is a description of which class of code
3406# points applies to it. This object helps implement this by storing the
3407# defaults, and for all but that final default, an eval string that generates
3408# the class that it applies to.
3409
3410
3411{ # Closure
3412
3413 main::setup_package();
3414
3415 my %class_defaults;
3416 # The defaults structure for the classes
3417 main::set_access('class_defaults', \%class_defaults);
3418
3419 my %other_default;
3420 # The default that applies to everything left over.
3421 main::set_access('other_default', \%other_default, 'r');
3422
3423
3424 sub new {
3425 # The constructor is called with default => eval pairs, terminated by
3426 # the left-over default. e.g.
3427 # Multi_Default->new(
3428 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
3429 # - 0x200D',
3430 # 'R' => 'some other expression that evaluates to code points',
3431 # .
3432 # .
3433 # .
3434 # 'U'));
68c3923f
KW
3435 # It is best to leave the final value be the one that matches the
3436 # above-Unicode code points.
99870f4d
KW
3437
3438 my $class = shift;
3439
3440 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 3441 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3442
3443 while (@_ > 1) {
3444 my $default = shift;
3445 my $eval = shift;
3446 $class_defaults{$addr}->{$default} = $eval;
3447 }
3448
3449 $other_default{$addr} = shift;
3450
3451 return $self;
3452 }
3453
3454 sub get_next_defaults {
3455 # Iterates and returns the next class of defaults.
3456 my $self = shift;
3457 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3458
ffe43484 3459 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3460
3461 return each %{$class_defaults{$addr}};
3462 }
3463}
3464
3465package Alias;
3466
3467# An alias is one of the names that a table goes by. This class defines them
3468# including some attributes. Everything is currently setup in the
3469# constructor.
3470
3471
3472{ # Closure
3473
3474 main::setup_package();
3475
3476 my %name;
3477 main::set_access('name', \%name, 'r');
3478
3479 my %loose_match;
c12f2655 3480 # Should this name match loosely or not.
99870f4d
KW
3481 main::set_access('loose_match', \%loose_match, 'r');
3482
33e96e72
KW
3483 my %make_re_pod_entry;
3484 # Some aliases should not get their own entries in the re section of the
3485 # pod, because they are covered by a wild-card, and some we want to
3486 # discourage use of. Binary
f82fe4ba 3487 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r', 's');
99870f4d 3488
fd1e3e84
KW
3489 my %ucd;
3490 # Is this documented to be accessible via Unicode::UCD
3491 main::set_access('ucd', \%ucd, 'r', 's');
3492
99870f4d
KW
3493 my %status;
3494 # Aliases have a status, like deprecated, or even suppressed (which means
3495 # they don't appear in documentation). Enum
3496 main::set_access('status', \%status, 'r');
3497
0eac1e20 3498 my %ok_as_filename;
99870f4d
KW
3499 # Similarly, some aliases should not be considered as usable ones for
3500 # external use, such as file names, or we don't want documentation to
3501 # recommend them. Boolean
0eac1e20 3502 main::set_access('ok_as_filename', \%ok_as_filename, 'r');
99870f4d
KW
3503
3504 sub new {
3505 my $class = shift;
3506
3507 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3508 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3509
3510 $name{$addr} = shift;
3511 $loose_match{$addr} = shift;
33e96e72 3512 $make_re_pod_entry{$addr} = shift;
0eac1e20 3513 $ok_as_filename{$addr} = shift;
99870f4d 3514 $status{$addr} = shift;
fd1e3e84 3515 $ucd{$addr} = shift;
99870f4d
KW
3516
3517 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3518
3519 # Null names are never ok externally
0eac1e20 3520 $ok_as_filename{$addr} = 0 if $name{$addr} eq "";
99870f4d
KW
3521
3522 return $self;
3523 }
3524}
3525
3526package Range;
3527
3528# A range is the basic unit for storing code points, and is described in the
3529# comments at the beginning of the program. Each range has a starting code
3530# point; an ending code point (not less than the starting one); a value
3531# that applies to every code point in between the two end-points, inclusive;
3532# and an enum type that applies to the value. The type is for the user's
3533# convenience, and has no meaning here, except that a non-zero type is
3534# considered to not obey the normal Unicode rules for having standard forms.
3535#
3536# The same structure is used for both map and match tables, even though in the
3537# latter, the value (and hence type) is irrelevant and could be used as a
3538# comment. In map tables, the value is what all the code points in the range
3539# map to. Type 0 values have the standardized version of the value stored as
3540# well, so as to not have to recalculate it a lot.
3541
3542sub trace { return main::trace(@_); }
3543
3544{ # Closure
3545
3546 main::setup_package();
3547
3548 my %start;
3549 main::set_access('start', \%start, 'r', 's');
3550
3551 my %end;
3552 main::set_access('end', \%end, 'r', 's');
3553
3554 my %value;
3555 main::set_access('value', \%value, 'r');
3556
3557 my %type;
3558 main::set_access('type', \%type, 'r');
3559
3560 my %standard_form;
3561 # The value in internal standard form. Defined only if the type is 0.
3562 main::set_access('standard_form', \%standard_form);
3563
3564 # Note that if these fields change, the dump() method should as well
3565
3566 sub new {
3567 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
3568 my $class = shift;
3569
3570 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3571 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3572
3573 $start{$addr} = shift;
3574 $end{$addr} = shift;
3575
3576 my %args = @_;
3577
3578 my $value = delete $args{'Value'}; # Can be 0
3579 $value = "" unless defined $value;
3580 $value{$addr} = $value;
3581
3582 $type{$addr} = delete $args{'Type'} || 0;
3583
3584 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3585
99870f4d
KW
3586 return $self;
3587 }
3588
3589 use overload
3590 fallback => 0,
3591 qw("") => "_operator_stringify",
3592 "." => \&main::_operator_dot,
1285127e 3593 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3594 ;
3595
3596 sub _operator_stringify {
3597 my $self = shift;
ffe43484 3598 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3599
3600 # Output it like '0041..0065 (value)'
3601 my $return = sprintf("%04X", $start{$addr})
3602 . '..'
3603 . sprintf("%04X", $end{$addr});
3604 my $value = $value{$addr};
3605 my $type = $type{$addr};
3606 $return .= ' (';
3607 $return .= "$value";
3608 $return .= ", Type=$type" if $type != 0;
3609 $return .= ')';
3610
3611 return $return;
3612 }
3613
3614 sub standard_form {
c292d35a
NC
3615 # Calculate the standard form only if needed, and cache the result.
3616 # The standard form is the value itself if the type is special.
3617 # This represents a considerable CPU and memory saving - at the time
3618 # of writing there are 368676 non-special objects, but the standard
3619 # form is only requested for 22047 of them - ie about 6%.
99870f4d
KW
3620
3621 my $self = shift;
3622 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3623
ffe43484 3624 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3625
3626 return $standard_form{$addr} if defined $standard_form{$addr};
c292d35a
NC
3627
3628 my $value = $value{$addr};
3629 return $value if $type{$addr};
3630 return $standard_form{$addr} = main::standardize($value);
99870f4d
KW
3631 }
3632
3633 sub dump {
3634 # Human, not machine readable. For machine readable, comment out this
3635 # entire routine and let the standard one take effect.
3636 my $self = shift;
3637 my $indent = shift;
3638 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3639
ffe43484 3640 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3641
3642 my $return = $indent
3643 . sprintf("%04X", $start{$addr})
3644 . '..'
3645 . sprintf("%04X", $end{$addr})
3646 . " '$value{$addr}';";
3647 if (! defined $standard_form{$addr}) {
3648 $return .= "(type=$type{$addr})";
3649 }
3650 elsif ($standard_form{$addr} ne $value{$addr}) {
3651 $return .= "(standard '$standard_form{$addr}')";
3652 }
3653 return $return;
3654 }
3655} # End closure
3656
3657package _Range_List_Base;
3658
3659# Base class for range lists. A range list is simply an ordered list of
3660# ranges, so that the ranges with the lowest starting numbers are first in it.
3661#
3662# When a new range is added that is adjacent to an existing range that has the
3663# same value and type, it merges with it to form a larger range.
3664#
3665# Ranges generally do not overlap, except that there can be multiple entries
3666# of single code point ranges. This is because of NameAliases.txt.
3667#
3668# In this program, there is a standard value such that if two different
3669# values, have the same standard value, they are considered equivalent. This
3670# value was chosen so that it gives correct results on Unicode data
3671
3672# There are a number of methods to manipulate range lists, and some operators
3673# are overloaded to handle them.
3674
99870f4d
KW
3675sub trace { return main::trace(@_); }
3676
3677{ # Clos