This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
mktables: Handle platforms with 3 digit exponents
[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.
8a6698d7 48my $version_of_mk_invlist_bounds = v11.0.0;
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
7fdcbfeb 138# well, which will always be 0), and arbitrarily set to the empty string.
99870f4d
KW
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
82bfada9 655# "" in the $string_compare_versions definition just below to a string
8f21889c
KW
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
c0734505
KW
898# Obsoleted
899if ($v_version ge v11.0.0) {
900 push @tables_that_may_be_empty, qw(
901 Grapheme_Cluster_Break=E_Base
902 Grapheme_Cluster_Break=E_Base_GAZ
903 Grapheme_Cluster_Break=E_Modifier
904 Grapheme_Cluster_Break=Glue_After_Zwj
905 Word_Break=E_Base
906 Word_Break=E_Base_GAZ
907 Word_Break=E_Modifier
908 Word_Break=Glue_After_Zwj);
909}
910
9be22ca5
KW
911# Enum values for to_output_map() method in the Map_Table package. (0 is don't
912# output)
272501f6
KW
913my $EXTERNAL_MAP = 1;
914my $INTERNAL_MAP = 2;
ce712c88 915my $OUTPUT_ADJUSTED = 3;
272501f6 916
fcf1973c
KW
917# To override computed values for writing the map tables for these properties.
918# The default for enum map tables is to write them out, so that the Unicode
919# .txt files can be removed, but all the data to compute any property value
920# for any code point is available in a more compact form.
921my %global_to_output_map = (
922 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
923 # get stuck supporting it if things change. Since it is a STRING
924 # property, it normally would be listed in the pod, but INTERNAL_MAP
925 # suppresses that.
fcf1973c
KW
926 Unicode_1_Name => $INTERNAL_MAP,
927
928 Present_In => 0, # Suppress, as easily computed from Age
74cd47d0
KW
929 Block => (NON_ASCII_PLATFORM) ? 1 : 0, # Suppress, as Blocks.txt is
930 # retained, but needed for
931 # non-ASCII
53d34b6c
KW
932
933 # Suppress, as mapping can be found instead from the
934 # Perl_Decomposition_Mapping file
935 Decomposition_Type => 0,
fcf1973c
KW
936);
937
99870f4d
KW
938# There are several types of obsolete properties defined by Unicode. These
939# must be hand-edited for every new Unicode release.
940my %why_deprecated; # Generates a deprecated warning message if used.
941my %why_stabilized; # Documentation only
942my %why_obsolete; # Documentation only
943
944{ # Closure
8364f9eb 945 my $simple = 'Perl uses the more complete version';
99870f4d
KW
946 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
947
948 my $other_properties = 'other properties';
949 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 950 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
951
952 %why_deprecated = (
25c180b8 953 'Grapheme_Link' => 'Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d 954 'Jamo_Short_Name' => $contributory,
25c180b8 955 'Line_Break=Surrogate' => 'Surrogates should never appear in well-formed text, and therefore shouldn\'t be the basis for line breaking',
99870f4d
KW
956 'Other_Alphabetic' => $contributory,
957 'Other_Default_Ignorable_Code_Point' => $contributory,
958 'Other_Grapheme_Extend' => $contributory,
959 'Other_ID_Continue' => $contributory,
960 'Other_ID_Start' => $contributory,
961 'Other_Lowercase' => $contributory,
962 'Other_Math' => $contributory,
963 'Other_Uppercase' => $contributory,
e22aaf5c
KW
964 'Expands_On_NFC' => $why_no_expand,
965 'Expands_On_NFD' => $why_no_expand,
966 'Expands_On_NFKC' => $why_no_expand,
967 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
968 );
969
970 %why_suppressed = (
5f7264c7 971 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
972 # contains the same information, but without the algorithmically
973 # determinable Hangul syllables'. This file is not published, so it's
974 # existence is not noted in the comment.
12fee290 975 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize or prop_invmap() or charprop() in Unicode::UCD::',
99870f4d 976
5f8d1a89
KW
977 # Don't suppress ISO_Comment, as otherwise special handling is needed
978 # to differentiate between it and gc=c, which can be written as 'isc',
979 # which is the same characters as ISO_Comment's short name.
99870f4d 980
12fee290 981 'Name' => "Accessible via \\N{...} or 'use charnames;' or charprop() or prop_invmap() in Unicode::UCD::",
e0b29447 982
12fee290
KW
983 'Simple_Case_Folding' => "$simple. Can access this through casefold(), charprop(), or prop_invmap() in Unicode::UCD",
984 'Simple_Lowercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
985 'Simple_Titlecase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
986 'Simple_Uppercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
99870f4d 987
dac6f618 988 FC_NFKC_Closure => 'Deprecated by Unicode, and supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
989 );
990
1704a0ea
KW
991 foreach my $property (
992
993 # The following are suppressed because they were made contributory
994 # or deprecated by Unicode before Perl ever thought about
995 # supporting them.
996 'Jamo_Short_Name',
997 'Grapheme_Link',
998 'Expands_On_NFC',
999 'Expands_On_NFD',
1000 'Expands_On_NFKC',
1001 'Expands_On_NFKD',
1002
1003 # The following are suppressed because they have been marked
1004 # as deprecated for a sufficient amount of time
1005 'Other_Alphabetic',
1006 'Other_Default_Ignorable_Code_Point',
1007 'Other_Grapheme_Extend',
1008 'Other_ID_Continue',
1009 'Other_ID_Start',
1010 'Other_Lowercase',
1011 'Other_Math',
1012 'Other_Uppercase',
e22aaf5c 1013 ) {
99870f4d
KW
1014 $why_suppressed{$property} = $why_deprecated{$property};
1015 }
cf25bb62 1016
99870f4d
KW
1017 # Customize the message for all the 'Other_' properties
1018 foreach my $property (keys %why_deprecated) {
1019 next if (my $main_property = $property) !~ s/^Other_//;
1020 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
1021 }
1022}
1023
e9c4b4f8
KW
1024if ($write_Unicode_deprecated_tables) {
1025 foreach my $property (keys %why_suppressed) {
1026 delete $why_suppressed{$property} if $property =~
1027 / ^ Other | Grapheme /x;
1028 }
1029}
1030
99870f4d
KW
1031if ($v_version ge 4.0.0) {
1032 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
1033 if ($v_version ge 6.0.0) {
1034 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
1035 }
99870f4d 1036}
5f7264c7 1037if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 1038 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 1039 if ($v_version ge 6.0.0) {
63f74647 1040 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 1041 }
99870f4d
KW
1042}
1043
1044# Probably obsolete forever
1045if ($v_version ge v4.1.0) {
82aed44a
KW
1046 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
1047}
1048if ($v_version ge v6.0.0) {
caa75395 1049 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana" (or both)';
2b352efd 1050 $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
1051}
1052
1053# This program can create files for enumerated-like properties, such as
1054# 'Numeric_Type'. This file would be the same format as for a string
1055# property, with a mapping from code point to its value, so you could look up,
1056# for example, the script a code point is in. But no one so far wants this
1057# mapping, or they have found another way to get it since this is a new
1058# feature. So no file is generated except if it is in this list.
1059my @output_mapped_properties = split "\n", <<END;
1060END
1061
b9e57660
KW
1062# If you want more Unihan properties than the default, you need to add them to
1063# these arrays. Depending on the property type, @missing lines might have to
1064# be added to the second array. A sample entry would be (including the '#'):
1065# @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
99870f4d 1066my @cjk_properties = split "\n", <<'END';
99870f4d 1067END
99870f4d 1068my @cjk_property_values = split "\n", <<'END';
99870f4d
KW
1069END
1070
1071# The input files don't list every code point. Those not listed are to be
1072# defaulted to some value. Below are hard-coded what those values are for
1073# non-binary properties as of 5.1. Starting in 5.0, there are
caa75395 1074# machine-parsable comment lines in the files that give the defaults; so this
99870f4d
KW
1075# list shouldn't have to be extended. The claim is that all missing entries
1076# for binary properties will default to 'N'. Unicode tried to change that in
1077# 5.2, but the beta period produced enough protest that they backed off.
1078#
1079# The defaults for the fields that appear in UnicodeData.txt in this hash must
1080# be in the form that it expects. The others may be synonyms.
1081my $CODE_POINT = '<code point>';
1082my %default_mapping = (
1083 Age => "Unassigned",
1084 # Bidi_Class => Complicated; set in code
1085 Bidi_Mirroring_Glyph => "",
1086 Block => 'No_Block',
1087 Canonical_Combining_Class => 0,
1088 Case_Folding => $CODE_POINT,
1089 Decomposition_Mapping => $CODE_POINT,
1090 Decomposition_Type => 'None',
1091 East_Asian_Width => "Neutral",
1092 FC_NFKC_Closure => $CODE_POINT,
ec2cd995 1093 General_Category => ($v_version le 6.3.0) ? 'Cn' : 'Unassigned',
99870f4d
KW
1094 Grapheme_Cluster_Break => 'Other',
1095 Hangul_Syllable_Type => 'NA',
1096 ISO_Comment => "",
1097 Jamo_Short_Name => "",
1098 Joining_Group => "No_Joining_Group",
1099 # Joining_Type => Complicated; set in code
1100 kIICore => 'N', # Is converted to binary
1101 #Line_Break => Complicated; set in code
1102 Lowercase_Mapping => $CODE_POINT,
1103 Name => "",
1104 Name_Alias => "",
1105 NFC_QC => 'Yes',
1106 NFD_QC => 'Yes',
1107 NFKC_QC => 'Yes',
1108 NFKD_QC => 'Yes',
1109 Numeric_Type => 'None',
1110 Numeric_Value => 'NaN',
1111 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1112 Sentence_Break => 'Other',
1113 Simple_Case_Folding => $CODE_POINT,
1114 Simple_Lowercase_Mapping => $CODE_POINT,
1115 Simple_Titlecase_Mapping => $CODE_POINT,
1116 Simple_Uppercase_Mapping => $CODE_POINT,
1117 Titlecase_Mapping => $CODE_POINT,
1118 Unicode_1_Name => "",
1119 Unicode_Radical_Stroke => "",
1120 Uppercase_Mapping => $CODE_POINT,
1121 Word_Break => 'Other',
1122);
1123
678f13d5 1124### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1125
1126my $HEADER=<<"EOF";
1127# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85 1128# This file is machine-generated by $0 from the Unicode
8f21889c 1129# database, Version $unicode_version. Any changes made here will be lost!
cf25bb62
JH
1130EOF
1131
126c3d4e 1132my $INTERNAL_ONLY_HEADER = <<"EOF";
99870f4d
KW
1133
1134# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1135# This file is for internal use by core Perl only. The format and even the
1136# name or existence of this file are subject to change without notice. Don't
d9ae3878
KW
1137# use it directly. Use Unicode::UCD to access the Unicode character data
1138# base.
99870f4d
KW
1139EOF
1140
1141my $DEVELOPMENT_ONLY=<<"EOF";
1142# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1143# This file contains information artificially constrained to code points
1144# present in Unicode release $string_compare_versions.
1145# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1146# not be used for production.
b6922eda
KW
1147
1148EOF
1149
346a20cf
KW
1150my $MAX_UNICODE_CODEPOINT_STRING = ($v_version ge v2.0.0)
1151 ? "10FFFF"
1152 : "FFFF";
6189eadc
KW
1153my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1154my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d 1155
026d3e0c
KW
1156# We work with above-Unicode code points, up to IV_MAX, but we may want to use
1157# sentinels above that number. Therefore for internal use, we use a much
1158# smaller number, translating it to IV_MAX only for output. The exact number
7fdcbfeb
KW
1159# is immaterial (all above-Unicode code points are treated exactly the same),
1160# but the algorithm requires it to be at least
1161# 2 * $MAX_UNICODE_CODEPOINTS + 1
2d88a86a
KW
1162my $MAX_WORKING_CODEPOINTS= $MAX_UNICODE_CODEPOINT * 8;
1163my $MAX_WORKING_CODEPOINT = $MAX_WORKING_CODEPOINTS - 1;
1164my $MAX_WORKING_CODEPOINT_STRING = sprintf("%X", $MAX_WORKING_CODEPOINT);
1165
026d3e0c 1166my $MAX_PLATFORM_CODEPOINT = ~0 >> 1;
2d88a86a 1167
99870f4d
KW
1168# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1169# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1170# to decrease backtracking. The first regex allows the code point to be at
1171# the end of a word, but to work properly, the word shouldn't end with a valid
1172# hex character. The second one won't match a code point at the end of a
1173# word, and doesn't have the run-on issue
8c32d378
KW
1174my $run_on_code_point_re =
1175 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1176my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d 1177
44b0f8c1 1178# This matches the beginning of the line in the Unicode DB files that give the
99870f4d
KW
1179# defaults for code points not listed (i.e., missing) in the file. The code
1180# depends on this ending with a semi-colon, so it can assume it is a valid
1181# field when the line is split() by semi-colons
346a20cf 1182my $missing_defaults_prefix = qr/^#\s+\@missing:\s+0000\.\.10FFFF\s*;/;
99870f4d
KW
1183
1184# Property types. Unicode has more types, but these are sufficient for our
1185# purposes.
1186my $UNKNOWN = -1; # initialized to illegal value
1187my $NON_STRING = 1; # Either binary or enum
1188my $BINARY = 2;
06f26c45
KW
1189my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1190 # tables, additional true and false tables are
1191 # generated so that false is anything matching the
1192 # default value, and true is everything else.
1193my $ENUM = 4; # Include catalog
1194my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1195
1196# Some input files have lines that give default values for code points not
1197# contained in the file. Sometimes these should be ignored.
1198my $NO_DEFAULTS = 0; # Must evaluate to false
1199my $NOT_IGNORED = 1;
1200my $IGNORED = 2;
1201
1202# Range types. Each range has a type. Most ranges are type 0, for normal,
1203# and will appear in the main body of the tables in the output files, but
1204# there are other types of ranges as well, listed below, that are specially
1205# handled. There are pseudo-types as well that will never be stored as a
1206# type, but will affect the calculation of the type.
1207
1208# 0 is for normal, non-specials
1209my $MULTI_CP = 1; # Sequence of more than code point
1210my $HANGUL_SYLLABLE = 2;
1211my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1212my $NULL = 4; # The map is to the null string; utf8.c can't
1213 # handle these, nor is there an accepted syntax
1214 # for them in \p{} constructs
f86864ac 1215my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1216 # otherwise be $MULTI_CP type are instead type 0
1217
1218# process_generic_property_file() can accept certain overrides in its input.
1219# Each of these must begin AND end with $CMD_DELIM.
1220my $CMD_DELIM = "\a";
1221my $REPLACE_CMD = 'replace'; # Override the Replace
1222my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1223
1224my $NO = 0;
1225my $YES = 1;
1226
1227# Values for the Replace argument to add_range.
1228# $NO # Don't replace; add only the code points not
1229 # already present.
1230my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1231 # the comments at the subroutine definition.
1232my $UNCONDITIONALLY = 2; # Replace without conditions.
9470941f 1233my $MULTIPLE_BEFORE = 4; # Don't replace, but add a duplicate record if
99870f4d 1234 # already there
7f4b1e25
KW
1235my $MULTIPLE_AFTER = 5; # Don't replace, but add a duplicate record if
1236 # already there
1237my $CROAK = 6; # Die with an error if is already there
99870f4d
KW
1238
1239# Flags to give property statuses. The phrases are to remind maintainers that
1240# if the flag is changed, the indefinite article referring to it in the
1241# documentation may need to be as well.
1242my $NORMAL = "";
99870f4d
KW
1243my $DEPRECATED = 'D';
1244my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1245my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1246my $DISCOURAGED = 'X';
1247my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1248my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1249my $STRICTER = 'T';
1250my $a_bold_stricter = "a 'B<$STRICTER>'";
1251my $A_bold_stricter = "A 'B<$STRICTER>'";
1252my $STABILIZED = 'S';
1253my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1254my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1255my $OBSOLETE = 'O';
1256my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1257my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1258
5c6d9947
KW
1259# Aliases can also have an extra status:
1260my $INTERNAL_ALIAS = 'P';
1261
99870f4d
KW
1262my %status_past_participles = (
1263 $DISCOURAGED => 'discouraged',
99870f4d
KW
1264 $STABILIZED => 'stabilized',
1265 $OBSOLETE => 'obsolete',
37e2e78e 1266 $DEPRECATED => 'deprecated',
5c6d9947 1267 $INTERNAL_ALIAS => 'reserved for Perl core internal use only',
99870f4d
KW
1268);
1269
395dfc19
KW
1270# Table fates. These are somewhat ordered, so that fates < $MAP_PROXIED should be
1271# externally documented.
301ba948 1272my $ORDINARY = 0; # The normal fate.
395dfc19
KW
1273my $MAP_PROXIED = 1; # The map table for the property isn't written out,
1274 # but there is a file written that can be used to
1275 # reconstruct this table
3cdaf629 1276my $INTERNAL_ONLY = 2; # The file for this table is written out, but it is
301ba948 1277 # for Perl's internal use only
277b7b16
KW
1278my $LEGACY_ONLY = 3; # Like $INTERNAL_ONLY, but not actually used by Perl.
1279 # Is for backwards compatibility for applications that
1280 # read the file directly, so it's format is
1281 # unchangeable.
1282my $SUPPRESSED = 4; # The file for this table is not written out, and as a
3cdaf629
KW
1283 # result, we don't bother to do many computations on
1284 # it.
277b7b16 1285my $PLACEHOLDER = 5; # Like $SUPPRESSED, but we go through all the
3cdaf629
KW
1286 # computations anyway, as the values are needed for
1287 # things to work. This happens when we have Perl
1288 # extensions that depend on Unicode tables that
1289 # wouldn't normally be in a given Unicode version.
301ba948 1290
f5817e0a
KW
1291# The format of the values of the tables:
1292my $EMPTY_FORMAT = "";
99870f4d
KW
1293my $BINARY_FORMAT = 'b';
1294my $DECIMAL_FORMAT = 'd';
1295my $FLOAT_FORMAT = 'f';
1296my $INTEGER_FORMAT = 'i';
1297my $HEX_FORMAT = 'x';
1298my $RATIONAL_FORMAT = 'r';
1299my $STRING_FORMAT = 's';
d11155ec 1300my $ADJUST_FORMAT = 'a';
24303724 1301my $HEX_ADJUST_FORMAT = 'ax';
a14f3cb1 1302my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1303my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1304
1305my %map_table_formats = (
1306 $BINARY_FORMAT => 'binary',
1307 $DECIMAL_FORMAT => 'single decimal digit',
1308 $FLOAT_FORMAT => 'floating point number',
1309 $INTEGER_FORMAT => 'integer',
add63c13 1310 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1311 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1312 $STRING_FORMAT => 'string',
d11155ec 1313 $ADJUST_FORMAT => 'some entries need adjustment',
24303724 1314 $HEX_ADJUST_FORMAT => 'mapped value in hex; some entries need adjustment',
92f9d56c 1315 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1316 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1317);
1318
1319# Unicode didn't put such derived files in a separate directory at first.
1320my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1321my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1322my $AUXILIARY = 'auxiliary';
1323
3854b4b8
KW
1324# Hashes and arrays that will eventually go into Heavy.pl for the use of
1325# utf8_heavy.pl and into UCD.pl for the use of UCD.pm
99870f4d
KW
1326my %loose_to_file_of; # loosely maps table names to their respective
1327 # files
1328my %stricter_to_file_of; # same; but for stricter mapping.
315bfd4e 1329my %loose_property_to_file_of; # Maps a loose property name to its map file
e1dc048d 1330my %strict_property_to_file_of; # Same, but strict
3854b4b8
KW
1331my @inline_definitions = "V0"; # Each element gives a definition of a unique
1332 # inversion list. When a definition is inlined,
1333 # its value in the hash it's in (one of the two
1334 # defined just above) will include an index into
1335 # this array. The 0th element is initialized to
9be22ca5 1336 # the definition for a zero length inversion list
89cf10cc
KW
1337my %file_to_swash_name; # Maps the file name to its corresponding key name
1338 # in the hash %utf8::SwashInfo
99870f4d
KW
1339my %nv_floating_to_rational; # maps numeric values floating point numbers to
1340 # their rational equivalent
c12f2655
KW
1341my %loose_property_name_of; # Loosely maps (non_string) property names to
1342 # standard form
e1dc048d
KW
1343my %strict_property_name_of; # Strictly maps (non_string) property names to
1344 # standard form
86a52d1e 1345my %string_property_loose_to_name; # Same, for string properties.
c15fda25
KW
1346my %loose_defaults; # keys are of form "prop=value", where 'prop' is
1347 # the property name in standard loose form, and
1348 # 'value' is the default value for that property,
1349 # also in standard loose form.
9e4a1e86
KW
1350my %loose_to_standard_value; # loosely maps table names to the canonical
1351 # alias for them
2df7880f
KW
1352my %ambiguous_names; # keys are alias names (in standard form) that
1353 # have more than one possible meaning.
d963b40d
KW
1354my %combination_property; # keys are alias names (in standard form) that
1355 # have both a map table, and a binary one that
1356 # yields true for all non-null maps.
5d1df013
KW
1357my %prop_aliases; # Keys are standard property name; values are each
1358 # one's aliases
1e863613
KW
1359my %prop_value_aliases; # Keys of top level are standard property name;
1360 # values are keys to another hash, Each one is
1361 # one of the property's values, in standard form.
1362 # The values are that prop-val's aliases.
b9e57660 1363my %skipped_files; # List of files that we skip
2df7880f 1364my %ucd_pod; # Holds entries that will go into the UCD section of the pod
99870f4d 1365
d867ccfb
KW
1366# Most properties are immune to caseless matching, otherwise you would get
1367# nonsensical results, as properties are a function of a code point, not
1368# everything that is caselessly equivalent to that code point. For example,
1369# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1370# be true because 's' and 'S' are equivalent caselessly. However,
1371# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1372# extend that concept to those very few properties that are like this. Each
1373# such property will match the full range caselessly. They are hard-coded in
1374# the program; it's not worth trying to make it general as it's extremely
1375# unlikely that they will ever change.
1376my %caseless_equivalent_to;
1377
53adf6a2
KW
1378# This is the range of characters that were in Release 1 of Unicode, and
1379# removed in Release 2 (replaced with the current Hangul syllables starting at
1380# U+AC00). The range was reused starting in Release 3 for other purposes.
1381my $FIRST_REMOVED_HANGUL_SYLLABLE = 0x3400;
1382my $FINAL_REMOVED_HANGUL_SYLLABLE = 0x4DFF;
1383
99870f4d
KW
1384# These constants names and values were taken from the Unicode standard,
1385# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1386# syllables. The '_string' versions are so generated tables can retain the
1387# hex format, which is the more familiar value
1388my $SBase_string = "0xAC00";
1389my $SBase = CORE::hex $SBase_string;
1390my $LBase_string = "0x1100";
1391my $LBase = CORE::hex $LBase_string;
1392my $VBase_string = "0x1161";
1393my $VBase = CORE::hex $VBase_string;
1394my $TBase_string = "0x11A7";
1395my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1396my $SCount = 11172;
1397my $LCount = 19;
1398my $VCount = 21;
1399my $TCount = 28;
1400my $NCount = $VCount * $TCount;
1401
1402# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1403# with the above published constants.
1404my %Jamo;
1405my %Jamo_L; # Leading consonants
1406my %Jamo_V; # Vowels
1407my %Jamo_T; # Trailing consonants
1408
bb1dd3da
KW
1409# For code points whose name contains its ordinal as a '-ABCD' suffix.
1410# The key is the base name of the code point, and the value is an
1411# array giving all the ranges that use this base name. Each range
1412# is actually a hash giving the 'low' and 'high' values of it.
1413my %names_ending_in_code_point;
1414my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1415 # removed from the names
1416# Inverse mapping. The list of ranges that have these kinds of
1417# names. Each element contains the low, high, and base names in an
1418# anonymous hash.
1419my @code_points_ending_in_code_point;
1420
6b5ab373
KW
1421# To hold Unicode's normalization test suite
1422my @normalization_tests;
1423
bb1dd3da
KW
1424# Boolean: does this Unicode version have the hangul syllables, and are we
1425# writing out a table for them?
1426my $has_hangul_syllables = 0;
1427
1428# Does this Unicode version have code points whose names end in their
1429# respective code points, and are we writing out a table for them? 0 for no;
1430# otherwise points to first property that a table is needed for them, so that
1431# if multiple tables are needed, we don't create duplicates
1432my $needing_code_points_ending_in_code_point = 0;
1433
37e2e78e 1434my @backslash_X_tests; # List of tests read in for testing \X
6b659339 1435my @LB_tests; # List of tests read in for testing \b{lb}
06ae2722 1436my @SB_tests; # List of tests read in for testing \b{sb}
ae3bb8ea 1437my @WB_tests; # List of tests read in for testing \b{wb}
99870f4d
KW
1438my @unhandled_properties; # Will contain a list of properties found in
1439 # the input that we didn't process.
f86864ac 1440my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1441 # listed in the pod
1442my @map_properties; # Properties that get map files written
1443my @named_sequences; # NamedSequences.txt contents.
1444my %potential_files; # Generated list of all .txt files in the directory
1445 # structure so we can warn if something is being
1446 # ignored.
a0204bb1
KW
1447my @missing_early_files; # Generated list of absent files that we need to
1448 # proceed in compiling this early Unicode version
99870f4d
KW
1449my @files_actually_output; # List of files we generated.
1450my @more_Names; # Some code point names are compound; this is used
1451 # to store the extra components of them.
0426f635
KW
1452my $E_FLOAT_PRECISION = 2; # The minimum number of digits after the decimal
1453 # point of a normalized floating point number
1454 # needed to match before we consider it equivalent
1455 # to a candidate rational
99870f4d
KW
1456
1457# These store references to certain commonly used property objects
1254636b 1458my $age;
3c88a801 1459my $ccc;
99870f4d
KW
1460my $gc;
1461my $perl;
1462my $block;
3e20195b
KW
1463my $perl_charname;
1464my $print;
ac7dbdb6 1465my $All;
bc0c431d 1466my $Assigned; # All assigned characters in this Unicode release
e47e66b9
KW
1467my $DI; # Default_Ignorable_Code_Point property
1468my $NChar; # Noncharacter_Code_Point property
359523e2 1469my $script;
8fb542fb 1470my $scx; # Script_Extensions property
99870f4d
KW
1471
1472# Are there conflicting names because of beginning with 'In_', or 'Is_'
1473my $has_In_conflicts = 0;
1474my $has_Is_conflicts = 0;
1475
1476sub internal_file_to_platform ($) {
1477 # Convert our file paths which have '/' separators to those of the
1478 # platform.
1479
1480 my $file = shift;
1481 return undef unless defined $file;
1482
1483 return File::Spec->join(split '/', $file);
d07a55ed 1484}
5beb625e 1485
99870f4d
KW
1486sub file_exists ($) { # platform independent '-e'. This program internally
1487 # uses slash as a path separator.
1488 my $file = shift;
1489 return 0 if ! defined $file;
1490 return -e internal_file_to_platform($file);
1491}
5beb625e 1492
99870f4d 1493sub objaddr($) {
23e33b60
KW
1494 # Returns the address of the blessed input object.
1495 # It doesn't check for blessedness because that would do a string eval
1496 # every call, and the program is structured so that this is never called
1497 # for a non-blessed object.
99870f4d 1498
23e33b60 1499 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1500
1501 # Numifying a ref gives its address.
051df77b 1502 return pack 'J', $_[0];
99870f4d
KW
1503}
1504
558712cf 1505# These are used only if $annotate is true.
c4019d52
KW
1506# The entire range of Unicode characters is examined to populate these
1507# after all the input has been processed. But most can be skipped, as they
1508# have the same descriptive phrases, such as being unassigned
1509my @viacode; # Contains the 1 million character names
4b06b8de 1510my @age; # And their ages ("" if none)
c4019d52
KW
1511my @printable; # boolean: And are those characters printable?
1512my @annotate_char_type; # Contains a type of those characters, specifically
1513 # for the purposes of annotation.
1514my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1515 # name for the purposes of annotation. They map to the
c4019d52
KW
1516 # upper edge of the range, so that the end point can
1517 # be immediately found. This is used to skip ahead to
1518 # the end of a range, and avoid processing each
1519 # individual code point in it.
1520my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1521 # characters, but excluding those which are
1522 # also noncharacter code points
1523
1524# The annotation types are an extension of the regular range types, though
1525# some of the latter are folded into one. Make the new types negative to
1526# avoid conflicting with the regular types
1527my $SURROGATE_TYPE = -1;
1528my $UNASSIGNED_TYPE = -2;
1529my $PRIVATE_USE_TYPE = -3;
1530my $NONCHARACTER_TYPE = -4;
1531my $CONTROL_TYPE = -5;
2d88a86a
KW
1532my $ABOVE_UNICODE_TYPE = -6;
1533my $UNKNOWN_TYPE = -7; # Used only if there is a bug in this program
c4019d52
KW
1534
1535sub populate_char_info ($) {
558712cf 1536 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1537 # input code point's info that are needed for outputting more detailed
1538 # comments. If calling context wants a return, it is the end point of
1539 # any contiguous range of characters that share essentially the same info
1540
1541 my $i = shift;
1542 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1543
1544 $viacode[$i] = $perl_charname->value_of($i) || "";
1254636b 1545 $age[$i] = (defined $age)
272af058 1546 ? (($age->value_of($i) =~ / ^ \d+ \. \d+ $ /x)
1254636b
KW
1547 ? $age->value_of($i)
1548 : "")
1549 : "";
c4019d52
KW
1550
1551 # A character is generally printable if Unicode says it is,
1552 # but below we make sure that most Unicode general category 'C' types
1553 # aren't.
1554 $printable[$i] = $print->contains($i);
1555
1254636b
KW
1556 # But the characters in this range were removed in v2.0 and replaced by
1557 # different ones later. Modern fonts will be for the replacement
1558 # characters, so suppress printing them.
1559 if (($v_version lt v2.0
1560 || ($compare_versions && $compare_versions lt v2.0))
1561 && ( $i >= $FIRST_REMOVED_HANGUL_SYLLABLE
1562 && $i <= $FINAL_REMOVED_HANGUL_SYLLABLE))
1563 {
1564 $printable[$i] = 0;
1565 }
1566
c4019d52
KW
1567 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1568
1569 # Only these two regular types are treated specially for annotations
1570 # purposes
1571 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1572 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1573
1574 # Give a generic name to all code points that don't have a real name.
1575 # We output ranges, if applicable, for these. Also calculate the end
1576 # point of the range.
1577 my $end;
1578 if (! $viacode[$i]) {
2d88a86a
KW
1579 if ($i > $MAX_UNICODE_CODEPOINT) {
1580 $viacode[$i] = 'Above-Unicode';
1581 $annotate_char_type[$i] = $ABOVE_UNICODE_TYPE;
1582 $printable[$i] = 0;
1583 $end = $MAX_WORKING_CODEPOINT;
1584 }
1585 elsif ($gc-> table('Private_use')->contains($i)) {
c4019d52
KW
1586 $viacode[$i] = 'Private Use';
1587 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1588 $printable[$i] = 0;
1589 $end = $gc->table('Private_Use')->containing_range($i)->end;
1590 }
e47e66b9 1591 elsif ($NChar->contains($i)) {
c4019d52
KW
1592 $viacode[$i] = 'Noncharacter';
1593 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1594 $printable[$i] = 0;
e47e66b9 1595 $end = $NChar->containing_range($i)->end;
c4019d52
KW
1596 }
1597 elsif ($gc-> table('Control')->contains($i)) {
1254636b
KW
1598 my $name_ref = property_ref('Name_Alias');
1599 $name_ref = property_ref('Unicode_1_Name') if ! defined $name_ref;
1600 $viacode[$i] = (defined $name_ref)
1601 ? $name_ref->value_of($i)
1602 : 'Control';
c4019d52
KW
1603 $annotate_char_type[$i] = $CONTROL_TYPE;
1604 $printable[$i] = 0;
c4019d52
KW
1605 }
1606 elsif ($gc-> table('Unassigned')->contains($i)) {
c4019d52
KW
1607 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1608 $printable[$i] = 0;
1254636b
KW
1609 $viacode[$i] = 'Unassigned';
1610
349edad4 1611 if (defined $block) { # No blocks in earliest releases
1254636b 1612 $viacode[$i] .= ', block=' . $block-> value_of($i);
1d025d66 1613 $end = $gc-> table('Unassigned')->containing_range($i)->end;
c4019d52 1614
bf06c733
KW
1615 # Because we name the unassigned by the blocks they are in, it
1616 # can't go past the end of that block, and it also can't go
1617 # past the unassigned range it is in. The special table makes
1618 # sure that the non-characters, which are unassigned, are
1619 # separated out.
1620 $end = min($block->containing_range($i)->end,
1621 $unassigned_sans_noncharacters->
1622 containing_range($i)->end);
1d025d66 1623 }
1254636b
KW
1624 else {
1625 $end = $i + 1;
1626 while ($unassigned_sans_noncharacters->contains($end)) {
1627 $end++;
1628 }
1629 $end--;
1630 }
1d025d66 1631 }
35e2b416 1632 elsif ($perl->table('_Perl_Surrogate')->contains($i)) {
1d025d66
KW
1633 $viacode[$i] = 'Surrogate';
1634 $annotate_char_type[$i] = $SURROGATE_TYPE;
1635 $printable[$i] = 0;
1636 $end = $gc->table('Surrogate')->containing_range($i)->end;
13ca76ff
KW
1637 }
1638 else {
1639 Carp::my_carp_bug("Can't figure out how to annotate "
1640 . sprintf("U+%04X", $i)
1641 . ". Proceeding anyway.");
c4019d52
KW
1642 $viacode[$i] = 'UNKNOWN';
1643 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1644 $printable[$i] = 0;
1645 }
1646 }
1647
1648 # Here, has a name, but if it's one in which the code point number is
1649 # appended to the name, do that.
1650 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1651 $viacode[$i] .= sprintf("-%04X", $i);
4b06b8de 1652
4b06b8de 1653 my $limit = $perl_charname->containing_range($i)->end;
1254636b
KW
1654 if (defined $age) {
1655 # Do all these as groups of the same age, instead of individually,
1656 # because their names are so meaningless, and there are typically
1657 # large quantities of them.
1658 $end = $i + 1;
1659 while ($end <= $limit && $age->value_of($end) == $age[$i]) {
1660 $end++;
1661 }
1662 $end--;
1663 }
1664 else {
1665 $end = $limit;
4b06b8de 1666 }
c4019d52
KW
1667 }
1668
1669 # And here, has a name, but if it's a hangul syllable one, replace it with
1670 # the correct name from the Unicode algorithm
1671 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1672 use integer;
1673 my $SIndex = $i - $SBase;
1674 my $L = $LBase + $SIndex / $NCount;
1675 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1676 my $T = $TBase + $SIndex % $TCount;
1677 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1678 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1679 $end = $perl_charname->containing_range($i)->end;
1680 }
1681
1682 return if ! defined wantarray;
1683 return $i if ! defined $end; # If not a range, return the input
1684
1685 # Save this whole range so can find the end point quickly
1686 $annotate_ranges->add_map($i, $end, $end);
1687
1688 return $end;
1689}
1690
23e33b60
KW
1691# Commented code below should work on Perl 5.8.
1692## This 'require' doesn't necessarily work in miniperl, and even if it does,
1693## the native perl version of it (which is what would operate under miniperl)
1694## is extremely slow, as it does a string eval every call.
7e017d6d 1695#my $has_fast_scalar_util = $^X !~ /miniperl/
23e33b60
KW
1696# && defined eval "require Scalar::Util";
1697#
1698#sub objaddr($) {
1699# # Returns the address of the blessed input object. Uses the XS version if
1700# # available. It doesn't check for blessedness because that would do a
1701# # string eval every call, and the program is structured so that this is
1702# # never called for a non-blessed object.
1703#
1704# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1705#
1706# # Check at least that is a ref.
1707# my $pkg = ref($_[0]) or return undef;
1708#
1709# # Change to a fake package to defeat any overloaded stringify
1710# bless $_[0], 'main::Fake';
1711#
1712# # Numifying a ref gives its address.
051df77b 1713# my $addr = pack 'J', $_[0];
23e33b60
KW
1714#
1715# # Return to original class
1716# bless $_[0], $pkg;
1717# return $addr;
1718#}
1719
99870f4d
KW
1720sub max ($$) {
1721 my $a = shift;
1722 my $b = shift;
1723 return $a if $a >= $b;
1724 return $b;
1725}
1726
1727sub min ($$) {
1728 my $a = shift;
1729 my $b = shift;
1730 return $a if $a <= $b;
1731 return $b;
1732}
1733
1734sub clarify_number ($) {
1735 # This returns the input number with underscores inserted every 3 digits
1736 # in large (5 digits or more) numbers. Input must be entirely digits, not
1737 # checked.
1738
1739 my $number = shift;
1740 my $pos = length($number) - 3;
1741 return $number if $pos <= 1;
1742 while ($pos > 0) {
1743 substr($number, $pos, 0) = '_';
1744 $pos -= 3;
5beb625e 1745 }
99870f4d 1746 return $number;
99598c8c
JH
1747}
1748
731cb813
KW
1749sub clarify_code_point_count ($) {
1750 # This is like clarify_number(), but the input is assumed to be a count of
1751 # code points, rather than a generic number.
1752
2d88a86a
KW
1753 my $append = "";
1754
1755 my $number = shift;
1756 if ($number > $MAX_UNICODE_CODEPOINTS) {
1757 $number -= ($MAX_WORKING_CODEPOINTS - $MAX_UNICODE_CODEPOINTS);
1758 return "All above-Unicode code points" if $number == 0;
1759 $append = " + all above-Unicode code points";
1760 }
1761 return clarify_number($number) . $append;
731cb813 1762}
12ac2576 1763
99870f4d 1764package Carp;
7ebf06b3 1765
99870f4d
KW
1766# These routines give a uniform treatment of messages in this program. They
1767# are placed in the Carp package to cause the stack trace to not include them,
1768# although an alternative would be to use another package and set @CARP_NOT
1769# for it.
12ac2576 1770
99870f4d 1771our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1772
99f78760
KW
1773# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1774# and overload trying to load Scalar:Util under miniperl. See
1775# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1776undef $overload::VERSION;
1777
99870f4d
KW
1778sub my_carp {
1779 my $message = shift || "";
1780 my $nofold = shift || 0;
7ebf06b3 1781
99870f4d
KW
1782 if ($message) {
1783 $message = main::join_lines($message);
1784 $message =~ s/^$0: *//; # Remove initial program name
1785 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1786 $message = "\n$0: $message;";
12ac2576 1787
99870f4d
KW
1788 # Fold the message with program name, semi-colon end punctuation
1789 # (which looks good with the message that carp appends to it), and a
1790 # hanging indent for continuation lines.
1791 $message = main::simple_fold($message, "", 4) unless $nofold;
1792 $message =~ s/\n$//; # Remove the trailing nl so what carp
1793 # appends is to the same line
1794 }
12ac2576 1795
99870f4d 1796 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1797
99870f4d
KW
1798 carp $message;
1799 return;
1800}
7ebf06b3 1801
99870f4d
KW
1802sub my_carp_bug {
1803 # This is called when it is clear that the problem is caused by a bug in
1804 # this program.
7ebf06b3 1805
99870f4d
KW
1806 my $message = shift;
1807 $message =~ s/^$0: *//;
1808 $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");
1809 carp $message;
1810 return;
1811}
7ebf06b3 1812
99870f4d
KW
1813sub carp_too_few_args {
1814 if (@_ != 2) {
1815 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1816 return;
12ac2576 1817 }
7ebf06b3 1818
99870f4d
KW
1819 my $args_ref = shift;
1820 my $count = shift;
7ebf06b3 1821
99870f4d
KW
1822 my_carp_bug("Need at least $count arguments to "
1823 . (caller 1)[3]
1824 . ". Instead got: '"
1825 . join ', ', @$args_ref
1826 . "'. No action taken.");
1827 return;
12ac2576
JP
1828}
1829
99870f4d
KW
1830sub carp_extra_args {
1831 my $args_ref = shift;
1832 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1833
99870f4d
KW
1834 unless (ref $args_ref) {
1835 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1836 return;
1837 }
1838 my ($package, $file, $line) = caller;
1839 my $subroutine = (caller 1)[3];
cf25bb62 1840
99870f4d
KW
1841 my $list;
1842 if (ref $args_ref eq 'HASH') {
1843 foreach my $key (keys %$args_ref) {
1844 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1845 }
99870f4d 1846 $list = join ', ', each %{$args_ref};
cf25bb62 1847 }
99870f4d
KW
1848 elsif (ref $args_ref eq 'ARRAY') {
1849 foreach my $arg (@$args_ref) {
1850 $arg = $UNDEF unless defined $arg;
1851 }
1852 $list = join ', ', @$args_ref;
1853 }
1854 else {
1855 my_carp_bug("Can't cope with ref "
1856 . ref($args_ref)
1857 . " . argument to 'carp_extra_args'. Not checking arguments.");
1858 return;
1859 }
1860
1861 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1862 return;
d73e5302
JH
1863}
1864
99870f4d
KW
1865package main;
1866
1867{ # Closure
1868
1869 # This program uses the inside-out method for objects, as recommended in
1b6b3fa9
KW
1870 # "Perl Best Practices". (This is the best solution still, since this has
1871 # to run under miniperl.) This closure aids in generating those. There
99870f4d
KW
1872 # are two routines. setup_package() is called once per package to set
1873 # things up, and then set_access() is called for each hash representing a
1874 # field in the object. These routines arrange for the object to be
1875 # properly destroyed when no longer used, and for standard accessor
1876 # functions to be generated. If you need more complex accessors, just
1877 # write your own and leave those accesses out of the call to set_access().
1878 # More details below.
1879
1880 my %constructor_fields; # fields that are to be used in constructors; see
1881 # below
1882
1883 # The values of this hash will be the package names as keys to other
1884 # hashes containing the name of each field in the package as keys, and
1885 # references to their respective hashes as values.
1886 my %package_fields;
1887
1888 sub setup_package {
1889 # Sets up the package, creating standard DESTROY and dump methods
1890 # (unless already defined). The dump method is used in debugging by
1891 # simple_dumper().
1892 # The optional parameters are:
1893 # a) a reference to a hash, that gets populated by later
1894 # set_access() calls with one of the accesses being
1895 # 'constructor'. The caller can then refer to this, but it is
1896 # not otherwise used by these two routines.
1897 # b) a reference to a callback routine to call during destruction
1898 # of the object, before any fields are actually destroyed
1899
1900 my %args = @_;
1901 my $constructor_ref = delete $args{'Constructor_Fields'};
1902 my $destroy_callback = delete $args{'Destroy_Callback'};
1903 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1904
1905 my %fields;
1906 my $package = (caller)[0];
1907
1908 $package_fields{$package} = \%fields;
1909 $constructor_fields{$package} = $constructor_ref;
1910
1911 unless ($package->can('DESTROY')) {
1912 my $destroy_name = "${package}::DESTROY";
1913 no strict "refs";
1914
1915 # Use typeglob to give the anonymous subroutine the name we want
1916 *$destroy_name = sub {
1917 my $self = shift;
ffe43484 1918 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1919
1920 $self->$destroy_callback if $destroy_callback;
1921 foreach my $field (keys %{$package_fields{$package}}) {
1922 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1923 delete $package_fields{$package}{$field}{$addr};
1924 }
1925 return;
1926 }
1927 }
1928
1929 unless ($package->can('dump')) {
1930 my $dump_name = "${package}::dump";
1931 no strict "refs";
1932 *$dump_name = sub {
1933 my $self = shift;
1934 return dump_inside_out($self, $package_fields{$package}, @_);
1935 }
1936 }
1937 return;
1938 }
1939
1940 sub set_access {
1941 # Arrange for the input field to be garbage collected when no longer
1942 # needed. Also, creates standard accessor functions for the field
1943 # based on the optional parameters-- none if none of these parameters:
1944 # 'addable' creates an 'add_NAME()' accessor function.
1945 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1946 # function.
1947 # 'settable' creates a 'set_NAME()' accessor function.
1948 # 'constructor' doesn't create an accessor function, but adds the
1949 # field to the hash that was previously passed to
1950 # setup_package();
1951 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1952 # 'add' etc. all mean 'addable'.
1953 # The read accessor function will work on both array and scalar
1954 # values. If another accessor in the parameter list is 'a', the read
1955 # access assumes an array. You can also force it to be array access
1956 # by specifying 'readable_array' instead of 'readable'
1957 #
1958 # A sort-of 'protected' access can be set-up by preceding the addable,
1959 # readable or settable with some initial portion of 'protected_' (but,
1960 # the underscore is required), like 'p_a', 'pro_set', etc. The
1961 # "protection" is only by convention. All that happens is that the
1962 # accessor functions' names begin with an underscore. So instead of
1963 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1964 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1965 # end of each package, and then storing the __LINE__ ranges and
1966 # checking them on every accessor. But that is way overkill.)
1967
1968 # We create anonymous subroutines as the accessors and then use
1969 # typeglobs to assign them to the proper package and name
1970
1971 my $name = shift; # Name of the field
1972 my $field = shift; # Reference to the inside-out hash containing the
1973 # field
1974
1975 my $package = (caller)[0];
1976
1977 if (! exists $package_fields{$package}) {
1978 croak "$0: Must call 'setup_package' before 'set_access'";
1979 }
d73e5302 1980
99870f4d
KW
1981 # Stash the field so DESTROY can get it.
1982 $package_fields{$package}{$name} = $field;
cf25bb62 1983
99870f4d
KW
1984 # Remaining arguments are the accessors. For each...
1985 foreach my $access (@_) {
1986 my $access = lc $access;
cf25bb62 1987
99870f4d 1988 my $protected = "";
cf25bb62 1989
99870f4d
KW
1990 # Match the input as far as it goes.
1991 if ($access =~ /^(p[^_]*)_/) {
1992 $protected = $1;
1993 if (substr('protected_', 0, length $protected)
1994 eq $protected)
1995 {
1996
1997 # Add 1 for the underscore not included in $protected
1998 $access = substr($access, length($protected) + 1);
1999 $protected = '_';
2000 }
2001 else {
2002 $protected = "";
2003 }
2004 }
2005
2006 if (substr('addable', 0, length $access) eq $access) {
2007 my $subname = "${package}::${protected}add_$name";
2008 no strict "refs";
2009
2010 # add_ accessor. Don't add if already there, which we
2011 # determine using 'eq' for scalars and '==' otherwise.
2012 *$subname = sub {
2013 use strict "refs";
2014 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
2015 my $self = shift;
2016 my $value = shift;
ffe43484 2017 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2018 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2019 if (ref $value) {
f998e60c 2020 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
2021 }
2022 else {
f998e60c 2023 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 2024 }
f998e60c 2025 push @{$field->{$addr}}, $value;
99870f4d
KW
2026 return;
2027 }
2028 }
2029 elsif (substr('constructor', 0, length $access) eq $access) {
2030 if ($protected) {
2031 Carp::my_carp_bug("Can't set-up 'protected' constructors")
2032 }
2033 else {
2034 $constructor_fields{$package}{$name} = $field;
2035 }
2036 }
2037 elsif (substr('readable_array', 0, length $access) eq $access) {
2038
2039 # Here has read access. If one of the other parameters for
2040 # access is array, or this one specifies array (by being more
2041 # than just 'readable_'), then create a subroutine that
2042 # assumes the data is an array. Otherwise just a scalar
2043 my $subname = "${package}::${protected}$name";
2044 if (grep { /^a/i } @_
2045 or length($access) > length('readable_'))
2046 {
2047 no strict "refs";
2048 *$subname = sub {
2049 use strict "refs";
23e33b60 2050 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 2051 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
2052 if (ref $field->{$addr} ne 'ARRAY') {
2053 my $type = ref $field->{$addr};
2054 $type = 'scalar' unless $type;
2055 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
2056 return;
2057 }
2058 return scalar @{$field->{$addr}} unless wantarray;
2059
2060 # Make a copy; had problems with caller modifying the
2061 # original otherwise
2062 my @return = @{$field->{$addr}};
2063 return @return;
2064 }
2065 }
2066 else {
2067
2068 # Here not an array value, a simpler function.
2069 no strict "refs";
2070 *$subname = sub {
2071 use strict "refs";
23e33b60 2072 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 2073 no overloading;
051df77b 2074 return $field->{pack 'J', $_[0]};
99870f4d
KW
2075 }
2076 }
2077 }
2078 elsif (substr('settable', 0, length $access) eq $access) {
2079 my $subname = "${package}::${protected}set_$name";
2080 no strict "refs";
2081 *$subname = sub {
2082 use strict "refs";
23e33b60
KW
2083 if (main::DEBUG) {
2084 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
2085 Carp::carp_extra_args(\@_) if @_ > 2;
2086 }
2087 # $self is $_[0]; $value is $_[1]
f998e60c 2088 no overloading;
051df77b 2089 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
2090 return;
2091 }
2092 }
2093 else {
2094 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
2095 }
cf25bb62 2096 }
99870f4d 2097 return;
cf25bb62 2098 }
99870f4d
KW
2099}
2100
2101package Input_file;
2102
2103# All input files use this object, which stores various attributes about them,
2104# and provides for convenient, uniform handling. The run method wraps the
2105# processing. It handles all the bookkeeping of opening, reading, and closing
2106# the file, returning only significant input lines.
2107#
2108# Each object gets a handler which processes the body of the file, and is
74cd47d0
KW
2109# called by run(). All character property files must use the generic,
2110# default handler, which has code scrubbed to handle things you might not
2111# expect, including automatic EBCDIC handling. For files that don't deal with
2112# mapping code points to a property value, such as test files,
2113# PropertyAliases, PropValueAliases, and named sequences, you can override the
2114# handler to be a custom one. Such a handler should basically be a
2115# while(next_line()) {...} loop.
99870f4d
KW
2116#
2117# You can also set up handlers to
40045c2d 2118# 0) call during object construction time, after everything else is done
537124e4 2119# 1) call before the first line is read, for pre processing
83b68635
KW
2120# 2) call to adjust each line of the input before the main handler gets
2121# them. This can be automatically generated, if appropriately simple
44b0f8c1 2122# enough, by specifying a Properties parameter in the constructor.
99870f4d 2123# 3) call upon EOF before the main handler exits its loop
537124e4 2124# 4) call at the end, for post processing
99870f4d
KW
2125#
2126# $_ is used to store the input line, and is to be filtered by the
2127# each_line_handler()s. So, if the format of the line is not in the desired
2128# format for the main handler, these are used to do that adjusting. They can
2129# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
44b0f8c1 2130# so the $_ output of one is used as the input to the next. The EOF handler
3848e5b7
KW
2131# is also stackable, but none of the others are, but could easily be changed
2132# to be so.
99870f4d 2133#
a0204bb1
KW
2134# Some properties are used by the Perl core but aren't defined until later
2135# Unicode releases. The perl interpreter would have problems working when
2136# compiled with an earlier Unicode version that doesn't have them, so we need
2137# to define them somehow for those releases. The 'Early' constructor
2138# parameter can be used to automatically handle this. It is essentially
2139# ignored if the Unicode version being compiled has a data file for this
2140# property. Either code to execute or a file to read can be specified.
2141# Details are at the %early definition.
2142#
99870f4d
KW
2143# Most of the handlers can call insert_lines() or insert_adjusted_lines()
2144# which insert the parameters as lines to be processed before the next input
3848e5b7 2145# file line is read. This allows the EOF handler(s) to flush buffers, for
99870f4d
KW
2146# example. The difference between the two routines is that the lines inserted
2147# by insert_lines() are subjected to the each_line_handler()s. (So if you
9be22ca5
KW
2148# called it from such a handler, you would get infinite recursion without some
2149# mechanism to prevent that.) Lines inserted by insert_adjusted_lines() go
2150# directly to the main handler without any adjustments. If the
2151# post-processing handler calls any of these, there will be no effect. Some
2152# error checking for these conditions could be added, but it hasn't been done.
99870f4d
KW
2153#
2154# carp_bad_line() should be called to warn of bad input lines, which clears $_
2155# to prevent further processing of the line. This routine will output the
2156# message as a warning once, and then keep a count of the lines that have the
2157# same message, and output that count at the end of the file's processing.
2158# This keeps the number of messages down to a manageable amount.
2159#
2160# get_missings() should be called to retrieve any @missing input lines.
2161# Messages will be raised if this isn't done if the options aren't to ignore
2162# missings.
2163
2164sub trace { return main::trace(@_); }
2165
99870f4d
KW
2166{ # Closure
2167 # Keep track of fields that are to be put into the constructor.
2168 my %constructor_fields;
2169
2170 main::setup_package(Constructor_Fields => \%constructor_fields);
2171
2172 my %file; # Input file name, required
2173 main::set_access('file', \%file, qw{ c r });
2174
2175 my %first_released; # Unicode version file was first released in, required
2176 main::set_access('first_released', \%first_released, qw{ c r });
2177
2178 my %handler; # Subroutine to process the input file, defaults to
2179 # 'process_generic_property_file'
2180 main::set_access('handler', \%handler, qw{ c });
2181
2182 my %property;
2183 # name of property this file is for. defaults to none, meaning not
2184 # applicable, or is otherwise determinable, for example, from each line.
696609bf 2185 main::set_access('property', \%property, qw{ c r });
99870f4d
KW
2186
2187 my %optional;
3274d896
KW
2188 # This is either an unsigned number, or a list of property names. In the
2189 # former case, if it is non-zero, it means the file is optional, so if the
2190 # file is absent, no warning about that is output. In the latter case, it
2191 # is a list of properties that the file (exclusively) defines. If the
2192 # file is present, tables for those properties will be produced; if
2193 # absent, none will, even if they are listed elsewhere (namely
2194 # PropertyAliases.txt and PropValueAliases.txt) as being in this release,
2195 # and no warnings will be raised about them not being available. (And no
2196 # warning about the file itself will be raised.)
2197 main::set_access('optional', \%optional, qw{ c readable_array } );
99870f4d
KW
2198
2199 my %non_skip;
2200 # This is used for debugging, to skip processing of all but a few input
2201 # files. Add 'non_skip => 1' to the constructor for those files you want
2202 # processed when you set the $debug_skip global.
2203 main::set_access('non_skip', \%non_skip, 'c');
2204
37e2e78e 2205 my %skip;
0dee21d6
KW
2206 # This is used to skip processing of this input file (semi-) permanently.
2207 # The value should be the reason the file is being skipped. It is used
2208 # for files that we aren't planning to process anytime soon, but want to
2209 # allow to be in the directory and be checked for their names not
2210 # conflicting with any other files on a DOS 8.3 name filesystem, but to
2211 # not otherwise be processed, and to not raise a warning about not being
2212 # handled. In the constructor call, any value that evaluates to a numeric
2213 # 0 or undef means don't skip. Any other value is a string giving the
44b0f8c1 2214 # reason it is being skipped, and this will appear in generated pod.
0dee21d6
KW
2215 # However, an empty string reason will suppress the pod entry.
2216 # Internally, calls that evaluate to numeric 0 are changed into undef to
2217 # distinguish them from an empty string call.
2218 main::set_access('skip', \%skip, 'c', 'r');
37e2e78e 2219
99870f4d
KW
2220 my %each_line_handler;
2221 # list of subroutines to look at and filter each non-comment line in the
2222 # file. defaults to none. The subroutines are called in order, each is
2223 # to adjust $_ for the next one, and the final one adjusts it for
2224 # 'handler'
2225 main::set_access('each_line_handler', \%each_line_handler, 'c');
2226
4e058f44
KW
2227 my %retain_trailing_comments;
2228 # This is used to not discard the comments that end data lines. This
2229 # would be used only for files with non-typical syntax, and most code here
2230 # assumes that comments have been stripped, so special handlers would have
2231 # to be written. It is assumed that the code will use these in
2232 # single-quoted contexts, and so any "'" marks in the comment will be
2233 # prefixed by a backslash.
2234 main::set_access('retain_trailing_comments', \%retain_trailing_comments, 'c');
2235
83b68635
KW
2236 my %properties; # Optional ordered list of the properties that occur in each
2237 # meaningful line of the input file. If present, an appropriate
2238 # each_line_handler() is automatically generated and pushed onto the stack
2239 # of such handlers. This is useful when a file contains multiple
44b0f8c1 2240 # properties per line, but no other special considerations are necessary.
83b68635
KW
2241 # The special value "<ignored>" means to discard the corresponding input
2242 # field.
2243 # Any @missing lines in the file should also match this syntax; no such
2244 # files exist as of 6.3. But if it happens in a future release, the code
2245 # could be expanded to properly parse them.
2246 main::set_access('properties', \%properties, qw{ c r });
2247
99870f4d
KW
2248 my %has_missings_defaults;
2249 # ? Are there lines in the file giving default values for code points
2250 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
2251 # the norm, but IGNORED means it has such lines, but the handler doesn't
2252 # use them. Having these three states allows us to catch changes to the
83b68635
KW
2253 # UCD that this program should track. XXX This could be expanded to
2254 # specify the syntax for such lines, like %properties above.
99870f4d
KW
2255 main::set_access('has_missings_defaults',
2256 \%has_missings_defaults, qw{ c r });
2257
40045c2d
KW
2258 my %construction_time_handler;
2259 # Subroutine to call at the end of the new method. If undef, no such
2260 # handler is called.
2261 main::set_access('construction_time_handler',
2262 \%construction_time_handler, qw{ c });
2263
99870f4d
KW
2264 my %pre_handler;
2265 # Subroutine to call before doing anything else in the file. If undef, no
2266 # such handler is called.
2267 main::set_access('pre_handler', \%pre_handler, qw{ c });
2268
2269 my %eof_handler;
3848e5b7 2270 # Subroutines to call upon getting an EOF on the input file, but before
99870f4d
KW
2271 # that is returned to the main handler. This is to allow buffers to be
2272 # flushed. The handler is expected to call insert_lines() or
2273 # insert_adjusted() with the buffered material
3848e5b7 2274 main::set_access('eof_handler', \%eof_handler, qw{ c });
99870f4d
KW
2275
2276 my %post_handler;
2277 # Subroutine to call after all the lines of the file are read in and
9be22ca5
KW
2278 # processed. If undef, no such handler is called. Note that this cannot
2279 # add lines to be processed; instead use eof_handler
99870f4d
KW
2280 main::set_access('post_handler', \%post_handler, qw{ c });
2281
2282 my %progress_message;
2283 # Message to print to display progress in lieu of the standard one
2284 main::set_access('progress_message', \%progress_message, qw{ c });
2285
2286 my %handle;
2287 # cache open file handle, internal. Is undef if file hasn't been
2288 # processed at all, empty if has;
2289 main::set_access('handle', \%handle);
2290
2291 my %added_lines;
2292 # cache of lines added virtually to the file, internal
2293 main::set_access('added_lines', \%added_lines);
2294
74cd47d0
KW
2295 my %remapped_lines;
2296 # cache of lines added virtually to the file, internal
2297 main::set_access('remapped_lines', \%remapped_lines);
2298
99870f4d
KW
2299 my %errors;
2300 # cache of errors found, internal
2301 main::set_access('errors', \%errors);
2302
2303 my %missings;
2304 # storage of '@missing' defaults lines
2305 main::set_access('missings', \%missings);
2306
a0204bb1
KW
2307 my %early;
2308 # Used for properties that must be defined (for Perl's purposes) on
2309 # versions of Unicode earlier than Unicode itself defines them. The
2310 # parameter is an array (it would be better to be a hash, but not worth
2311 # bothering about due to its rare use).
2312 #
2313 # The first element is either a code reference to call when in a release
2314 # earlier than the Unicode file is available in, or it is an alternate
2315 # file to use instead of the non-existent one. This file must have been
2316 # plunked down in the same directory as mktables. Should you be compiling
2317 # on a release that needs such a file, mktables will abort the
2318 # compilation, and tell you where to get the necessary file(s), and what
2319 # name(s) to use to store them as.
2320 # In the case of specifying an alternate file, the array must contain two
2321 # further elements:
2322 #
2323 # [1] is the name of the property that will be generated by this file.
2324 # The class automatically takes the input file and excludes any code
2325 # points in it that were not assigned in the Unicode version being
2326 # compiled. It then uses this result to define the property in the given
2327 # version. Since the property doesn't actually exist in the Unicode
2328 # version being compiled, this should be a name accessible only by core
2329 # perl. If it is the same name as the regular property, the constructor
2330 # will mark the output table as a $PLACEHOLDER so that it doesn't actually
2331 # get output, and so will be unusable by non-core code. Otherwise it gets
2332 # marked as $INTERNAL_ONLY.
2333 #
2334 # [2] is a property value to assign (only when compiling Unicode 1.1.5) to
2335 # the Hangul syllables in that release (which were ripped out in version
2336 # 2) for the given property . (Hence it is ignored except when compiling
2337 # version 1. You only get one value that applies to all of them, which
2338 # may not be the actual reality, but probably nobody cares anyway for
2339 # these obsolete characters.)
2340 #
3d36fe12
KW
2341 # [3] if present is the default value for the property to assign for code
2342 # points not given in the input. If not present, the default from the
2343 # normal property is used
2344 #
4ed2b786
KW
2345 # [-1] If there is an extra final element that is the string 'ONLY_EARLY'.
2346 # it means to not add the name in [1] as an alias to the property name
2347 # used for these. Normally, when compiling Unicode versions that don't
2348 # invoke the early handling, the name is added as a synonym.
cbdbe9d4 2349 #
a0204bb1
KW
2350 # Not all files can be handled in the above way, and so the code ref
2351 # alternative is available. It can do whatever it needs to. The other
2352 # array elements are optional in this case, and the code is free to use or
2353 # ignore them if they are present.
2354 #
2355 # Internally, the constructor unshifts a 0 or 1 onto this array to
2356 # indicate if an early alternative is actually being used or not. This
2357 # makes for easier testing later on.
2358 main::set_access('early', \%early, 'c');
2359
4ed2b786
KW
2360 my %only_early;
2361 main::set_access('only_early', \%only_early, 'c');
2362
0dee21d6
KW
2363 my %required_even_in_debug_skip;
2364 # debug_skip is used to speed up compilation during debugging by skipping
2365 # processing files that are not needed for the task at hand. However,
2366 # some files pretty much can never be skipped, and this is used to specify
2367 # that this is one of them. In order to skip this file, the call to the
2368 # constructor must be edited to comment out this parameter.
2369 main::set_access('required_even_in_debug_skip',
2370 \%required_even_in_debug_skip, 'c');
2371
d8062573
KW
2372 my %withdrawn;
2373 # Some files get removed from the Unicode DB. This is a version object
2374 # giving the first release without this file.
2375 main::set_access('withdrawn', \%withdrawn, 'c');
2376
0dee21d6
KW
2377 my %in_this_release;
2378 # Calculated value from %first_released and %withdrawn. Are we compiling
2379 # a Unicode release which includes this file?
2380 main::set_access('in_this_release', \%in_this_release);
2381
74cd47d0
KW
2382 sub _next_line;
2383 sub _next_line_with_remapped_range;
2384
99870f4d
KW
2385 sub new {
2386 my $class = shift;
2387
2388 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2389 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2390
2391 # Set defaults
2392 $handler{$addr} = \&main::process_generic_property_file;
4e058f44 2393 $retain_trailing_comments{$addr} = 0;
99870f4d 2394 $non_skip{$addr} = 0;
0dee21d6 2395 $skip{$addr} = undef;
99870f4d
KW
2396 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2397 $handle{$addr} = undef;
2398 $added_lines{$addr} = [ ];
74cd47d0 2399 $remapped_lines{$addr} = [ ];
99870f4d 2400 $each_line_handler{$addr} = [ ];
3848e5b7 2401 $eof_handler{$addr} = [ ];
99870f4d
KW
2402 $errors{$addr} = { };
2403 $missings{$addr} = [ ];
a0204bb1 2404 $early{$addr} = [ ];
3274d896 2405 $optional{$addr} = [ ];
99870f4d
KW
2406
2407 # Two positional parameters.
99f78760 2408 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2409 $file{$addr} = main::internal_file_to_platform(shift);
2410 $first_released{$addr} = shift;
2411
2412 # The rest of the arguments are key => value pairs
2413 # %constructor_fields has been set up earlier to list all possible
2414 # ones. Either set or push, depending on how the default has been set
2415 # up just above.
2416 my %args = @_;
2417 foreach my $key (keys %args) {
2418 my $argument = $args{$key};
2419
2420 # Note that the fields are the lower case of the constructor keys
2421 my $hash = $constructor_fields{lc $key};
2422 if (! defined $hash) {
2423 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2424 next;
2425 }
2426 if (ref $hash->{$addr} eq 'ARRAY') {
2427 if (ref $argument eq 'ARRAY') {
2428 foreach my $argument (@{$argument}) {
2429 next if ! defined $argument;
2430 push @{$hash->{$addr}}, $argument;
2431 }
2432 }
2433 else {
2434 push @{$hash->{$addr}}, $argument if defined $argument;
2435 }
2436 }
2437 else {
2438 $hash->{$addr} = $argument;
2439 }
2440 delete $args{$key};
2441 };
2442
0dee21d6
KW
2443 $non_skip{$addr} = 1 if $required_even_in_debug_skip{$addr};
2444
2445 # Convert 0 (meaning don't skip) to undef
2446 undef $skip{$addr} unless $skip{$addr};
2447
3274d896
KW
2448 # Handle the case where this file is optional
2449 my $pod_message_for_non_existent_optional = "";
2450 if ($optional{$addr}->@*) {
2451
2452 # First element is the pod message
2453 $pod_message_for_non_existent_optional
2454 = shift $optional{$addr}->@*;
2455 # Convert a 0 'Optional' argument to an empty list to make later
2456 # code more concise.
2457 if ( $optional{$addr}->@*
2458 && $optional{$addr}->@* == 1
2459 && $optional{$addr}[0] ne ""
2460 && $optional{$addr}[0] !~ /\D/
2461 && $optional{$addr}[0] == 0)
2462 {
2463 $optional{$addr} = [ ];
2464 }
2465 else { # But if the only element doesn't evaluate to 0, make sure
2466 # that this file is indeed considered optional below.
2467 unshift $optional{$addr}->@*, 1;
2468 }
2469 }
2470
0dee21d6 2471 my $progress;
a0204bb1
KW
2472 my $function_instead_of_file = 0;
2473
4ed2b786
KW
2474 if ($early{$addr}->@* && $early{$addr}[-1] eq 'ONLY_EARLY') {
2475 $only_early{$addr} = 1;
2476 pop $early{$addr}->@*;
cbdbe9d4
KW
2477 }
2478
a0204bb1
KW
2479 # If we are compiling a Unicode release earlier than the file became
2480 # available, the constructor may have supplied a substitute
2481 if ($first_released{$addr} gt $v_version && $early{$addr}->@*) {
2482
2483 # Yes, we have a substitute, that we will use; mark it so
2484 unshift $early{$addr}->@*, 1;
2485
2486 # See the definition of %early for what the array elements mean.
4ed2b786
KW
2487 # Note that we have just unshifted onto the array, so the numbers
2488 # below are +1 of those in the %early description.
a0204bb1
KW
2489 # If we have a property this defines, create a table and default
2490 # map for it now (at essentially compile time), so that it will be
2491 # available for the whole of run time. (We will want to add this
2492 # name as an alias when we are using the official property name;
2493 # but this must be deferred until run(), because at construction
2494 # time the official names have yet to be defined.)
2495 if ($early{$addr}[2]) {
2496 my $fate = ($property{$addr}
2497 && $property{$addr} eq $early{$addr}[2])
2498 ? $PLACEHOLDER
2499 : $INTERNAL_ONLY;
2500 my $prop_object = Property->new($early{$addr}[2],
2501 Fate => $fate,
2502 Perl_Extension => 1,
2503 );
2504
3d36fe12
KW
2505 # If not specified by the constructor, use the default mapping
2506 # for the regular property for this substitute one.
4ed2b786
KW
2507 if ($early{$addr}[4]) {
2508 $prop_object->set_default_map($early{$addr}[4]);
3d36fe12
KW
2509 }
2510 elsif ( defined $property{$addr}
2511 && defined $default_mapping{$property{$addr}})
a0204bb1
KW
2512 {
2513 $prop_object
2514 ->set_default_map($default_mapping{$property{$addr}});
2515 }
2516 }
2517
2518 if (ref $early{$addr}[1] eq 'CODE') {
2519 $function_instead_of_file = 1;
2520
2521 # If the first element of the array is a code ref, the others
2522 # are optional.
2523 $handler{$addr} = $early{$addr}[1];
2524 $property{$addr} = $early{$addr}[2]
2525 if defined $early{$addr}[2];
2526 $progress = "substitute $file{$addr}";
2527
2528 undef $file{$addr};
2529 }
2530 else { # Specifying a substitute file
2531
2532 if (! main::file_exists($early{$addr}[1])) {
2533
2534 # If we don't see the substitute file, generate an error
2535 # message giving the needed things, and add it to the list
2536 # of such to output before actual processing happens
2537 # (hence the user finds out all of them in one run).
ce8d64d9
KW
2538 # Instead of creating a general method for NameAliases,
2539 # hard-code it here, as there is unlikely to ever be a
2540 # second one which needs special handling.
2541 my $string_version = ($file{$addr} eq "NameAliases.txt")
2542 ? 'at least 6.1 (the later, the better)'
2543 : sprintf "%vd", $first_released{$addr};
a0204bb1
KW
2544 push @missing_early_files, <<END;
2545'$file{$addr}' version $string_version should be copied to '$early{$addr}[1]'.
2546END
2547 ;
2548 return;
2549 }
2550 $progress = $early{$addr}[1];
2551 $progress .= ", substituting for $file{$addr}" if $file{$addr};
2552 $file{$addr} = $early{$addr}[1];
2553 $property{$addr} = $early{$addr}[2];
0dee21d6 2554
a0204bb1
KW
2555 # Ignore code points not in the version being compiled
2556 push $each_line_handler{$addr}->@*, \&_exclude_unassigned;
2557
2558 if ( $v_version lt v2.0 # Hanguls in this release ...
4ed2b786 2559 && defined $early{$addr}[3]) # ... need special treatment
a0204bb1
KW
2560 {
2561 push $eof_handler{$addr}->@*, \&_fixup_obsolete_hanguls;
2562 }
2563 }
2564
2565 # And this substitute is valid for all releases.
2566 $first_released{$addr} = v0;
2567 }
2568 else { # Normal behavior
0dee21d6 2569 $progress = $file{$addr};
a0204bb1 2570 unshift $early{$addr}->@*, 0; # No substitute
99870f4d
KW
2571 }
2572
0dee21d6
KW
2573 my $file = $file{$addr};
2574 $progress_message{$addr} = "Processing $progress"
2575 unless $progress_message{$addr};
2576
d8062573
KW
2577 # A file should be there if it is within the window of versions for
2578 # which Unicode supplies it
2579 if ($withdrawn{$addr} && $withdrawn{$addr} le $v_version) {
2580 $in_this_release{$addr} = 0;
2581 $skip{$addr} = "";
2582 }
2583 else {
0dee21d6
KW
2584 $in_this_release{$addr} = $first_released{$addr} le $v_version;
2585
a0204bb1
KW
2586 # Check that the file for this object (possibly using a substitute
2587 # for early releases) exists or we have a function alternative
2588 if ( ! $function_instead_of_file
2589 && ! main::file_exists($file))
0dee21d6
KW
2590 {
2591 # Here there is nothing available for this release. This is
2592 # fine if we aren't expecting anything in this release.
2593 if (! $in_this_release{$addr}) {
2594 $skip{$addr} = ""; # Don't remark since we expected
2595 # nothing and got nothing
2596 }
3274d896 2597 elsif ($optional{$addr}->@*) {
0dee21d6 2598
3274d896
KW
2599 # Here the file is optional in this release; Use the
2600 # passed in text to document this case in the pod.
2601 $skip{$addr} = $pod_message_for_non_existent_optional;
0dee21d6
KW
2602 }
2603 elsif ( $in_this_release{$addr}
2604 && ! defined $skip{$addr}
2605 && defined $file)
2606 { # Doesn't exist but should.
2607 $skip{$addr} = "'$file' not found. Possibly Big problems";
2608 Carp::my_carp($skip{$addr});
2609 }
2610 }
2611 elsif ($debug_skip && ! defined $skip{$addr} && ! $non_skip{$addr})
2612 {
2613
2614 # The file exists; if not skipped for another reason, and we are
2615 # skipping most everything during debugging builds, use that as
2616 # the skip reason.
2617 $skip{$addr} = '$debug_skip is on'
2618 }
d8062573 2619 }
0dee21d6
KW
2620
2621 if ( ! $debug_skip
2622 && $non_skip{$addr}
2623 && ! $required_even_in_debug_skip{$addr}
2624 && $verbosity)
2625 {
2626 print "Warning: " . __PACKAGE__ . " constructor for $file has useless 'non_skip' in it\n";
a3a8c5f0 2627 }
99870f4d 2628
0dee21d6 2629 # Here, we have figured out if we will be skipping this file or not.
3274d896
KW
2630 # If so, we add any single property it defines to any passed in
2631 # optional property list. These will be dealt with at run time.
0dee21d6 2632 if (defined $skip{$addr}) {
3274d896
KW
2633 if ($property{$addr}) {
2634 push $optional{$addr}->@*, $property{$addr};
2635 }
2636 } # Otherwise, are going to process the file.
0dee21d6
KW
2637 elsif ($property{$addr}) {
2638
2639 # If the file has a property defined in the constructor for it, it
2640 # means that the property is not listed in the file's entries. So
2641 # add a handler (to the list of line handlers) to insert the
2642 # property name into the lines, to provide a uniform interface to
2643 # the final processing subroutine.
2644 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
09ca89ce 2645 }
83b68635
KW
2646 elsif ($properties{$addr}) {
2647
9be22ca5
KW
2648 # Similarly, there may be more than one property represented on
2649 # each line, with no clue but the constructor input what those
2650 # might be. Add a handler for each line in the input so that it
2651 # creates a separate input line for each property in those input
2652 # lines, thus making them suitable to handle generically.
83b68635
KW
2653
2654 push @{$each_line_handler{$addr}},
2655 sub {
2656 my $file = shift;
2657 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2658
2659 my @fields = split /\s*;\s*/, $_, -1;
2660
2661 if (@fields - 1 > @{$properties{$addr}}) {
2662 $file->carp_bad_line('Extra fields');
2663 $_ = "";
2664 return;
2665 }
2666 my $range = shift @fields; # 0th element is always the
2667 # range
2668
2669 # The next fields in the input line correspond
2670 # respectively to the stored properties.
2671 for my $i (0 .. @{$properties{$addr}} - 1) {
2672 my $property_name = $properties{$addr}[$i];
2673 next if $property_name eq '<ignored>';
2674 $file->insert_adjusted_lines(
2675 "$range; $property_name; $fields[$i]");
2676 }
2677 $_ = "";
2678
2679 return;
2680 };
2681 }
37e2e78e 2682
9be22ca5 2683 { # On non-ascii platforms, we use a special pre-handler
74cd47d0
KW
2684 no strict;
2685 no warnings 'once';
2686 *next_line = (main::NON_ASCII_PLATFORM)
2687 ? *_next_line_with_remapped_range
2688 : *_next_line;
2689 }
2690
40045c2d
KW
2691 &{$construction_time_handler{$addr}}($self)
2692 if $construction_time_handler{$addr};
2693
99870f4d 2694 return $self;
d73e5302
JH
2695 }
2696
cf25bb62 2697
99870f4d
KW
2698 use overload
2699 fallback => 0,
2700 qw("") => "_operator_stringify",
2701 "." => \&main::_operator_dot,
1285127e 2702 ".=" => \&main::_operator_dot_equal,
99870f4d 2703 ;
cf25bb62 2704
99870f4d
KW
2705 sub _operator_stringify {
2706 my $self = shift;
cf25bb62 2707
99870f4d 2708 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2709 }
d73e5302 2710
99870f4d
KW
2711 sub run {
2712 # Process the input object $self. This opens and closes the file and
2713 # calls all the handlers for it. Currently, this can only be called
3848e5b7 2714 # once per file, as it destroy's the EOF handlers
d73e5302 2715
0dee21d6 2716 # flag to make sure extracted files are processed early
f8b69004 2717 state $seen_non_extracted = 0;
0dee21d6 2718
99870f4d
KW
2719 my $self = shift;
2720 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2721
ffe43484 2722 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2723
99870f4d 2724 my $file = $file{$addr};
d73e5302 2725
0dee21d6
KW
2726 if (! $file) {
2727 $handle{$addr} = 'pretend_is_open';
99870f4d
KW
2728 }
2729 else {
f8b69004 2730 if ($seen_non_extracted) {
9be22ca5
KW
2731 if ($file =~ /$EXTRACTED/i) # Some platforms may change the
2732 # case of the file's name
2733 {
1675ea0d 2734 Carp::my_carp_bug(main::join_lines(<<END
99f78760 2735$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2736anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2737have subtle problems
2738END
2739 ));
2740 }
2741 }
2742 elsif ($EXTRACTED_DIR
0dee21d6
KW
2743
2744 # We only do this check for generic property files
2745 && $handler{$addr} == \&main::process_generic_property_file
2746
f8b69004 2747 && $file !~ /$EXTRACTED/i)
99870f4d
KW
2748 {
2749 # We don't set this (by the 'if' above) if we have no
2750 # extracted directory, so if running on an early version,
2751 # this test won't work. Not worth worrying about.
f8b69004 2752 $seen_non_extracted = 1;
99870f4d
KW
2753 }
2754
9be22ca5 2755 # Mark the file as having being processed, and warn if it
99870f4d
KW
2756 # isn't a file we are expecting. As we process the files,
2757 # they are deleted from the hash, so any that remain at the
2758 # end of the program are files that we didn't process.
517956bf 2759 my $fkey = File::Spec->rel2abs($file);
0dee21d6 2760 my $exists = delete $potential_files{lc($fkey)};
faf3cf6b 2761
0dee21d6
KW
2762 Carp::my_carp("Was not expecting '$file'.")
2763 if $exists && ! $in_this_release{$addr};
2764
a0204bb1
KW
2765 # If there is special handling for compiling Unicode releases
2766 # earlier than the first one in which Unicode defines this
2767 # property ...
2768 if ($early{$addr}->@* > 1) {
2769
2770 # Mark as processed any substitute file that would be used in
2771 # such a release
2772 $fkey = File::Spec->rel2abs($early{$addr}[1]);
2773 delete $potential_files{lc($fkey)};
2774
2775 # As commented in the constructor code, when using the
2776 # official property, we still have to allow the publicly
2777 # inaccessible early name so that the core code which uses it
2778 # will work regardless.
4ed2b786
KW
2779 if ( ! $only_early{$addr}
2780 && ! $early{$addr}[0]
2781 && $early{$addr}->@* > 2)
2782 {
a0204bb1 2783 my $early_property_name = $early{$addr}[2];
4ed2b786 2784 if ($property{$addr} ne $early_property_name) {
a0204bb1
KW
2785 main::property_ref($property{$addr})
2786 ->add_alias($early_property_name);
2787 }
2788 }
2789 }
2790
0dee21d6
KW
2791 # We may be skipping this file ...
2792 if (defined $skip{$addr}) {
2793
2794 # If the file isn't supposed to be in this release, there is
2795 # nothing to do
2796 if ($in_this_release{$addr}) {
2797
2798 # But otherwise, we may print a message
2799 if ($debug_skip) {
2800 print STDERR "Skipping input file '$file'",
2801 " because '$skip{$addr}'\n";
2802 }
2803
2804 # And add it to the list of skipped files, which is later
2805 # used to make the pod
2806 $skipped_files{$file} = $skip{$addr};
3274d896
KW
2807
2808 # The 'optional' list contains properties that are also to
2809 # be skipped along with the file. (There may also be
2810 # digits which are just placeholders to make sure it isn't
2811 # an empty list
2812 foreach my $property ($optional{$addr}->@*) {
2813 next unless $property =~ /\D/;
2814 my $prop_object = main::property_ref($property);
2815 next unless defined $prop_object;
2816 $prop_object->set_fate($SUPPRESSED, $skip{$addr});
2817 }
0dee21d6 2818 }
99870f4d 2819
37e2e78e
KW
2820 return;
2821 }
2822
9be22ca5
KW
2823 # Here, we are going to process the file. Open it, converting the
2824 # slashes used in this program into the proper form for the OS
99870f4d
KW
2825 my $file_handle;
2826 if (not open $file_handle, "<", $file) {
2827 Carp::my_carp("Can't open $file. Skipping: $!");
d3664eb9 2828 return;
99870f4d
KW
2829 }
2830 $handle{$addr} = $file_handle; # Cache the open file handle
9e65c3f4 2831
0dee21d6
KW
2832 # If possible, make sure that the file is the correct version.
2833 # (This data isn't available on early Unicode releases or in
2834 # UnicodeData.txt.) We don't do this check if we are using a
2835 # substitute file instead of the official one (though the code
2836 # could be extended to do so).
2837 if ($in_this_release{$addr}
a0204bb1 2838 && ! $early{$addr}[0]
0dee21d6
KW
2839 && lc($file) ne 'unicodedata.txt')
2840 {
96f226dc 2841 if ($file !~ /^Unihan/i) {
0dee21d6
KW
2842
2843 # The non-Unihan files started getting version numbers in
2844 # 3.2, but some files in 4.0 are unchanged from 3.2, and
2845 # marked as 3.2. 4.0.1 is the first version where there
2846 # are no files marked as being from less than 4.0, though
2847 # some are marked as 4.0. In versions after that, the
2848 # numbers are correct.
2849 if ($v_version ge v4.0.1) {
2850 $_ = <$file_handle>; # The version number is in the
2851 # very first line
2852 if ($_ !~ / - $string_version \. /x) {
2853 chomp;
2854 $_ =~ s/^#\s*//;
2855
2856 # 4.0.1 had some valid files that weren't updated.
2857 if (! ($v_version eq v4.0.1 && $_ =~ /4\.0\.0/)) {
2858 die Carp::my_carp("File '$file' is version "
2859 . "'$_'. It should be "
2860 . "version $string_version");
2861 }
2862 }
cafe9cf0 2863 }
96f226dc 2864 }
0dee21d6
KW
2865 elsif ($v_version ge v6.0.0) { # Unihan
2866
2867 # Unihan files didn't get accurate version numbers until
2868 # 6.0. The version is somewhere in the first comment
2869 # block
96f226dc
KW
2870 while (<$file_handle>) {
2871 if ($_ !~ /^#/) {
0dee21d6
KW
2872 Carp::my_carp_bug("Could not find the expected "
2873 . "version info in file '$file'");
96f226dc
KW
2874 last;
2875 }
2876 chomp;
2877 $_ =~ s/^#\s*//;
2878 next if $_ !~ / version: /x;
2879 last if $_ =~ /$string_version/;
0dee21d6
KW
2880 die Carp::my_carp("File '$file' is version "
2881 . "'$_'. It should be "
2882 . "version $string_version");
96f226dc
KW
2883 }
2884 }
9e65c3f4 2885 }
99870f4d
KW
2886 }
2887
0dee21d6 2888 print "$progress_message{$addr}\n" if $verbosity >= $PROGRESS;
99870f4d
KW
2889
2890 # Call any special handler for before the file.
2891 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2892
2893 # Then the main handler
2894 &{$handler{$addr}}($self);
2895
2896 # Then any special post-file handler.
2897 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2898
2899 # If any errors have been accumulated, output the counts (as the first
2900 # error message in each class was output when it was encountered).
2901 if ($errors{$addr}) {
2902 my $total = 0;
2903 my $types = 0;
2904 foreach my $error (keys %{$errors{$addr}}) {
2905 $total += $errors{$addr}->{$error};
2906 delete $errors{$addr}->{$error};
2907 $types++;
2908 }
2909 if ($total > 1) {
2910 my $message
2911 = "A total of $total lines had errors in $file. ";
2912
2913 $message .= ($types == 1)
2914 ? '(Only the first one was displayed.)'
2915 : '(Only the first of each type was displayed.)';
2916 Carp::my_carp($message);
2917 }
2918 }
2919
2920 if (@{$missings{$addr}}) {
2921 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2922 }
2923
2924 # If a real file handle, close it.
2925 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2926 ref $handle{$addr};
2927 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2928 # the file, as opposed to undef
2929 return;
2930 }
2931
74cd47d0 2932 sub _next_line {
99870f4d
KW
2933 # Sets $_ to be the next logical input line, if any. Returns non-zero
2934 # if such a line exists. 'logical' means that any lines that have
2935 # been added via insert_lines() will be returned in $_ before the file
2936 # is read again.
2937
2938 my $self = shift;
2939 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2940
ffe43484 2941 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2942
2943 # Here the file is open (or if the handle is not a ref, is an open
2944 # 'virtual' file). Get the next line; any inserted lines get priority
2945 # over the file itself.
2946 my $adjusted;
2947
2948 LINE:
2949 while (1) { # Loop until find non-comment, non-empty line
2950 #local $to_trace = 1 if main::DEBUG;
2951 my $inserted_ref = shift @{$added_lines{$addr}};
2952 if (defined $inserted_ref) {
2953 ($adjusted, $_) = @{$inserted_ref};
2954 trace $adjusted, $_ if main::DEBUG && $to_trace;
2955 return 1 if $adjusted;
2956 }
2957 else {
2958 last if ! ref $handle{$addr}; # Don't read unless is real file
2959 last if ! defined ($_ = readline $handle{$addr});
2960 }
2961 chomp;
2962 trace $_ if main::DEBUG && $to_trace;
2963
2964 # See if this line is the comment line that defines what property
2965 # value that code points that are not listed in the file should
2966 # have. The format or existence of these lines is not guaranteed
2967 # by Unicode since they are comments, but the documentation says
2968 # that this was added for machine-readability, so probably won't
2969 # change. This works starting in Unicode Version 5.0. They look
2970 # like:
2971 #
2972 # @missing: 0000..10FFFF; Not_Reordered
2973 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2974 # @missing: 0000..10FFFF; ; NaN
2975 #
2976 # Save the line for a later get_missings() call.
2977 if (/$missing_defaults_prefix/) {
2978 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2979 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2980 }
2981 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2982 my @defaults = split /\s* ; \s*/x, $_;
2983
2984 # The first field is the @missing, which ends in a
2985 # semi-colon, so can safely shift.
2986 shift @defaults;
2987
2988 # Some of these lines may have empty field placeholders
2989 # which get in the way. An example is:
2990 # @missing: 0000..10FFFF; ; NaN
2991 # Remove them. Process starting from the top so the
2992 # splice doesn't affect things still to be looked at.
2993 for (my $i = @defaults - 1; $i >= 0; $i--) {
2994 next if $defaults[$i] ne "";
2995 splice @defaults, $i, 1;
2996 }
2997
2998 # What's left should be just the property (maybe) and the
2999 # default. Having only one element means it doesn't have
3000 # the property.
3001 my $default;
3002 my $property;
3003 if (@defaults >= 1) {
3004 if (@defaults == 1) {
3005 $default = $defaults[0];
3006 }
3007 else {
3008 $property = $defaults[0];
3009 $default = $defaults[1];
3010 }
3011 }
3012
3013 if (@defaults < 1
3014 || @defaults > 2
3015 || ($default =~ /^</
3016 && $default !~ /^<code *point>$/i
09f8d0ac
KW
3017 && $default !~ /^<none>$/i
3018 && $default !~ /^<script>$/i))
99870f4d
KW
3019 {
3020 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
3021 }
3022 else {
3023
3024 # If the property is missing from the line, it should
3025 # be the one for the whole file
3026 $property = $property{$addr} if ! defined $property;
3027
3028 # Change <none> to the null string, which is what it
3029 # really means. If the default is the code point
3030 # itself, set it to <code point>, which is what
3031 # Unicode uses (but sometimes they've forgotten the
3032 # space)
3033 if ($default =~ /^<none>$/i) {
3034 $default = "";
3035 }
3036 elsif ($default =~ /^<code *point>$/i) {
3037 $default = $CODE_POINT;
3038 }
09f8d0ac
KW
3039 elsif ($default =~ /^<script>$/i) {
3040
3041 # Special case this one. Currently is from
3042 # ScriptExtensions.txt, and means for all unlisted
3043 # code points, use their Script property values.
3044 # For the code points not listed in that file, the
3045 # default value is 'Unknown'.
3046 $default = "Unknown";
3047 }
99870f4d
KW
3048
3049 # Store them as a sub-arrays with both components.
3050 push @{$missings{$addr}}, [ $default, $property ];
3051 }
3052 }
3053
3054 # There is nothing for the caller to process on this comment
3055 # line.
3056 next;
3057 }
3058
4e058f44
KW
3059 # Unless to keep, remove comments. If to keep, ignore
3060 # comment-only lines
3061 if ($retain_trailing_comments{$addr}) {
3062 next if / ^ \s* \# /x;
3063
3064 # But escape any single quotes (done in both the comment and
3065 # non-comment portion; this could be a bug someday, but not
3066 # likely)
3067 s/'/\\'/g;
3068 }
3069 else {
3070 s/#.*//;
3071 }
3072
3073 # Remove trailing space, and skip this line if the result is empty
99870f4d
KW
3074 s/\s+$//;
3075 next if /^$/;
3076
3077 # Call any handlers for this line, and skip further processing of
3078 # the line if the handler sets the line to null.
3079 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
3080 &{$sub_ref}($self);
3081 next LINE if /^$/;
3082 }
3083
3084 # Here the line is ok. return success.
3085 return 1;
3086 } # End of looping through lines.
3087
3848e5b7 3088 # If there are EOF handlers, call each (only once) and if it generates
99870f4d 3089 # more lines to process go back in the loop to handle them.
3848e5b7
KW
3090 while ($eof_handler{$addr}->@*) {
3091 &{$eof_handler{$addr}[0]}($self);
3092 shift $eof_handler{$addr}->@*; # Currently only get one shot at it.
99870f4d
KW
3093 goto LINE if $added_lines{$addr};
3094 }
3095
3096 # Return failure -- no more lines.
3097 return 0;
3098
3099 }
3100
74cd47d0
KW
3101 sub _next_line_with_remapped_range {
3102 my $self = shift;
3103 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3104
3105 # like _next_line(), but for use on non-ASCII platforms. It sets $_
3106 # to be the next logical input line, if any. Returns non-zero if such
3107 # a line exists. 'logical' means that any lines that have been added
3108 # via insert_lines() will be returned in $_ before the file is read
3109 # again.
3110 #
3111 # The difference from _next_line() is that this remaps the Unicode
3112 # code points in the input to those of the native platform. Each
3113 # input line contains a single code point, or a single contiguous
3114 # range of them This routine splits each range into its individual
3115 # code points and caches them. It returns the cached values,
3116 # translated into their native equivalents, one at a time, for each
3117 # call, before reading the next line. Since native values can only be
3118 # a single byte wide, no translation is needed for code points above
3119 # 0xFF, and ranges that are entirely above that number are not split.
3120 # If an input line contains the range 254-1000, it would be split into
3121 # three elements: 254, 255, and 256-1000. (The downstream table
3122 # insertion code will sort and coalesce the individual code points
3123 # into appropriate ranges.)
3124
3125 my $addr = do { no overloading; pack 'J', $self; };
3126
3127 while (1) {
3128
3129 # Look in cache before reading the next line. Return any cached
3130 # value, translated
3131 my $inserted = shift @{$remapped_lines{$addr}};
3132 if (defined $inserted) {
3133 trace $inserted if main::DEBUG && $to_trace;
3134 $_ = $inserted =~ s/^ ( \d+ ) /sprintf("%04X", utf8::unicode_to_native($1))/xer;
3135 trace $_ if main::DEBUG && $to_trace;
3136 return 1;
3137 }
3138
3139 # Get the next line.
3140 return 0 unless _next_line($self);
3141
3142 # If there is a special handler for it, return the line,
3143 # untranslated. This should happen only for files that are
3144 # special, not being code-point related, such as property names.
3145 return 1 if $handler{$addr}
3146 != \&main::process_generic_property_file;
3147
3148 my ($range, $property_name, $map, @remainder)
3149 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3150
3151 if (@remainder
3152 || ! defined $property_name
3153 || $range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3154 {
3155 Carp::my_carp_bug("Unrecognized input line '$_'. Ignored");
3156 }
3157
3158 my $low = hex $1;
3159 my $high = (defined $2) ? hex $2 : $low;
3160
3161 # If the input maps the range to another code point, remap the
3162 # target if it is between 0 and 255.
3163 my $tail;
3164 if (defined $map) {
3165 $map =~ s/\b 00 ( [0-9A-F]{2} ) \b/sprintf("%04X", utf8::unicode_to_native(hex $1))/gxe;
3166 $tail = "$property_name; $map";
3167 $_ = "$range; $tail";
3168 }
3169 else {
3170 $tail = $property_name;
3171 }
3172
3173 # If entire range is above 255, just return it, unchanged (except
3174 # any mapped-to code point, already changed above)
3175 return 1 if $low > 255;
3176
3177 # Cache an entry for every code point < 255. For those in the
3178 # range above 255, return a dummy entry for just that portion of
3179 # the range. Note that this will be out-of-order, but that is not
3180 # a problem.
3181 foreach my $code_point ($low .. $high) {
3182 if ($code_point > 255) {
3183 $_ = sprintf "%04X..%04X; $tail", $code_point, $high;
3184 return 1;
3185 }
3186 push @{$remapped_lines{$addr}}, "$code_point; $tail";
3187 }
3188 } # End of looping through lines.
3189
3190 # NOTREACHED
3191 }
3192
99870f4d
KW
3193# Not currently used, not fully tested.
3194# sub peek {
f67a5002 3195# # Non-destructive lookahead one non-adjusted, non-comment, non-blank
99870f4d
KW
3196# # record. Not callable from an each_line_handler(), nor does it call
3197# # an each_line_handler() on the line.
3198#
3199# my $self = shift;
ffe43484 3200# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3201#
3202# foreach my $inserted_ref (@{$added_lines{$addr}}) {
3203# my ($adjusted, $line) = @{$inserted_ref};
3204# next if $adjusted;
3205#
3206# # Remove comments and trailing space, and return a non-empty
3207# # resulting line
3208# $line =~ s/#.*//;
3209# $line =~ s/\s+$//;
3210# return $line if $line ne "";
3211# }
3212#
3213# return if ! ref $handle{$addr}; # Don't read unless is real file
3214# while (1) { # Loop until find non-comment, non-empty line
3215# local $to_trace = 1 if main::DEBUG;
3216# trace $_ if main::DEBUG && $to_trace;
3217# return if ! defined (my $line = readline $handle{$addr});
3218# chomp $line;
3219# push @{$added_lines{$addr}}, [ 0, $line ];
3220#
3221# $line =~ s/#.*//;
3222# $line =~ s/\s+$//;
3223# return $line if $line ne "";
3224# }
3225#
3226# return;
3227# }
3228
3229
3230 sub insert_lines {
3231 # Lines can be inserted so that it looks like they were in the input
3232 # file at the place it was when this routine is called. See also
3233 # insert_adjusted_lines(). Lines inserted via this routine go through
3234 # any each_line_handler()
3235
3236 my $self = shift;
3237
3238 # Each inserted line is an array, with the first element being 0 to
3239 # indicate that this line hasn't been adjusted, and needs to be
3240 # processed.
f998e60c 3241 no overloading;
051df77b 3242 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
3243 return;
3244 }
3245
3246 sub insert_adjusted_lines {
3247 # Lines can be inserted so that it looks like they were in the input
3248 # file at the place it was when this routine is called. See also
3249 # insert_lines(). Lines inserted via this routine are already fully
3250 # adjusted, ready to be processed; each_line_handler()s handlers will
3251 # not be called. This means this is not a completely general
3252 # facility, as only the last each_line_handler on the stack should
3253 # call this. It could be made more general, by passing to each of the
3254 # line_handlers their position on the stack, which they would pass on
3255 # to this routine, and that would replace the boolean first element in
3256 # the anonymous array pushed here, so that the next_line routine could
3257 # use that to call only those handlers whose index is after it on the
3258 # stack. But this is overkill for what is needed now.
3259
3260 my $self = shift;
3261 trace $_[0] if main::DEBUG && $to_trace;
3262
3263 # Each inserted line is an array, with the first element being 1 to
3264 # indicate that this line has been adjusted
f998e60c 3265 no overloading;
051df77b 3266 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
3267 return;
3268 }
3269
3270 sub get_missings {
3271 # Returns the stored up @missings lines' values, and clears the list.
3272 # The values are in an array, consisting of the default in the first
3273 # element, and the property in the 2nd. However, since these lines
3274 # can be stacked up, the return is an array of all these arrays.
3275
3276 my $self = shift;
3277 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3278
ffe43484 3279 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3280
3281 # If not accepting a list return, just return the first one.
3282 return shift @{$missings{$addr}} unless wantarray;
3283
3284 my @return = @{$missings{$addr}};
3285 undef @{$missings{$addr}};
3286 return @return;
3287 }
3288
a0204bb1
KW
3289 sub _exclude_unassigned {
3290
3291 # Takes the range in $_ and excludes code points that aren't assigned
3292 # in this release
3293
3294 state $skip_inserted_count = 0;
3295
3296 # Ignore recursive calls.
3297 if ($skip_inserted_count) {
3298 $skip_inserted_count--;
3299 return;
3300 }
3301
3302 # Find what code points are assigned in this release
3303 main::calculate_Assigned() if ! defined $Assigned;
3304
3305 my $self = shift;
3306 my $addr = do { no overloading; pack 'J', $self; };
3307 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3308
3309 my ($range, @remainder)
3310 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3311
3312 # Examine the range.
3313 if ($range =~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3314 {
3315 my $low = hex $1;
3316 my $high = (defined $2) ? hex $2 : $low;
3317
3318 # Split the range into subranges of just those code points in it
3319 # that are assigned.
3320 my @ranges = (Range_List->new(Initialize
3321 => Range->new($low, $high)) & $Assigned)->ranges;
3322
3323 # Do nothing if nothing in the original range is assigned in this
3324 # release; handle normally if everything is in this release.
3325 if (! @ranges) {
3326 $_ = "";
3327 }
3328 elsif (@ranges != 1) {
3329
3330 # Here, some code points in the original range aren't in this
3331 # release; @ranges gives the ones that are. Create fake input
3332 # lines for each of the ranges, and set things up so that when
3333 # this routine is called on that fake input, it will do
3334 # nothing.
3335 $skip_inserted_count = @ranges;
3336 my $remainder = join ";", @remainder;
3337 for my $range (@ranges) {
3338 $self->insert_lines(sprintf("%04X..%04X;%s",
3339 $range->start, $range->end, $remainder));
3340 }
3341 $_ = ""; # The original range is now defunct.
3342 }
3343 }
3344
3345 return;
3346 }
3347
3348 sub _fixup_obsolete_hanguls {
3349
3350 # This is called only when compiling Unicode version 1. All Unicode
3351 # data for subsequent releases assumes that the code points that were
3352 # Hangul syllables in this release only are something else, so if
3353 # using such data, we have to override it
3354
3355 my $self = shift;
3356 my $addr = do { no overloading; pack 'J', $self; };
3357 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3358
3359 my $object = main::property_ref($property{$addr});
53adf6a2
KW
3360 $object->add_map($FIRST_REMOVED_HANGUL_SYLLABLE,
3361 $FINAL_REMOVED_HANGUL_SYLLABLE,
a0204bb1
KW
3362 $early{$addr}[3], # Passed-in value for these
3363 Replace => $UNCONDITIONALLY);
3364 }
3365
99870f4d
KW
3366 sub _insert_property_into_line {
3367 # Add a property field to $_, if this file requires it.
3368
f998e60c 3369 my $self = shift;
ffe43484 3370 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 3371 my $property = $property{$addr};
99870f4d
KW
3372 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3373
3374 $_ =~ s/(;|$)/; $property$1/;
3375 return;
3376 }
3377
3378 sub carp_bad_line {
3379 # Output consistent error messages, using either a generic one, or the
3380 # one given by the optional parameter. To avoid gazillions of the
3381 # same message in case the syntax of a file is way off, this routine
3382 # only outputs the first instance of each message, incrementing a
3383 # count so the totals can be output at the end of the file.
3384
3385 my $self = shift;
3386 my $message = shift;
3387 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3388
ffe43484 3389 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3390
3391 $message = 'Unexpected line' unless $message;
3392
3393 # No trailing punctuation so as to fit with our addenda.
3394 $message =~ s/[.:;,]$//;
3395
3396 # If haven't seen this exact message before, output it now. Otherwise
3397 # increment the count of how many times it has occurred
3398 unless ($errors{$addr}->{$message}) {
3399 Carp::my_carp("$message in '$_' in "
f998e60c 3400 . $file{$addr}
99870f4d
KW
3401 . " at line $.. Skipping this line;");
3402 $errors{$addr}->{$message} = 1;
3403 }
3404 else {
3405 $errors{$addr}->{$message}++;
3406 }
3407
3408 # Clear the line to prevent any further (meaningful) processing of it.
3409 $_ = "";
3410
3411 return;
3412 }
3413} # End closure
3414
3415package Multi_Default;
3416
3417# Certain properties in early versions of Unicode had more than one possible
3418# default for code points missing from the files. In these cases, one
3419# default applies to everything left over after all the others are applied,
3420# and for each of the others, there is a description of which class of code
3421# points applies to it. This object helps implement this by storing the
3422# defaults, and for all but that final default, an eval string that generates
3423# the class that it applies to.
3424
3425
3426{ # Closure
3427
3428 main::setup_package();
3429
3430 my %class_defaults;
3431 # The defaults structure for the classes
3432 main::set_access('class_defaults', \%class_defaults);
3433
3434 my %other_default;
3435 # The default that applies to everything left over.
3436 main::set_access('other_default', \%other_default, 'r');
3437
3438
3439 sub new {
3440 # The constructor is called with default => eval pairs, terminated by
3441 # the left-over default. e.g.
3442 # Multi_Default->new(
3443 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
3444 # - 0x200D',
3445 # 'R' => 'some other expression that evaluates to code points',
3446 # .
3447 # .
3448 # .
3449 # 'U'));
68c3923f
KW
3450 # It is best to leave the final value be the one that matches the
3451 # above-Unicode code points.
99870f4d
KW
3452
3453 my $class = shift;
3454
3455 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 3456 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3457
3458 while (@_ > 1) {
3459 my $default = shift;
3460 my $eval = shift;
3461 $class_defaults{$addr}->{$default} = $eval;
3462 }
3463
3464 $other_default{$addr} = shift;
3465
3466 return $self;
3467 }
3468
3469 sub get_next_defaults {
3470 # Iterates and returns the next class of defaults.
3471 my $self = shift;
3472 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3473
ffe43484 3474 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3475
3476 return each %{$class_defaults{$addr}};
3477 }
3478}
3479
3480package Alias;
3481
3482# An alias is one of the names that a table goes by. This class defines them
3483# including some attributes. Everything is currently setup in the
3484# constructor.
3485
3486
3487{ # Closure
3488
3489 main::setup_package();
3490
3491 my %name;
3492 main::set_access('name', \%name, 'r');
3493
3494 my %loose_match;
c12f2655 3495 # Should this name match loosely or not.
99870f4d
KW
3496 main::set_access('loose_match', \%loose_match, 'r');
3497
33e96e72
KW
3498 my %make_re_pod_entry;
3499 # Some aliases should not get their own entries in the re section of the
3500 # pod, because they are covered by a wild-card, and some we want to
3501 # discourage use of. Binary
f82fe4ba 3502 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r', 's');
99870f4d 3503
fd1e3e84
KW
3504 my %ucd;
3505 # Is this documented to be accessible via Unicode::UCD
3506 main::set_access('ucd', \%ucd, 'r', 's');
3507
99870f4d
KW
3508 my %status;
3509 # Aliases have a status, like deprecated, or even suppressed (which means
3510 # they don't appear in documentation). Enum
3511 main::set_access('status', \%status, 'r');
3512
0eac1e20 3513 my %ok_as_filename;
99870f4d
KW
3514 # Similarly, some aliases should not be considered as usable ones for
3515 # external use, such as file names, or we don't want documentation to
3516 # recommend them. Boolean
0eac1e20 3517 main::set_access('ok_as_filename', \%ok_as_filename, 'r');
99870f4d
KW
3518
3519 sub new {
3520 my $class = shift;
3521
3522 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3523 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3524
3525 $name{$addr} = shift;
3526 $loose_match{$addr} = shift;
33e96e72 3527 $make_re_pod_entry{$addr} = shift;
0eac1e20 3528 $ok_as_filename{$addr} = shift;
99870f4d 3529 $status{$addr} = shift;
fd1e3e84 3530 $ucd{$addr} = shift;
99870f4d
KW
3531
3532 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3533
3534 # Null names are never ok externally
0eac1e20 3535 $ok_as_filename{$addr} = 0 if $name{$addr} eq "";
99870f4d
KW
3536
3537 return $self;
3538 }
3539}
3540
3541package Range;
3542
3543# A range is the basic unit for storing code points, and is described in the
3544# comments at the beginning of the program. Each range has a starting code
3545# point; an ending code point (not less than the starting one); a value
3546# that applies to every code point in between the two end-points, inclusive;
3547# and an enum type that applies to the value. The type is for the user's
3548# convenience, and has no meaning here, except that a non-zero type is
3549# considered to not obey the normal Unicode rules for having standard forms.
3550#
3551# The same structure is used for both map and match tables, even though in the
3552# latter, the value (and hence type) is irrelevant and could be used as a
3553# comment. In map tables, the value is what all the code points in the range
3554# map to. Type 0 values have the standardized version of the value stored as
3555# well, so as to not have to recalculate it a lot.
3556
3557sub trace { return main::trace(@_); }
3558
3559{ # Closure
3560
3561 main::setup_package();
3562
3563 my %start;
3564 main::set_access('start', \%start, 'r', 's');
3565
3566 my %end;
3567 main::set_access('end', \%end, 'r', 's');
3568
3569 my %value;
eeeafd9c 3570 main::set_access('value', \%value, 'r', 's');
99870f4d
KW
3571
3572 my %type;
3573 main::set_access('type', \%type, 'r');
3574
3575 my %standard_form;
3576 # The value in internal standard form. Defined only if the type is 0.
3577 main::set_access('standard_form', \%standard_form);
3578
3579 # Note that if these fields change, the dump() method should as well
3580
3581 sub new {
3582 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
3583 my $class = shift;
3584
3585 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3586 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3587
3588 $start{$addr} = shift;
3589 $end{$addr} = shift;
3590
3591 my %args = @_;
3592
3593 my $value = delete $args{'Value'}; # Can be 0
3594 $value = "" unless defined $value;
3595 $value{$addr} = $value;
3596
3597 $type{$addr} = delete $args{'Type'} || 0;
3598
3599 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3600
99870f4d
KW
3601 return $self;
3602 }
3603
3604 use overload
3605 fallback => 0,
3606 qw("") => "_operator_stringify",
3607 "." => \&main::_operator_dot,
1285127e 3608 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3609 ;
3610
3611 sub _operator_stringify {
3612 my $self = shift;
ffe43484 3613 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3614
3615 # Output it like '0041..0065 (value)'
3616 my $return = sprintf("%04X", $start{$addr})
3617 . '..'
3618 . sprintf("%04X", $end{$addr});
3619 my $value = $value{$addr};
3620 my $type = $type{$addr};
3621 $return .= ' (';
3622 $return .= "$value";
3623 $return .= ", Type=$type" if $type != 0;
3624 $return .= ')';
3625
3626 return $return;
3627 }
3628
3629 sub standard_form {
c292d35a
NC
3630 # Calculate the standard form only if needed, and cache the result.
3631 # The standard form is the value itself if the type is special.
3632 # This represents a considerable CPU and memory saving - at the time
3633 # of writing there are 368676 non-special objects, but the standard
3634 # form is only requested for 22047 of them - ie about 6%.
99870f4d
KW
3635
3636 my $self = shift;
3637 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3638
ffe43484 3639 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3640
3641 return $standard_form{$addr} if defined $standard_form{$addr};
c292d35a
NC
3642
3643 my $value = $value{$addr};
3644 return $value if $type{$addr};
3645 return $standard_form{$addr} = main::standardize($value);
99870f4d
KW
3646 }
3647
3648 sub dump {
3649 # Human, not machine readable. For machine readable, comment out this
3650 # entire routine and let the standard one take effect.
3651 my $self = shift;
3652 my $indent = shift;
3653 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3654
ffe43484 3655 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3656
3657 my $return = $indent
3658 . sprintf("%04X", $start{$addr})
3659 . '..'
3660 . sprintf("%04X", $end{$addr})
3661 . " '$value{$addr}';";
3662 if (! defined $standard_form{$addr}) {
3663 $return .= "(type=$type{$addr})";
3664 }
3665 elsif ($standard_form{$addr} ne $value{$addr}) {
3666 $return .= "(standard '$standard_form{$addr}')";
3667 }