This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
mktables: Add guard against Unicode breakage
[perl5.git] / lib / unicore / mktables
CommitLineData
d73e5302 1#!/usr/bin/perl -w
99870f4d
KW
2
3# !!!!!!!!!!!!!! IF YOU MODIFY THIS FILE !!!!!!!!!!!!!!!!!!!!!!!!!
4# Any files created or read by this program should be listed in 'mktables.lst'
5# Use -makelist to regenerate it.
6
9be22ca5
KW
7# There was an attempt when this was first rewritten to make it 5.8
8# compatible, but that has now been abandoned, and newer constructs are used
9# as convenient.
6c68572b 10
80f41172
JH
11# NOTE: this script can run quite slowly in older/slower systems.
12# It can also consume a lot of memory (128 MB or more), you may need
13# to raise your process resource limits (e.g. in bash, "ulimit -a"
14# to inspect, and "ulimit -d ..." or "ulimit -m ..." to set)
15
cdcef19a 16my $start_time;
98dc9551 17BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
18 # get it as close as possible
19 $start_time= time;
20}
21
23e33b60 22require 5.010_001;
d73e5302 23use strict;
99870f4d 24use warnings;
cf25bb62 25use Carp;
bd9ebcfd 26use Config;
99870f4d
KW
27use File::Find;
28use File::Path;
d07a55ed 29use File::Spec;
99870f4d 30use Text::Tabs;
6b64c11c 31use re "/aa";
0dee21d6 32use feature 'state';
99870f4d
KW
33
34sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 35my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d 36
74cd47d0
KW
37sub NON_ASCII_PLATFORM { ord("A") != 65 }
38
6295dc14
KW
39# When a new version of Unicode is published, unfortunately the algorithms for
40# dealing with various bounds, like \b{gcb}, \b{lb} may have to be updated
41# manually. The changes may or may not be backward compatible with older
42# releases. The code is in regen/mk_invlist.pl and regexec.c. Make the
43# changes, then come back here and set the variable below to what version the
44# code is expecting. If a newer version of Unicode is being compiled than
45# expected, a warning will be generated. If an older version is being
46# compiled, any bounds tests that fail in the generated test file (-maketest
47# option) will be marked as TODO.
c492f156 48my $version_of_mk_invlist_bounds = v10.0.0;
6295dc14 49
99870f4d
KW
50##########################################################################
51#
52# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
53# from the Unicode database files (lib/unicore/.../*.txt), It also generates
232ed87f 54# a pod file and .t files, depending on option parameters.
99870f4d
KW
55#
56# The structure of this file is:
57# First these introductory comments; then
58# code needed for everywhere, such as debugging stuff; then
59# code to handle input parameters; then
60# data structures likely to be of external interest (some of which depend on
61# the input parameters, so follows them; then
62# more data structures and subroutine and package (class) definitions; then
63# the small actual loop to process the input files and finish up; then
64# a __DATA__ section, for the .t tests
65#
232ed87f
KW
66# This program works on all releases of Unicode so far. The outputs have been
67# scrutinized most intently for release 5.1. The others have been checked for
68# somewhat more than just sanity. It can handle all non-provisional Unicode
69# character properties in those releases.
99870f4d 70#
99870f4d
KW
71# This program is mostly about Unicode character (or code point) properties.
72# A property describes some attribute or quality of a code point, like if it
73# is lowercase or not, its name, what version of Unicode it was first defined
74# in, or what its uppercase equivalent is. Unicode deals with these disparate
75# possibilities by making all properties into mappings from each code point
76# into some corresponding value. In the case of it being lowercase or not,
77# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
78# property maps each Unicode code point to a single value, called a "property
232ed87f
KW
79# value". (Some more recently defined properties, map a code point to a set
80# of values.)
99870f4d
KW
81#
82# When using a property in a regular expression, what is desired isn't the
83# mapping of the code point to its property's value, but the reverse (or the
84# mathematical "inverse relation"): starting with the property value, "Does a
85# code point map to it?" These are written in a "compound" form:
86# \p{property=value}, e.g., \p{category=punctuation}. This program generates
87# files containing the lists of code points that map to each such regular
88# expression property value, one file per list
89#
90# There is also a single form shortcut that Perl adds for many of the commonly
91# used properties. This happens for all binary properties, plus script,
92# general_category, and block properties.
93#
94# Thus the outputs of this program are files. There are map files, mostly in
95# the 'To' directory; and there are list files for use in regular expression
96# matching, all in subdirectories of the 'lib' directory, with each
97# subdirectory being named for the property that the lists in it are for.
98# Bookkeeping, test, and documentation files are also generated.
99
100my $matches_directory = 'lib'; # Where match (\p{}) files go.
101my $map_directory = 'To'; # Where map files go.
102
103# DATA STRUCTURES
104#
105# The major data structures of this program are Property, of course, but also
106# Table. There are two kinds of tables, very similar to each other.
107# "Match_Table" is the data structure giving the list of code points that have
108# a particular property value, mentioned above. There is also a "Map_Table"
109# data structure which gives the property's mapping from code point to value.
110# There are two structures because the match tables need to be combined in
111# various ways, such as constructing unions, intersections, complements, etc.,
112# and the map ones don't. And there would be problems, perhaps subtle, if
113# a map table were inadvertently operated on in some of those ways.
114# The use of separate classes with operations defined on one but not the other
115# prevents accidentally confusing the two.
116#
117# At the heart of each table's data structure is a "Range_List", which is just
118# an ordered list of "Ranges", plus ancillary information, and methods to
119# operate on them. A Range is a compact way to store property information.
120# Each range has a starting code point, an ending code point, and a value that
121# is meant to apply to all the code points between the two end points,
122# inclusive. For a map table, this value is the property value for those
123# code points. Two such ranges could be written like this:
124# 0x41 .. 0x5A, 'Upper',
125# 0x61 .. 0x7A, 'Lower'
126#
127# Each range also has a type used as a convenience to classify the values.
128# Most ranges in this program will be Type 0, or normal, but there are some
129# ranges that have a non-zero type. These are used only in map tables, and
130# are for mappings that don't fit into the normal scheme of things. Mappings
131# that require a hash entry to communicate with utf8.c are one example;
132# another example is mappings for charnames.pm to use which indicate a name
232ed87f 133# that is algorithmically determinable from its code point (and the reverse).
99870f4d
KW
134# These are used to significantly compact these tables, instead of listing
135# each one of the tens of thousands individually.
136#
137# In a match table, the value of a range is irrelevant (and hence the type as
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
8f21889c
KW
655# 0 in the $string_compare_versions definition just below to a string
656# containing a SINGLE dotted Unicode release number (e.g. "2.1") corresponding
657# to the starting release. This program will then compile, but throw away all
658# code points introduced after the starting release. Finally use a diff tool
659# to compare the two directory structures. They include only the code points
660# common to both releases, and you can see the changes caused just by the
661# underlying release semantic changes. For versions earlier than 3.2, you
662# must copy a version of DAge.txt into the directory.
272af058 663my $string_compare_versions = DEBUG && ""; # e.g., "2.1";
99870f4d
KW
664my $compare_versions = DEBUG
665 && $string_compare_versions
666 && pack "C*", split /\./, $string_compare_versions;
667
668sub uniques {
669 # Returns non-duplicated input values. From "Perl Best Practices:
670 # Encapsulated Cleverness". p. 455 in first edition.
671
672 my %seen;
0e407844
NC
673 # Arguably this breaks encapsulation, if the goal is to permit multiple
674 # distinct objects to stringify to the same value, and be interchangeable.
675 # However, for this program, no two objects stringify identically, and all
676 # lists passed to this function are either objects or strings. So this
677 # doesn't affect correctness, but it does give a couple of percent speedup.
678 no overloading;
99870f4d
KW
679 return grep { ! $seen{$_}++ } @_;
680}
681
682$0 = File::Spec->canonpath($0);
683
684my $make_test_script = 0; # ? Should we output a test script
6b5ab373 685my $make_norm_test_script = 0; # ? Should we output a normalization test script
99870f4d
KW
686my $write_unchanged_files = 0; # ? Should we update the output files even if
687 # we don't think they have changed
688my $use_directory = ""; # ? Should we chdir somewhere.
689my $pod_directory; # input directory to store the pod file.
690my $pod_file = 'perluniprops';
691my $t_path; # Path to the .t test file
692my $file_list = 'mktables.lst'; # File to store input and output file names.
693 # This is used to speed up the build, by not
694 # executing the main body of the program if
695 # nothing on the list has changed since the
696 # previous build
697my $make_list = 1; # ? Should we write $file_list. Set to always
698 # make a list so that when the pumpking is
699 # preparing a release, s/he won't have to do
700 # special things
701my $glob_list = 0; # ? Should we try to include unknown .txt files
702 # in the input.
bd9ebcfd
KW
703my $output_range_counts = $debugging_build; # ? Should we include the number
704 # of code points in ranges in
705 # the output
558712cf 706my $annotate = 0; # ? Should character names be in the output
9ef2b94f 707
99870f4d
KW
708# Verbosity levels; 0 is quiet
709my $NORMAL_VERBOSITY = 1;
710my $PROGRESS = 2;
711my $VERBOSE = 3;
712
713my $verbosity = $NORMAL_VERBOSITY;
714
0458fbc1
KW
715# Stored in mktables.lst so that if this program is called with different
716# options, will regenerate even if the files otherwise look like they're
717# up-to-date.
718my $command_line_arguments = join " ", @ARGV;
719
99870f4d
KW
720# Process arguments
721while (@ARGV) {
cf25bb62
JH
722 my $arg = shift @ARGV;
723 if ($arg eq '-v') {
99870f4d
KW
724 $verbosity = $VERBOSE;
725 }
726 elsif ($arg eq '-p') {
727 $verbosity = $PROGRESS;
728 $| = 1; # Flush buffers as we go.
729 }
730 elsif ($arg eq '-q') {
731 $verbosity = 0;
732 }
733 elsif ($arg eq '-w') {
44b0f8c1
SF
734 # update the files even if they haven't changed
735 $write_unchanged_files = 1;
99870f4d
KW
736 }
737 elsif ($arg eq '-check') {
6ae7e459
YO
738 my $this = shift @ARGV;
739 my $ok = shift @ARGV;
740 if ($this ne $ok) {
741 print "Skipping as check params are not the same.\n";
742 exit(0);
743 }
00a8df5c 744 }
99870f4d
KW
745 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
746 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
747 }
3df51b85
KW
748 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
749 {
99870f4d 750 $make_test_script = 1;
99870f4d 751 }
6b5ab373
KW
752 elsif ($arg eq '-makenormtest')
753 {
754 $make_norm_test_script = 1;
755 }
99870f4d
KW
756 elsif ($arg eq '-makelist') {
757 $make_list = 1;
758 }
759 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
760 -d $use_directory or croak "Unknown directory '$use_directory'";
761 }
762 elsif ($arg eq '-L') {
763
764 # Existence not tested until have chdir'd
765 $file_list = shift;
766 }
767 elsif ($arg eq '-globlist') {
768 $glob_list = 1;
769 }
770 elsif ($arg eq '-c') {
771 $output_range_counts = ! $output_range_counts
772 }
b4a0206c 773 elsif ($arg eq '-annotate') {
558712cf 774 $annotate = 1;
bd9ebcfd
KW
775 $debugging_build = 1;
776 $output_range_counts = 1;
9ef2b94f 777 }
99870f4d
KW
778 else {
779 my $with_c = 'with';
780 $with_c .= 'out' if $output_range_counts; # Complements the state
781 croak <<END;
782usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
783 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
784 [-check A B ]
785 -c : Output comments $with_c number of code points in ranges
786 -q : Quiet Mode: Only output serious warnings.
787 -p : Set verbosity level to normal plus show progress.
788 -v : Set Verbosity level high: Show progress and non-serious
789 warnings
790 -w : Write files regardless
791 -C dir : Change to this directory before proceeding. All relative paths
792 except those specified by the -P and -T options will be done
793 with respect to this directory.
794 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 795 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
796 -L filelist : Use alternate 'filelist' instead of standard one
797 -globlist : Take as input all non-Test *.txt files in current and sub
798 directories
3df51b85
KW
799 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
800 overrides -T
99870f4d 801 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 802 -annotate : Output an annotation for each character in the table files;
97a8a595
KW
803 useful for debugging mktables, looking at diffs; but is slow
804 and memory intensive
99870f4d
KW
805 -check A B : Executes $0 only if A and B are the same
806END
807 }
808}
809
810# Stores the most-recently changed file. If none have changed, can skip the
811# build
aeab6150 812my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
813
814# Change directories now, because need to read 'version' early.
815if ($use_directory) {
3df51b85 816 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
817 $pod_directory = File::Spec->rel2abs($pod_directory);
818 }
3df51b85 819 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 820 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 821 }
99870f4d 822 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 823 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 824 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 825 }
3df51b85 826 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 827 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 828 }
00a8df5c
YO
829}
830
99870f4d
KW
831# Get Unicode version into regular and v-string. This is done now because
832# various tables below get populated based on it. These tables are populated
833# here to be near the top of the file, and so easily seeable by those needing
834# to modify things.
835open my $VERSION, "<", "version"
836 or croak "$0: can't open required file 'version': $!\n";
837my $string_version = <$VERSION>;
838close $VERSION;
839chomp $string_version;
840my $v_version = pack "C*", split /\./, $string_version; # v string
841
8f21889c
KW
842my $unicode_version = ($compare_versions)
843 ? ( "$string_compare_versions (using "
844 . "$string_version rules)")
845 : $string_version;
846
99870f4d
KW
847# The following are the complete names of properties with property values that
848# are known to not match any code points in some versions of Unicode, but that
849# may change in the future so they should be matchable, hence an empty file is
850# generated for them.
a9c9e371
KW
851my @tables_that_may_be_empty;
852push @tables_that_may_be_empty, 'Joining_Type=Left_Joining'
853 if $v_version lt v6.3.0;
99870f4d
KW
854push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
855push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
856push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
857 if $v_version ge v4.1.0;
82aed44a
KW
858push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
859 if $v_version ge v6.0.0;
f583b44c
KW
860push @tables_that_may_be_empty, 'Grapheme_Cluster_Break=Prepend'
861 if $v_version ge v6.1.0;
1e958ea9
KW
862push @tables_that_may_be_empty, 'Canonical_Combining_Class=CCC133'
863 if $v_version ge v6.2.0;
99870f4d
KW
864
865# The lists below are hashes, so the key is the item in the list, and the
866# value is the reason why it is in the list. This makes generation of
867# documentation easier.
868
869my %why_suppressed; # No file generated for these.
870
871# Files aren't generated for empty extraneous properties. This is arguable.
872# Extraneous properties generally come about because a property is no longer
873# used in a newer version of Unicode. If we generated a file without code
874# points, programs that used to work on that property will still execute
875# without errors. It just won't ever match (or will always match, with \P{}).
876# This means that the logic is now likely wrong. I (khw) think its better to
877# find this out by getting an error message. Just move them to the table
878# above to change this behavior
879my %why_suppress_if_empty_warn_if_not = (
880
881 # It is the only property that has ever officially been removed from the
882 # Standard. The database never contained any code points for it.
883 'Special_Case_Condition' => 'Obsolete',
884
885 # Apparently never official, but there were code points in some versions of
886 # old-style PropList.txt
887 'Non_Break' => 'Obsolete',
888);
889
890# These would normally go in the warn table just above, but they were changed
891# a long time before this program was written, so warnings about them are
892# moot.
893if ($v_version gt v3.2.0) {
894 push @tables_that_may_be_empty,
895 'Canonical_Combining_Class=Attached_Below_Left'
896}
897
9be22ca5
KW
898# Enum values for to_output_map() method in the Map_Table package. (0 is don't
899# output)
272501f6
KW
900my $EXTERNAL_MAP = 1;
901my $INTERNAL_MAP = 2;
ce712c88 902my $OUTPUT_ADJUSTED = 3;
272501f6 903
fcf1973c
KW
904# To override computed values for writing the map tables for these properties.
905# The default for enum map tables is to write them out, so that the Unicode
906# .txt files can be removed, but all the data to compute any property value
907# for any code point is available in a more compact form.
908my %global_to_output_map = (
909 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
910 # get stuck supporting it if things change. Since it is a STRING
911 # property, it normally would be listed in the pod, but INTERNAL_MAP
912 # suppresses that.
fcf1973c
KW
913 Unicode_1_Name => $INTERNAL_MAP,
914
915 Present_In => 0, # Suppress, as easily computed from Age
74cd47d0
KW
916 Block => (NON_ASCII_PLATFORM) ? 1 : 0, # Suppress, as Blocks.txt is
917 # retained, but needed for
918 # non-ASCII
53d34b6c
KW
919
920 # Suppress, as mapping can be found instead from the
921 # Perl_Decomposition_Mapping file
922 Decomposition_Type => 0,
fcf1973c
KW
923);
924
99870f4d
KW
925# There are several types of obsolete properties defined by Unicode. These
926# must be hand-edited for every new Unicode release.
927my %why_deprecated; # Generates a deprecated warning message if used.
928my %why_stabilized; # Documentation only
929my %why_obsolete; # Documentation only
930
931{ # Closure
8364f9eb 932 my $simple = 'Perl uses the more complete version';
99870f4d
KW
933 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
934
935 my $other_properties = 'other properties';
936 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 937 my $why_no_expand = "Deprecated by Unicode. These are characters that expand to more than one character in the specified normalization form, but whether they actually take up more bytes or not depends on the encoding being used. For example, a UTF-8 encoded character may expand to a different number of bytes than a UTF-32 encoded character.";
99870f4d
KW
938
939 %why_deprecated = (
25c180b8 940 'Grapheme_Link' => 'Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d 941 'Jamo_Short_Name' => $contributory,
25c180b8 942 'Line_Break=Surrogate' => 'Surrogates should never appear in well-formed text, and therefore shouldn\'t be the basis for line breaking',
99870f4d
KW
943 'Other_Alphabetic' => $contributory,
944 'Other_Default_Ignorable_Code_Point' => $contributory,
945 'Other_Grapheme_Extend' => $contributory,
946 'Other_ID_Continue' => $contributory,
947 'Other_ID_Start' => $contributory,
948 'Other_Lowercase' => $contributory,
949 'Other_Math' => $contributory,
950 'Other_Uppercase' => $contributory,
e22aaf5c
KW
951 'Expands_On_NFC' => $why_no_expand,
952 'Expands_On_NFD' => $why_no_expand,
953 'Expands_On_NFKC' => $why_no_expand,
954 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
955 );
956
957 %why_suppressed = (
5f7264c7 958 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
959 # contains the same information, but without the algorithmically
960 # determinable Hangul syllables'. This file is not published, so it's
961 # existence is not noted in the comment.
12fee290 962 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize or prop_invmap() or charprop() in Unicode::UCD::',
99870f4d 963
5f8d1a89
KW
964 # Don't suppress ISO_Comment, as otherwise special handling is needed
965 # to differentiate between it and gc=c, which can be written as 'isc',
966 # which is the same characters as ISO_Comment's short name.
99870f4d 967
12fee290 968 'Name' => "Accessible via \\N{...} or 'use charnames;' or charprop() or prop_invmap() in Unicode::UCD::",
e0b29447 969
12fee290
KW
970 'Simple_Case_Folding' => "$simple. Can access this through casefold(), charprop(), or prop_invmap() in Unicode::UCD",
971 'Simple_Lowercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
972 'Simple_Titlecase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
973 'Simple_Uppercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
99870f4d 974
dac6f618 975 FC_NFKC_Closure => 'Deprecated by Unicode, and supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
976 );
977
1704a0ea
KW
978 foreach my $property (
979
980 # The following are suppressed because they were made contributory
981 # or deprecated by Unicode before Perl ever thought about
982 # supporting them.
983 'Jamo_Short_Name',
984 'Grapheme_Link',
985 'Expands_On_NFC',
986 'Expands_On_NFD',
987 'Expands_On_NFKC',
988 'Expands_On_NFKD',
989
990 # The following are suppressed because they have been marked
991 # as deprecated for a sufficient amount of time
992 'Other_Alphabetic',
993 'Other_Default_Ignorable_Code_Point',
994 'Other_Grapheme_Extend',
995 'Other_ID_Continue',
996 'Other_ID_Start',
997 'Other_Lowercase',
998 'Other_Math',
999 'Other_Uppercase',
e22aaf5c 1000 ) {
99870f4d
KW
1001 $why_suppressed{$property} = $why_deprecated{$property};
1002 }
cf25bb62 1003
99870f4d
KW
1004 # Customize the message for all the 'Other_' properties
1005 foreach my $property (keys %why_deprecated) {
1006 next if (my $main_property = $property) !~ s/^Other_//;
1007 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
1008 }
1009}
1010
e9c4b4f8
KW
1011if ($write_Unicode_deprecated_tables) {
1012 foreach my $property (keys %why_suppressed) {
1013 delete $why_suppressed{$property} if $property =~
1014 / ^ Other | Grapheme /x;
1015 }
1016}
1017
99870f4d
KW
1018if ($v_version ge 4.0.0) {
1019 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
1020 if ($v_version ge 6.0.0) {
1021 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
1022 }
99870f4d 1023}
5f7264c7 1024if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 1025 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 1026 if ($v_version ge 6.0.0) {
63f74647 1027 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 1028 }
99870f4d
KW
1029}
1030
1031# Probably obsolete forever
1032if ($v_version ge v4.1.0) {
82aed44a
KW
1033 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
1034}
1035if ($v_version ge v6.0.0) {
caa75395 1036 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana" (or both)';
2b352efd 1037 $why_suppressed{'Script_Extensions=Katakana_Or_Hiragana'} = 'All code points that would be matched by this are matched by either "Script_Extensions=Katakana" or "Script_Extensions=Hiragana"';
99870f4d
KW
1038}
1039
1040# This program can create files for enumerated-like properties, such as
1041# 'Numeric_Type'. This file would be the same format as for a string
1042# property, with a mapping from code point to its value, so you could look up,
1043# for example, the script a code point is in. But no one so far wants this
1044# mapping, or they have found another way to get it since this is a new
1045# feature. So no file is generated except if it is in this list.
1046my @output_mapped_properties = split "\n", <<END;
1047END
1048
b9e57660
KW
1049# If you want more Unihan properties than the default, you need to add them to
1050# these arrays. Depending on the property type, @missing lines might have to
1051# be added to the second array. A sample entry would be (including the '#'):
1052# @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
99870f4d 1053my @cjk_properties = split "\n", <<'END';
99870f4d 1054END
99870f4d 1055my @cjk_property_values = split "\n", <<'END';
99870f4d
KW
1056END
1057
1058# The input files don't list every code point. Those not listed are to be
1059# defaulted to some value. Below are hard-coded what those values are for
1060# non-binary properties as of 5.1. Starting in 5.0, there are
caa75395 1061# machine-parsable comment lines in the files that give the defaults; so this
99870f4d
KW
1062# list shouldn't have to be extended. The claim is that all missing entries
1063# for binary properties will default to 'N'. Unicode tried to change that in
1064# 5.2, but the beta period produced enough protest that they backed off.
1065#
1066# The defaults for the fields that appear in UnicodeData.txt in this hash must
1067# be in the form that it expects. The others may be synonyms.
1068my $CODE_POINT = '<code point>';
1069my %default_mapping = (
1070 Age => "Unassigned",
1071 # Bidi_Class => Complicated; set in code
1072 Bidi_Mirroring_Glyph => "",
1073 Block => 'No_Block',
1074 Canonical_Combining_Class => 0,
1075 Case_Folding => $CODE_POINT,
1076 Decomposition_Mapping => $CODE_POINT,
1077 Decomposition_Type => 'None',
1078 East_Asian_Width => "Neutral",
1079 FC_NFKC_Closure => $CODE_POINT,
ec2cd995 1080 General_Category => ($v_version le 6.3.0) ? 'Cn' : 'Unassigned',
99870f4d
KW
1081 Grapheme_Cluster_Break => 'Other',
1082 Hangul_Syllable_Type => 'NA',
1083 ISO_Comment => "",
1084 Jamo_Short_Name => "",
1085 Joining_Group => "No_Joining_Group",
1086 # Joining_Type => Complicated; set in code
1087 kIICore => 'N', # Is converted to binary
1088 #Line_Break => Complicated; set in code
1089 Lowercase_Mapping => $CODE_POINT,
1090 Name => "",
1091 Name_Alias => "",
1092 NFC_QC => 'Yes',
1093 NFD_QC => 'Yes',
1094 NFKC_QC => 'Yes',
1095 NFKD_QC => 'Yes',
1096 Numeric_Type => 'None',
1097 Numeric_Value => 'NaN',
1098 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1099 Sentence_Break => 'Other',
1100 Simple_Case_Folding => $CODE_POINT,
1101 Simple_Lowercase_Mapping => $CODE_POINT,
1102 Simple_Titlecase_Mapping => $CODE_POINT,
1103 Simple_Uppercase_Mapping => $CODE_POINT,
1104 Titlecase_Mapping => $CODE_POINT,
1105 Unicode_1_Name => "",
1106 Unicode_Radical_Stroke => "",
1107 Uppercase_Mapping => $CODE_POINT,
1108 Word_Break => 'Other',
1109);
1110
678f13d5 1111### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1112
1113my $HEADER=<<"EOF";
1114# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85 1115# This file is machine-generated by $0 from the Unicode
8f21889c 1116# database, Version $unicode_version. Any changes made here will be lost!
cf25bb62
JH
1117EOF
1118
126c3d4e 1119my $INTERNAL_ONLY_HEADER = <<"EOF";
99870f4d
KW
1120
1121# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1122# This file is for internal use by core Perl only. The format and even the
1123# name or existence of this file are subject to change without notice. Don't
d9ae3878
KW
1124# use it directly. Use Unicode::UCD to access the Unicode character data
1125# base.
99870f4d
KW
1126EOF
1127
1128my $DEVELOPMENT_ONLY=<<"EOF";
1129# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1130# This file contains information artificially constrained to code points
1131# present in Unicode release $string_compare_versions.
1132# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1133# not be used for production.
b6922eda
KW
1134
1135EOF
1136
346a20cf
KW
1137my $MAX_UNICODE_CODEPOINT_STRING = ($v_version ge v2.0.0)
1138 ? "10FFFF"
1139 : "FFFF";
6189eadc
KW
1140my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1141my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d 1142
026d3e0c
KW
1143# We work with above-Unicode code points, up to IV_MAX, but we may want to use
1144# sentinels above that number. Therefore for internal use, we use a much
1145# smaller number, translating it to IV_MAX only for output. The exact number
7fdcbfeb
KW
1146# is immaterial (all above-Unicode code points are treated exactly the same),
1147# but the algorithm requires it to be at least
1148# 2 * $MAX_UNICODE_CODEPOINTS + 1
2d88a86a
KW
1149my $MAX_WORKING_CODEPOINTS= $MAX_UNICODE_CODEPOINT * 8;
1150my $MAX_WORKING_CODEPOINT = $MAX_WORKING_CODEPOINTS - 1;
1151my $MAX_WORKING_CODEPOINT_STRING = sprintf("%X", $MAX_WORKING_CODEPOINT);
1152
026d3e0c 1153my $MAX_PLATFORM_CODEPOINT = ~0 >> 1;
2d88a86a 1154
99870f4d
KW
1155# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1156# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1157# to decrease backtracking. The first regex allows the code point to be at
1158# the end of a word, but to work properly, the word shouldn't end with a valid
1159# hex character. The second one won't match a code point at the end of a
1160# word, and doesn't have the run-on issue
8c32d378
KW
1161my $run_on_code_point_re =
1162 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1163my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d 1164
44b0f8c1 1165# This matches the beginning of the line in the Unicode DB files that give the
99870f4d
KW
1166# defaults for code points not listed (i.e., missing) in the file. The code
1167# depends on this ending with a semi-colon, so it can assume it is a valid
1168# field when the line is split() by semi-colons
346a20cf 1169my $missing_defaults_prefix = qr/^#\s+\@missing:\s+0000\.\.10FFFF\s*;/;
99870f4d
KW
1170
1171# Property types. Unicode has more types, but these are sufficient for our
1172# purposes.
1173my $UNKNOWN = -1; # initialized to illegal value
1174my $NON_STRING = 1; # Either binary or enum
1175my $BINARY = 2;
06f26c45
KW
1176my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1177 # tables, additional true and false tables are
1178 # generated so that false is anything matching the
1179 # default value, and true is everything else.
1180my $ENUM = 4; # Include catalog
1181my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1182
1183# Some input files have lines that give default values for code points not
1184# contained in the file. Sometimes these should be ignored.
1185my $NO_DEFAULTS = 0; # Must evaluate to false
1186my $NOT_IGNORED = 1;
1187my $IGNORED = 2;
1188
1189# Range types. Each range has a type. Most ranges are type 0, for normal,
1190# and will appear in the main body of the tables in the output files, but
1191# there are other types of ranges as well, listed below, that are specially
1192# handled. There are pseudo-types as well that will never be stored as a
1193# type, but will affect the calculation of the type.
1194
1195# 0 is for normal, non-specials
1196my $MULTI_CP = 1; # Sequence of more than code point
1197my $HANGUL_SYLLABLE = 2;
1198my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1199my $NULL = 4; # The map is to the null string; utf8.c can't
1200 # handle these, nor is there an accepted syntax
1201 # for them in \p{} constructs
f86864ac 1202my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1203 # otherwise be $MULTI_CP type are instead type 0
1204
1205# process_generic_property_file() can accept certain overrides in its input.
1206# Each of these must begin AND end with $CMD_DELIM.
1207my $CMD_DELIM = "\a";
1208my $REPLACE_CMD = 'replace'; # Override the Replace
1209my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1210
1211my $NO = 0;
1212my $YES = 1;
1213
1214# Values for the Replace argument to add_range.
1215# $NO # Don't replace; add only the code points not
1216 # already present.
1217my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1218 # the comments at the subroutine definition.
1219my $UNCONDITIONALLY = 2; # Replace without conditions.
9470941f 1220my $MULTIPLE_BEFORE = 4; # Don't replace, but add a duplicate record if
99870f4d 1221 # already there
7f4b1e25
KW
1222my $MULTIPLE_AFTER = 5; # Don't replace, but add a duplicate record if
1223 # already there
1224my $CROAK = 6; # Die with an error if is already there
99870f4d
KW
1225
1226# Flags to give property statuses. The phrases are to remind maintainers that
1227# if the flag is changed, the indefinite article referring to it in the
1228# documentation may need to be as well.
1229my $NORMAL = "";
99870f4d
KW
1230my $DEPRECATED = 'D';
1231my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1232my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1233my $DISCOURAGED = 'X';
1234my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1235my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1236my $STRICTER = 'T';
1237my $a_bold_stricter = "a 'B<$STRICTER>'";
1238my $A_bold_stricter = "A 'B<$STRICTER>'";
1239my $STABILIZED = 'S';
1240my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1241my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1242my $OBSOLETE = 'O';
1243my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1244my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1245
5c6d9947
KW
1246# Aliases can also have an extra status:
1247my $INTERNAL_ALIAS = 'P';
1248
99870f4d
KW
1249my %status_past_participles = (
1250 $DISCOURAGED => 'discouraged',
99870f4d
KW
1251 $STABILIZED => 'stabilized',
1252 $OBSOLETE => 'obsolete',
37e2e78e 1253 $DEPRECATED => 'deprecated',
5c6d9947 1254 $INTERNAL_ALIAS => 'reserved for Perl core internal use only',
99870f4d
KW
1255);
1256
395dfc19
KW
1257# Table fates. These are somewhat ordered, so that fates < $MAP_PROXIED should be
1258# externally documented.
301ba948 1259my $ORDINARY = 0; # The normal fate.
395dfc19
KW
1260my $MAP_PROXIED = 1; # The map table for the property isn't written out,
1261 # but there is a file written that can be used to
1262 # reconstruct this table
3cdaf629 1263my $INTERNAL_ONLY = 2; # The file for this table is written out, but it is
301ba948 1264 # for Perl's internal use only
277b7b16
KW
1265my $LEGACY_ONLY = 3; # Like $INTERNAL_ONLY, but not actually used by Perl.
1266 # Is for backwards compatibility for applications that
1267 # read the file directly, so it's format is
1268 # unchangeable.
1269my $SUPPRESSED = 4; # The file for this table is not written out, and as a
3cdaf629
KW
1270 # result, we don't bother to do many computations on
1271 # it.
277b7b16 1272my $PLACEHOLDER = 5; # Like $SUPPRESSED, but we go through all the
3cdaf629
KW
1273 # computations anyway, as the values are needed for
1274 # things to work. This happens when we have Perl
1275 # extensions that depend on Unicode tables that
1276 # wouldn't normally be in a given Unicode version.
301ba948 1277
f5817e0a
KW
1278# The format of the values of the tables:
1279my $EMPTY_FORMAT = "";
99870f4d
KW
1280my $BINARY_FORMAT = 'b';
1281my $DECIMAL_FORMAT = 'd';
1282my $FLOAT_FORMAT = 'f';
1283my $INTEGER_FORMAT = 'i';
1284my $HEX_FORMAT = 'x';
1285my $RATIONAL_FORMAT = 'r';
1286my $STRING_FORMAT = 's';
d11155ec 1287my $ADJUST_FORMAT = 'a';
24303724 1288my $HEX_ADJUST_FORMAT = 'ax';
a14f3cb1 1289my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1290my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1291
1292my %map_table_formats = (
1293 $BINARY_FORMAT => 'binary',
1294 $DECIMAL_FORMAT => 'single decimal digit',
1295 $FLOAT_FORMAT => 'floating point number',
1296 $INTEGER_FORMAT => 'integer',
add63c13 1297 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1298 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1299 $STRING_FORMAT => 'string',
d11155ec 1300 $ADJUST_FORMAT => 'some entries need adjustment',
24303724 1301 $HEX_ADJUST_FORMAT => 'mapped value in hex; some entries need adjustment',
92f9d56c 1302 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1303 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1304);
1305
1306# Unicode didn't put such derived files in a separate directory at first.
1307my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1308my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1309my $AUXILIARY = 'auxiliary';
1310
3854b4b8
KW
1311# Hashes and arrays that will eventually go into Heavy.pl for the use of
1312# utf8_heavy.pl and into UCD.pl for the use of UCD.pm
99870f4d
KW
1313my %loose_to_file_of; # loosely maps table names to their respective
1314 # files
1315my %stricter_to_file_of; # same; but for stricter mapping.
315bfd4e 1316my %loose_property_to_file_of; # Maps a loose property name to its map file
e1dc048d 1317my %strict_property_to_file_of; # Same, but strict
3854b4b8
KW
1318my @inline_definitions = "V0"; # Each element gives a definition of a unique
1319 # inversion list. When a definition is inlined,
1320 # its value in the hash it's in (one of the two
1321 # defined just above) will include an index into
1322 # this array. The 0th element is initialized to
9be22ca5 1323 # the definition for a zero length inversion list
89cf10cc
KW
1324my %file_to_swash_name; # Maps the file name to its corresponding key name
1325 # in the hash %utf8::SwashInfo
99870f4d
KW
1326my %nv_floating_to_rational; # maps numeric values floating point numbers to
1327 # their rational equivalent
c12f2655
KW
1328my %loose_property_name_of; # Loosely maps (non_string) property names to
1329 # standard form
e1dc048d
KW
1330my %strict_property_name_of; # Strictly maps (non_string) property names to
1331 # standard form
86a52d1e 1332my %string_property_loose_to_name; # Same, for string properties.
c15fda25
KW
1333my %loose_defaults; # keys are of form "prop=value", where 'prop' is
1334 # the property name in standard loose form, and
1335 # 'value' is the default value for that property,
1336 # also in standard loose form.
9e4a1e86
KW
1337my %loose_to_standard_value; # loosely maps table names to the canonical
1338 # alias for them
2df7880f
KW
1339my %ambiguous_names; # keys are alias names (in standard form) that
1340 # have more than one possible meaning.
d963b40d
KW
1341my %combination_property; # keys are alias names (in standard form) that
1342 # have both a map table, and a binary one that
1343 # yields true for all non-null maps.
5d1df013
KW
1344my %prop_aliases; # Keys are standard property name; values are each
1345 # one's aliases
1e863613
KW
1346my %prop_value_aliases; # Keys of top level are standard property name;
1347 # values are keys to another hash, Each one is
1348 # one of the property's values, in standard form.
1349 # The values are that prop-val's aliases.
b9e57660 1350my %skipped_files; # List of files that we skip
2df7880f 1351my %ucd_pod; # Holds entries that will go into the UCD section of the pod
99870f4d 1352
d867ccfb
KW
1353# Most properties are immune to caseless matching, otherwise you would get
1354# nonsensical results, as properties are a function of a code point, not
1355# everything that is caselessly equivalent to that code point. For example,
1356# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1357# be true because 's' and 'S' are equivalent caselessly. However,
1358# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1359# extend that concept to those very few properties that are like this. Each
1360# such property will match the full range caselessly. They are hard-coded in
1361# the program; it's not worth trying to make it general as it's extremely
1362# unlikely that they will ever change.
1363my %caseless_equivalent_to;
1364
53adf6a2
KW
1365# This is the range of characters that were in Release 1 of Unicode, and
1366# removed in Release 2 (replaced with the current Hangul syllables starting at
1367# U+AC00). The range was reused starting in Release 3 for other purposes.
1368my $FIRST_REMOVED_HANGUL_SYLLABLE = 0x3400;
1369my $FINAL_REMOVED_HANGUL_SYLLABLE = 0x4DFF;
1370
99870f4d
KW
1371# These constants names and values were taken from the Unicode standard,
1372# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1373# syllables. The '_string' versions are so generated tables can retain the
1374# hex format, which is the more familiar value
1375my $SBase_string = "0xAC00";
1376my $SBase = CORE::hex $SBase_string;
1377my $LBase_string = "0x1100";
1378my $LBase = CORE::hex $LBase_string;
1379my $VBase_string = "0x1161";
1380my $VBase = CORE::hex $VBase_string;
1381my $TBase_string = "0x11A7";
1382my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1383my $SCount = 11172;
1384my $LCount = 19;
1385my $VCount = 21;
1386my $TCount = 28;
1387my $NCount = $VCount * $TCount;
1388
1389# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1390# with the above published constants.
1391my %Jamo;
1392my %Jamo_L; # Leading consonants
1393my %Jamo_V; # Vowels
1394my %Jamo_T; # Trailing consonants
1395
bb1dd3da
KW
1396# For code points whose name contains its ordinal as a '-ABCD' suffix.
1397# The key is the base name of the code point, and the value is an
1398# array giving all the ranges that use this base name. Each range
1399# is actually a hash giving the 'low' and 'high' values of it.
1400my %names_ending_in_code_point;
1401my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1402 # removed from the names
1403# Inverse mapping. The list of ranges that have these kinds of
1404# names. Each element contains the low, high, and base names in an
1405# anonymous hash.
1406my @code_points_ending_in_code_point;
1407
6b5ab373
KW
1408# To hold Unicode's normalization test suite
1409my @normalization_tests;
1410
bb1dd3da
KW
1411# Boolean: does this Unicode version have the hangul syllables, and are we
1412# writing out a table for them?
1413my $has_hangul_syllables = 0;
1414
1415# Does this Unicode version have code points whose names end in their
1416# respective code points, and are we writing out a table for them? 0 for no;
1417# otherwise points to first property that a table is needed for them, so that
1418# if multiple tables are needed, we don't create duplicates
1419my $needing_code_points_ending_in_code_point = 0;
1420
37e2e78e 1421my @backslash_X_tests; # List of tests read in for testing \X
6b659339 1422my @LB_tests; # List of tests read in for testing \b{lb}
06ae2722 1423my @SB_tests; # List of tests read in for testing \b{sb}
ae3bb8ea 1424my @WB_tests; # List of tests read in for testing \b{wb}
99870f4d
KW
1425my @unhandled_properties; # Will contain a list of properties found in
1426 # the input that we didn't process.
f86864ac 1427my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1428 # listed in the pod
1429my @map_properties; # Properties that get map files written
1430my @named_sequences; # NamedSequences.txt contents.
1431my %potential_files; # Generated list of all .txt files in the directory
1432 # structure so we can warn if something is being
1433 # ignored.
a0204bb1
KW
1434my @missing_early_files; # Generated list of absent files that we need to
1435 # proceed in compiling this early Unicode version
99870f4d
KW
1436my @files_actually_output; # List of files we generated.
1437my @more_Names; # Some code point names are compound; this is used
1438 # to store the extra components of them.
0426f635
KW
1439my $E_FLOAT_PRECISION = 2; # The minimum number of digits after the decimal
1440 # point of a normalized floating point number
1441 # needed to match before we consider it equivalent
1442 # to a candidate rational
99870f4d
KW
1443
1444# These store references to certain commonly used property objects
1254636b 1445my $age;
3c88a801 1446my $ccc;
99870f4d
KW
1447my $gc;
1448my $perl;
1449my $block;
3e20195b
KW
1450my $perl_charname;
1451my $print;
ac7dbdb6 1452my $All;
bc0c431d 1453my $Assigned; # All assigned characters in this Unicode release
e47e66b9
KW
1454my $DI; # Default_Ignorable_Code_Point property
1455my $NChar; # Noncharacter_Code_Point property
359523e2 1456my $script;
8fb542fb 1457my $scx; # Script_Extensions property
99870f4d
KW
1458
1459# Are there conflicting names because of beginning with 'In_', or 'Is_'
1460my $has_In_conflicts = 0;
1461my $has_Is_conflicts = 0;
1462
1463sub internal_file_to_platform ($) {
1464 # Convert our file paths which have '/' separators to those of the
1465 # platform.
1466
1467 my $file = shift;
1468 return undef unless defined $file;
1469
1470 return File::Spec->join(split '/', $file);
d07a55ed 1471}
5beb625e 1472
99870f4d
KW
1473sub file_exists ($) { # platform independent '-e'. This program internally
1474 # uses slash as a path separator.
1475 my $file = shift;
1476 return 0 if ! defined $file;
1477 return -e internal_file_to_platform($file);
1478}
5beb625e 1479
99870f4d 1480sub objaddr($) {
23e33b60
KW
1481 # Returns the address of the blessed input object.
1482 # It doesn't check for blessedness because that would do a string eval
1483 # every call, and the program is structured so that this is never called
1484 # for a non-blessed object.
99870f4d 1485
23e33b60 1486 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1487
1488 # Numifying a ref gives its address.
051df77b 1489 return pack 'J', $_[0];
99870f4d
KW
1490}
1491
558712cf 1492# These are used only if $annotate is true.
c4019d52
KW
1493# The entire range of Unicode characters is examined to populate these
1494# after all the input has been processed. But most can be skipped, as they
1495# have the same descriptive phrases, such as being unassigned
1496my @viacode; # Contains the 1 million character names
4b06b8de 1497my @age; # And their ages ("" if none)
c4019d52
KW
1498my @printable; # boolean: And are those characters printable?
1499my @annotate_char_type; # Contains a type of those characters, specifically
1500 # for the purposes of annotation.
1501my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1502 # name for the purposes of annotation. They map to the
c4019d52
KW
1503 # upper edge of the range, so that the end point can
1504 # be immediately found. This is used to skip ahead to
1505 # the end of a range, and avoid processing each
1506 # individual code point in it.
1507my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1508 # characters, but excluding those which are
1509 # also noncharacter code points
1510
1511# The annotation types are an extension of the regular range types, though
1512# some of the latter are folded into one. Make the new types negative to
1513# avoid conflicting with the regular types
1514my $SURROGATE_TYPE = -1;
1515my $UNASSIGNED_TYPE = -2;
1516my $PRIVATE_USE_TYPE = -3;
1517my $NONCHARACTER_TYPE = -4;
1518my $CONTROL_TYPE = -5;
2d88a86a
KW
1519my $ABOVE_UNICODE_TYPE = -6;
1520my $UNKNOWN_TYPE = -7; # Used only if there is a bug in this program
c4019d52
KW
1521
1522sub populate_char_info ($) {
558712cf 1523 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1524 # input code point's info that are needed for outputting more detailed
1525 # comments. If calling context wants a return, it is the end point of
1526 # any contiguous range of characters that share essentially the same info
1527
1528 my $i = shift;
1529 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1530
1531 $viacode[$i] = $perl_charname->value_of($i) || "";
1254636b 1532 $age[$i] = (defined $age)
272af058 1533 ? (($age->value_of($i) =~ / ^ \d+ \. \d+ $ /x)
1254636b
KW
1534 ? $age->value_of($i)
1535 : "")
1536 : "";
c4019d52
KW
1537
1538 # A character is generally printable if Unicode says it is,
1539 # but below we make sure that most Unicode general category 'C' types
1540 # aren't.
1541 $printable[$i] = $print->contains($i);
1542
1254636b
KW
1543 # But the characters in this range were removed in v2.0 and replaced by
1544 # different ones later. Modern fonts will be for the replacement
1545 # characters, so suppress printing them.
1546 if (($v_version lt v2.0
1547 || ($compare_versions && $compare_versions lt v2.0))
1548 && ( $i >= $FIRST_REMOVED_HANGUL_SYLLABLE
1549 && $i <= $FINAL_REMOVED_HANGUL_SYLLABLE))
1550 {
1551 $printable[$i] = 0;
1552 }
1553
c4019d52
KW
1554 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1555
1556 # Only these two regular types are treated specially for annotations
1557 # purposes
1558 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1559 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1560
1561 # Give a generic name to all code points that don't have a real name.
1562 # We output ranges, if applicable, for these. Also calculate the end
1563 # point of the range.
1564 my $end;
1565 if (! $viacode[$i]) {
2d88a86a
KW
1566 if ($i > $MAX_UNICODE_CODEPOINT) {
1567 $viacode[$i] = 'Above-Unicode';
1568 $annotate_char_type[$i] = $ABOVE_UNICODE_TYPE;
1569 $printable[$i] = 0;
1570 $end = $MAX_WORKING_CODEPOINT;
1571 }
1572 elsif ($gc-> table('Private_use')->contains($i)) {
c4019d52
KW
1573 $viacode[$i] = 'Private Use';
1574 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1575 $printable[$i] = 0;
1576 $end = $gc->table('Private_Use')->containing_range($i)->end;
1577 }
e47e66b9 1578 elsif ($NChar->contains($i)) {
c4019d52
KW
1579 $viacode[$i] = 'Noncharacter';
1580 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1581 $printable[$i] = 0;
e47e66b9 1582 $end = $NChar->containing_range($i)->end;
c4019d52
KW
1583 }
1584 elsif ($gc-> table('Control')->contains($i)) {
1254636b
KW
1585 my $name_ref = property_ref('Name_Alias');
1586 $name_ref = property_ref('Unicode_1_Name') if ! defined $name_ref;
1587 $viacode[$i] = (defined $name_ref)
1588 ? $name_ref->value_of($i)
1589 : 'Control';
c4019d52
KW
1590 $annotate_char_type[$i] = $CONTROL_TYPE;
1591 $printable[$i] = 0;
c4019d52
KW
1592 }
1593 elsif ($gc-> table('Unassigned')->contains($i)) {
c4019d52
KW
1594 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1595 $printable[$i] = 0;
1254636b
KW
1596 $viacode[$i] = 'Unassigned';
1597
349edad4 1598 if (defined $block) { # No blocks in earliest releases
1254636b 1599 $viacode[$i] .= ', block=' . $block-> value_of($i);
1d025d66 1600 $end = $gc-> table('Unassigned')->containing_range($i)->end;
c4019d52 1601
bf06c733
KW
1602 # Because we name the unassigned by the blocks they are in, it
1603 # can't go past the end of that block, and it also can't go
1604 # past the unassigned range it is in. The special table makes
1605 # sure that the non-characters, which are unassigned, are
1606 # separated out.
1607 $end = min($block->containing_range($i)->end,
1608 $unassigned_sans_noncharacters->
1609 containing_range($i)->end);
1d025d66 1610 }
1254636b
KW
1611 else {
1612 $end = $i + 1;
1613 while ($unassigned_sans_noncharacters->contains($end)) {
1614 $end++;
1615 }
1616 $end--;
1617 }
1d025d66 1618 }
35e2b416 1619 elsif ($perl->table('_Perl_Surrogate')->contains($i)) {
1d025d66
KW
1620 $viacode[$i] = 'Surrogate';
1621 $annotate_char_type[$i] = $SURROGATE_TYPE;
1622 $printable[$i] = 0;
1623 $end = $gc->table('Surrogate')->containing_range($i)->end;
13ca76ff
KW
1624 }
1625 else {
1626 Carp::my_carp_bug("Can't figure out how to annotate "
1627 . sprintf("U+%04X", $i)
1628 . ". Proceeding anyway.");
c4019d52
KW
1629 $viacode[$i] = 'UNKNOWN';
1630 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1631 $printable[$i] = 0;
1632 }
1633 }
1634
1635 # Here, has a name, but if it's one in which the code point number is
1636 # appended to the name, do that.
1637 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1638 $viacode[$i] .= sprintf("-%04X", $i);
4b06b8de 1639
4b06b8de 1640 my $limit = $perl_charname->containing_range($i)->end;
1254636b
KW
1641 if (defined $age) {
1642 # Do all these as groups of the same age, instead of individually,
1643 # because their names are so meaningless, and there are typically
1644 # large quantities of them.
1645 $end = $i + 1;
1646 while ($end <= $limit && $age->value_of($end) == $age[$i]) {
1647 $end++;
1648 }
1649 $end--;
1650 }
1651 else {
1652 $end = $limit;
4b06b8de 1653 }
c4019d52
KW
1654 }
1655
1656 # And here, has a name, but if it's a hangul syllable one, replace it with
1657 # the correct name from the Unicode algorithm
1658 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1659 use integer;
1660 my $SIndex = $i - $SBase;
1661 my $L = $LBase + $SIndex / $NCount;
1662 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1663 my $T = $TBase + $SIndex % $TCount;
1664 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1665 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1666 $end = $perl_charname->containing_range($i)->end;
1667 }
1668
1669 return if ! defined wantarray;
1670 return $i if ! defined $end; # If not a range, return the input
1671
1672 # Save this whole range so can find the end point quickly
1673 $annotate_ranges->add_map($i, $end, $end);
1674
1675 return $end;
1676}
1677
23e33b60
KW
1678# Commented code below should work on Perl 5.8.
1679## This 'require' doesn't necessarily work in miniperl, and even if it does,
1680## the native perl version of it (which is what would operate under miniperl)
1681## is extremely slow, as it does a string eval every call.
7e017d6d 1682#my $has_fast_scalar_util = $^X !~ /miniperl/
23e33b60
KW
1683# && defined eval "require Scalar::Util";
1684#
1685#sub objaddr($) {
1686# # Returns the address of the blessed input object. Uses the XS version if
1687# # available. It doesn't check for blessedness because that would do a
1688# # string eval every call, and the program is structured so that this is
1689# # never called for a non-blessed object.
1690#
1691# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1692#
1693# # Check at least that is a ref.
1694# my $pkg = ref($_[0]) or return undef;
1695#
1696# # Change to a fake package to defeat any overloaded stringify
1697# bless $_[0], 'main::Fake';
1698#
1699# # Numifying a ref gives its address.
051df77b 1700# my $addr = pack 'J', $_[0];
23e33b60
KW
1701#
1702# # Return to original class
1703# bless $_[0], $pkg;
1704# return $addr;
1705#}
1706
99870f4d
KW
1707sub max ($$) {
1708 my $a = shift;
1709 my $b = shift;
1710 return $a if $a >= $b;
1711 return $b;
1712}
1713
1714sub min ($$) {
1715 my $a = shift;
1716 my $b = shift;
1717 return $a if $a <= $b;
1718 return $b;
1719}
1720
1721sub clarify_number ($) {
1722 # This returns the input number with underscores inserted every 3 digits
1723 # in large (5 digits or more) numbers. Input must be entirely digits, not
1724 # checked.
1725
1726 my $number = shift;
1727 my $pos = length($number) - 3;
1728 return $number if $pos <= 1;
1729 while ($pos > 0) {
1730 substr($number, $pos, 0) = '_';
1731 $pos -= 3;
5beb625e 1732 }
99870f4d 1733 return $number;
99598c8c
JH
1734}
1735
731cb813
KW
1736sub clarify_code_point_count ($) {
1737 # This is like clarify_number(), but the input is assumed to be a count of
1738 # code points, rather than a generic number.
1739
2d88a86a
KW
1740 my $append = "";
1741
1742 my $number = shift;
1743 if ($number > $MAX_UNICODE_CODEPOINTS) {
1744 $number -= ($MAX_WORKING_CODEPOINTS - $MAX_UNICODE_CODEPOINTS);
1745 return "All above-Unicode code points" if $number == 0;
1746 $append = " + all above-Unicode code points";
1747 }
1748 return clarify_number($number) . $append;
731cb813 1749}
12ac2576 1750
99870f4d 1751package Carp;
7ebf06b3 1752
99870f4d
KW
1753# These routines give a uniform treatment of messages in this program. They
1754# are placed in the Carp package to cause the stack trace to not include them,
1755# although an alternative would be to use another package and set @CARP_NOT
1756# for it.
12ac2576 1757
99870f4d 1758our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1759
99f78760
KW
1760# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1761# and overload trying to load Scalar:Util under miniperl. See
1762# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1763undef $overload::VERSION;
1764
99870f4d
KW
1765sub my_carp {
1766 my $message = shift || "";
1767 my $nofold = shift || 0;
7ebf06b3 1768
99870f4d
KW
1769 if ($message) {
1770 $message = main::join_lines($message);
1771 $message =~ s/^$0: *//; # Remove initial program name
1772 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1773 $message = "\n$0: $message;";
12ac2576 1774
99870f4d
KW
1775 # Fold the message with program name, semi-colon end punctuation
1776 # (which looks good with the message that carp appends to it), and a
1777 # hanging indent for continuation lines.
1778 $message = main::simple_fold($message, "", 4) unless $nofold;
1779 $message =~ s/\n$//; # Remove the trailing nl so what carp
1780 # appends is to the same line
1781 }
12ac2576 1782
99870f4d 1783 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1784
99870f4d
KW
1785 carp $message;
1786 return;
1787}
7ebf06b3 1788
99870f4d
KW
1789sub my_carp_bug {
1790 # This is called when it is clear that the problem is caused by a bug in
1791 # this program.
7ebf06b3 1792
99870f4d
KW
1793 my $message = shift;
1794 $message =~ s/^$0: *//;
1795 $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");
1796 carp $message;
1797 return;
1798}
7ebf06b3 1799
99870f4d
KW
1800sub carp_too_few_args {
1801 if (@_ != 2) {
1802 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1803 return;
12ac2576 1804 }
7ebf06b3 1805
99870f4d
KW
1806 my $args_ref = shift;
1807 my $count = shift;
7ebf06b3 1808
99870f4d
KW
1809 my_carp_bug("Need at least $count arguments to "
1810 . (caller 1)[3]
1811 . ". Instead got: '"
1812 . join ', ', @$args_ref
1813 . "'. No action taken.");
1814 return;
12ac2576
JP
1815}
1816
99870f4d
KW
1817sub carp_extra_args {
1818 my $args_ref = shift;
1819 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1820
99870f4d
KW
1821 unless (ref $args_ref) {
1822 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1823 return;
1824 }
1825 my ($package, $file, $line) = caller;
1826 my $subroutine = (caller 1)[3];
cf25bb62 1827
99870f4d
KW
1828 my $list;
1829 if (ref $args_ref eq 'HASH') {
1830 foreach my $key (keys %$args_ref) {
1831 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1832 }
99870f4d 1833 $list = join ', ', each %{$args_ref};
cf25bb62 1834 }
99870f4d
KW
1835 elsif (ref $args_ref eq 'ARRAY') {
1836 foreach my $arg (@$args_ref) {
1837 $arg = $UNDEF unless defined $arg;
1838 }
1839 $list = join ', ', @$args_ref;
1840 }
1841 else {
1842 my_carp_bug("Can't cope with ref "
1843 . ref($args_ref)
1844 . " . argument to 'carp_extra_args'. Not checking arguments.");
1845 return;
1846 }
1847
1848 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1849 return;
d73e5302
JH
1850}
1851
99870f4d
KW
1852package main;
1853
1854{ # Closure
1855
1856 # This program uses the inside-out method for objects, as recommended in
1b6b3fa9
KW
1857 # "Perl Best Practices". (This is the best solution still, since this has
1858 # to run under miniperl.) This closure aids in generating those. There
99870f4d
KW
1859 # are two routines. setup_package() is called once per package to set
1860 # things up, and then set_access() is called for each hash representing a
1861 # field in the object. These routines arrange for the object to be
1862 # properly destroyed when no longer used, and for standard accessor
1863 # functions to be generated. If you need more complex accessors, just
1864 # write your own and leave those accesses out of the call to set_access().
1865 # More details below.
1866
1867 my %constructor_fields; # fields that are to be used in constructors; see
1868 # below
1869
1870 # The values of this hash will be the package names as keys to other
1871 # hashes containing the name of each field in the package as keys, and
1872 # references to their respective hashes as values.
1873 my %package_fields;
1874
1875 sub setup_package {
1876 # Sets up the package, creating standard DESTROY and dump methods
1877 # (unless already defined). The dump method is used in debugging by
1878 # simple_dumper().
1879 # The optional parameters are:
1880 # a) a reference to a hash, that gets populated by later
1881 # set_access() calls with one of the accesses being
1882 # 'constructor'. The caller can then refer to this, but it is
1883 # not otherwise used by these two routines.
1884 # b) a reference to a callback routine to call during destruction
1885 # of the object, before any fields are actually destroyed
1886
1887 my %args = @_;
1888 my $constructor_ref = delete $args{'Constructor_Fields'};
1889 my $destroy_callback = delete $args{'Destroy_Callback'};
1890 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1891
1892 my %fields;
1893 my $package = (caller)[0];
1894
1895 $package_fields{$package} = \%fields;
1896 $constructor_fields{$package} = $constructor_ref;
1897
1898 unless ($package->can('DESTROY')) {
1899 my $destroy_name = "${package}::DESTROY";
1900 no strict "refs";
1901
1902 # Use typeglob to give the anonymous subroutine the name we want
1903 *$destroy_name = sub {
1904 my $self = shift;
ffe43484 1905 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1906
1907 $self->$destroy_callback if $destroy_callback;
1908 foreach my $field (keys %{$package_fields{$package}}) {
1909 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1910 delete $package_fields{$package}{$field}{$addr};
1911 }
1912 return;
1913 }
1914 }
1915
1916 unless ($package->can('dump')) {
1917 my $dump_name = "${package}::dump";
1918 no strict "refs";
1919 *$dump_name = sub {
1920 my $self = shift;
1921 return dump_inside_out($self, $package_fields{$package}, @_);
1922 }
1923 }
1924 return;
1925 }
1926
1927 sub set_access {
1928 # Arrange for the input field to be garbage collected when no longer
1929 # needed. Also, creates standard accessor functions for the field
1930 # based on the optional parameters-- none if none of these parameters:
1931 # 'addable' creates an 'add_NAME()' accessor function.
1932 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1933 # function.
1934 # 'settable' creates a 'set_NAME()' accessor function.
1935 # 'constructor' doesn't create an accessor function, but adds the
1936 # field to the hash that was previously passed to
1937 # setup_package();
1938 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1939 # 'add' etc. all mean 'addable'.
1940 # The read accessor function will work on both array and scalar
1941 # values. If another accessor in the parameter list is 'a', the read
1942 # access assumes an array. You can also force it to be array access
1943 # by specifying 'readable_array' instead of 'readable'
1944 #
1945 # A sort-of 'protected' access can be set-up by preceding the addable,
1946 # readable or settable with some initial portion of 'protected_' (but,
1947 # the underscore is required), like 'p_a', 'pro_set', etc. The
1948 # "protection" is only by convention. All that happens is that the
1949 # accessor functions' names begin with an underscore. So instead of
1950 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1951 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1952 # end of each package, and then storing the __LINE__ ranges and
1953 # checking them on every accessor. But that is way overkill.)
1954
1955 # We create anonymous subroutines as the accessors and then use
1956 # typeglobs to assign them to the proper package and name
1957
1958 my $name = shift; # Name of the field
1959 my $field = shift; # Reference to the inside-out hash containing the
1960 # field
1961
1962 my $package = (caller)[0];
1963
1964 if (! exists $package_fields{$package}) {
1965 croak "$0: Must call 'setup_package' before 'set_access'";
1966 }
d73e5302 1967
99870f4d
KW
1968 # Stash the field so DESTROY can get it.
1969 $package_fields{$package}{$name} = $field;
cf25bb62 1970
99870f4d
KW
1971 # Remaining arguments are the accessors. For each...
1972 foreach my $access (@_) {
1973 my $access = lc $access;
cf25bb62 1974
99870f4d 1975 my $protected = "";
cf25bb62 1976
99870f4d
KW
1977 # Match the input as far as it goes.
1978 if ($access =~ /^(p[^_]*)_/) {
1979 $protected = $1;
1980 if (substr('protected_', 0, length $protected)
1981 eq $protected)
1982 {
1983
1984 # Add 1 for the underscore not included in $protected
1985 $access = substr($access, length($protected) + 1);
1986 $protected = '_';
1987 }
1988 else {
1989 $protected = "";
1990 }
1991 }
1992
1993 if (substr('addable', 0, length $access) eq $access) {
1994 my $subname = "${package}::${protected}add_$name";
1995 no strict "refs";
1996
1997 # add_ accessor. Don't add if already there, which we
1998 # determine using 'eq' for scalars and '==' otherwise.
1999 *$subname = sub {
2000 use strict "refs";
2001 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
2002 my $self = shift;
2003 my $value = shift;
ffe43484 2004 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2005 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2006 if (ref $value) {
f998e60c 2007 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
2008 }
2009 else {
f998e60c 2010 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 2011 }
f998e60c 2012 push @{$field->{$addr}}, $value;
99870f4d
KW
2013 return;
2014 }
2015 }
2016 elsif (substr('constructor', 0, length $access) eq $access) {
2017 if ($protected) {
2018 Carp::my_carp_bug("Can't set-up 'protected' constructors")
2019 }
2020 else {
2021 $constructor_fields{$package}{$name} = $field;
2022 }
2023 }
2024 elsif (substr('readable_array', 0, length $access) eq $access) {
2025
2026 # Here has read access. If one of the other parameters for
2027 # access is array, or this one specifies array (by being more
2028 # than just 'readable_'), then create a subroutine that
2029 # assumes the data is an array. Otherwise just a scalar
2030 my $subname = "${package}::${protected}$name";
2031 if (grep { /^a/i } @_
2032 or length($access) > length('readable_'))
2033 {
2034 no strict "refs";
2035 *$subname = sub {
2036 use strict "refs";
23e33b60 2037 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 2038 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
2039 if (ref $field->{$addr} ne 'ARRAY') {
2040 my $type = ref $field->{$addr};
2041 $type = 'scalar' unless $type;
2042 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
2043 return;
2044 }
2045 return scalar @{$field->{$addr}} unless wantarray;
2046
2047 # Make a copy; had problems with caller modifying the
2048 # original otherwise
2049 my @return = @{$field->{$addr}};
2050 return @return;
2051 }
2052 }
2053 else {
2054
2055 # Here not an array value, a simpler function.
2056 no strict "refs";
2057 *$subname = sub {
2058 use strict "refs";
23e33b60 2059 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 2060 no overloading;
051df77b 2061 return $field->{pack 'J', $_[0]};
99870f4d
KW
2062 }
2063 }
2064 }
2065 elsif (substr('settable', 0, length $access) eq $access) {
2066 my $subname = "${package}::${protected}set_$name";
2067 no strict "refs";
2068 *$subname = sub {
2069 use strict "refs";
23e33b60
KW
2070 if (main::DEBUG) {
2071 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
2072 Carp::carp_extra_args(\@_) if @_ > 2;
2073 }
2074 # $self is $_[0]; $value is $_[1]
f998e60c 2075 no overloading;
051df77b 2076 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
2077 return;
2078 }
2079 }
2080 else {
2081 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
2082 }
cf25bb62 2083 }
99870f4d 2084 return;
cf25bb62 2085 }
99870f4d
KW
2086}
2087
2088package Input_file;
2089
2090# All input files use this object, which stores various attributes about them,
2091# and provides for convenient, uniform handling. The run method wraps the
2092# processing. It handles all the bookkeeping of opening, reading, and closing
2093# the file, returning only significant input lines.
2094#
2095# Each object gets a handler which processes the body of the file, and is
74cd47d0
KW
2096# called by run(). All character property files must use the generic,
2097# default handler, which has code scrubbed to handle things you might not
2098# expect, including automatic EBCDIC handling. For files that don't deal with
2099# mapping code points to a property value, such as test files,
2100# PropertyAliases, PropValueAliases, and named sequences, you can override the
2101# handler to be a custom one. Such a handler should basically be a
2102# while(next_line()) {...} loop.
99870f4d
KW
2103#
2104# You can also set up handlers to
40045c2d 2105# 0) call during object construction time, after everything else is done
537124e4 2106# 1) call before the first line is read, for pre processing
83b68635
KW
2107# 2) call to adjust each line of the input before the main handler gets
2108# them. This can be automatically generated, if appropriately simple
44b0f8c1 2109# enough, by specifying a Properties parameter in the constructor.
99870f4d 2110# 3) call upon EOF before the main handler exits its loop
537124e4 2111# 4) call at the end, for post processing
99870f4d
KW
2112#
2113# $_ is used to store the input line, and is to be filtered by the
2114# each_line_handler()s. So, if the format of the line is not in the desired
2115# format for the main handler, these are used to do that adjusting. They can
2116# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
44b0f8c1 2117# so the $_ output of one is used as the input to the next. The EOF handler
3848e5b7
KW
2118# is also stackable, but none of the others are, but could easily be changed
2119# to be so.
99870f4d 2120#
a0204bb1
KW
2121# Some properties are used by the Perl core but aren't defined until later
2122# Unicode releases. The perl interpreter would have problems working when
2123# compiled with an earlier Unicode version that doesn't have them, so we need
2124# to define them somehow for those releases. The 'Early' constructor
2125# parameter can be used to automatically handle this. It is essentially
2126# ignored if the Unicode version being compiled has a data file for this
2127# property. Either code to execute or a file to read can be specified.
2128# Details are at the %early definition.
2129#
99870f4d
KW
2130# Most of the handlers can call insert_lines() or insert_adjusted_lines()
2131# which insert the parameters as lines to be processed before the next input
3848e5b7 2132# file line is read. This allows the EOF handler(s) to flush buffers, for
99870f4d
KW
2133# example. The difference between the two routines is that the lines inserted
2134# by insert_lines() are subjected to the each_line_handler()s. (So if you
9be22ca5
KW
2135# called it from such a handler, you would get infinite recursion without some
2136# mechanism to prevent that.) Lines inserted by insert_adjusted_lines() go
2137# directly to the main handler without any adjustments. If the
2138# post-processing handler calls any of these, there will be no effect. Some
2139# error checking for these conditions could be added, but it hasn't been done.
99870f4d
KW
2140#
2141# carp_bad_line() should be called to warn of bad input lines, which clears $_
2142# to prevent further processing of the line. This routine will output the
2143# message as a warning once, and then keep a count of the lines that have the
2144# same message, and output that count at the end of the file's processing.
2145# This keeps the number of messages down to a manageable amount.
2146#
2147# get_missings() should be called to retrieve any @missing input lines.
2148# Messages will be raised if this isn't done if the options aren't to ignore
2149# missings.
2150
2151sub trace { return main::trace(@_); }
2152
99870f4d
KW
2153{ # Closure
2154 # Keep track of fields that are to be put into the constructor.
2155 my %constructor_fields;
2156
2157 main::setup_package(Constructor_Fields => \%constructor_fields);
2158
2159 my %file; # Input file name, required
2160 main::set_access('file', \%file, qw{ c r });
2161
2162 my %first_released; # Unicode version file was first released in, required
2163 main::set_access('first_released', \%first_released, qw{ c r });
2164
2165 my %handler; # Subroutine to process the input file, defaults to
2166 # 'process_generic_property_file'
2167 main::set_access('handler', \%handler, qw{ c });
2168
2169 my %property;
2170 # name of property this file is for. defaults to none, meaning not
2171 # applicable, or is otherwise determinable, for example, from each line.
696609bf 2172 main::set_access('property', \%property, qw{ c r });
99870f4d
KW
2173
2174 my %optional;
3274d896
KW
2175 # This is either an unsigned number, or a list of property names. In the
2176 # former case, if it is non-zero, it means the file is optional, so if the
2177 # file is absent, no warning about that is output. In the latter case, it
2178 # is a list of properties that the file (exclusively) defines. If the
2179 # file is present, tables for those properties will be produced; if
2180 # absent, none will, even if they are listed elsewhere (namely
2181 # PropertyAliases.txt and PropValueAliases.txt) as being in this release,
2182 # and no warnings will be raised about them not being available. (And no
2183 # warning about the file itself will be raised.)
2184 main::set_access('optional', \%optional, qw{ c readable_array } );
99870f4d
KW
2185
2186 my %non_skip;
2187 # This is used for debugging, to skip processing of all but a few input
2188 # files. Add 'non_skip => 1' to the constructor for those files you want
2189 # processed when you set the $debug_skip global.
2190 main::set_access('non_skip', \%non_skip, 'c');
2191
37e2e78e 2192 my %skip;
0dee21d6
KW
2193 # This is used to skip processing of this input file (semi-) permanently.
2194 # The value should be the reason the file is being skipped. It is used
2195 # for files that we aren't planning to process anytime soon, but want to
2196 # allow to be in the directory and be checked for their names not
2197 # conflicting with any other files on a DOS 8.3 name filesystem, but to
2198 # not otherwise be processed, and to not raise a warning about not being
2199 # handled. In the constructor call, any value that evaluates to a numeric
2200 # 0 or undef means don't skip. Any other value is a string giving the
44b0f8c1 2201 # reason it is being skipped, and this will appear in generated pod.
0dee21d6
KW
2202 # However, an empty string reason will suppress the pod entry.
2203 # Internally, calls that evaluate to numeric 0 are changed into undef to
2204 # distinguish them from an empty string call.
2205 main::set_access('skip', \%skip, 'c', 'r');
37e2e78e 2206
99870f4d
KW
2207 my %each_line_handler;
2208 # list of subroutines to look at and filter each non-comment line in the
2209 # file. defaults to none. The subroutines are called in order, each is
2210 # to adjust $_ for the next one, and the final one adjusts it for
2211 # 'handler'
2212 main::set_access('each_line_handler', \%each_line_handler, 'c');
2213
4e058f44
KW
2214 my %retain_trailing_comments;
2215 # This is used to not discard the comments that end data lines. This
2216 # would be used only for files with non-typical syntax, and most code here
2217 # assumes that comments have been stripped, so special handlers would have
2218 # to be written. It is assumed that the code will use these in
2219 # single-quoted contexts, and so any "'" marks in the comment will be
2220 # prefixed by a backslash.
2221 main::set_access('retain_trailing_comments', \%retain_trailing_comments, 'c');
2222
83b68635
KW
2223 my %properties; # Optional ordered list of the properties that occur in each
2224 # meaningful line of the input file. If present, an appropriate
2225 # each_line_handler() is automatically generated and pushed onto the stack
2226 # of such handlers. This is useful when a file contains multiple
44b0f8c1 2227 # properties per line, but no other special considerations are necessary.
83b68635
KW
2228 # The special value "<ignored>" means to discard the corresponding input
2229 # field.
2230 # Any @missing lines in the file should also match this syntax; no such
2231 # files exist as of 6.3. But if it happens in a future release, the code
2232 # could be expanded to properly parse them.
2233 main::set_access('properties', \%properties, qw{ c r });
2234
99870f4d
KW
2235 my %has_missings_defaults;
2236 # ? Are there lines in the file giving default values for code points
2237 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
2238 # the norm, but IGNORED means it has such lines, but the handler doesn't
2239 # use them. Having these three states allows us to catch changes to the
83b68635
KW
2240 # UCD that this program should track. XXX This could be expanded to
2241 # specify the syntax for such lines, like %properties above.
99870f4d
KW
2242 main::set_access('has_missings_defaults',
2243 \%has_missings_defaults, qw{ c r });
2244
40045c2d
KW
2245 my %construction_time_handler;
2246 # Subroutine to call at the end of the new method. If undef, no such
2247 # handler is called.
2248 main::set_access('construction_time_handler',
2249 \%construction_time_handler, qw{ c });
2250
99870f4d
KW
2251 my %pre_handler;
2252 # Subroutine to call before doing anything else in the file. If undef, no
2253 # such handler is called.
2254 main::set_access('pre_handler', \%pre_handler, qw{ c });
2255
2256 my %eof_handler;
3848e5b7 2257 # Subroutines to call upon getting an EOF on the input file, but before
99870f4d
KW
2258 # that is returned to the main handler. This is to allow buffers to be
2259 # flushed. The handler is expected to call insert_lines() or
2260 # insert_adjusted() with the buffered material
3848e5b7 2261 main::set_access('eof_handler', \%eof_handler, qw{ c });
99870f4d
KW
2262
2263 my %post_handler;
2264 # Subroutine to call after all the lines of the file are read in and
9be22ca5
KW
2265 # processed. If undef, no such handler is called. Note that this cannot
2266 # add lines to be processed; instead use eof_handler
99870f4d
KW
2267 main::set_access('post_handler', \%post_handler, qw{ c });
2268
2269 my %progress_message;
2270 # Message to print to display progress in lieu of the standard one
2271 main::set_access('progress_message', \%progress_message, qw{ c });
2272
2273 my %handle;
2274 # cache open file handle, internal. Is undef if file hasn't been
2275 # processed at all, empty if has;
2276 main::set_access('handle', \%handle);
2277
2278 my %added_lines;
2279 # cache of lines added virtually to the file, internal
2280 main::set_access('added_lines', \%added_lines);
2281
74cd47d0
KW
2282 my %remapped_lines;
2283 # cache of lines added virtually to the file, internal
2284 main::set_access('remapped_lines', \%remapped_lines);
2285
99870f4d
KW
2286 my %errors;
2287 # cache of errors found, internal
2288 main::set_access('errors', \%errors);
2289
2290 my %missings;
2291 # storage of '@missing' defaults lines
2292 main::set_access('missings', \%missings);
2293
a0204bb1
KW
2294 my %early;
2295 # Used for properties that must be defined (for Perl's purposes) on
2296 # versions of Unicode earlier than Unicode itself defines them. The
2297 # parameter is an array (it would be better to be a hash, but not worth
2298 # bothering about due to its rare use).
2299 #
2300 # The first element is either a code reference to call when in a release
2301 # earlier than the Unicode file is available in, or it is an alternate
2302 # file to use instead of the non-existent one. This file must have been
2303 # plunked down in the same directory as mktables. Should you be compiling
2304 # on a release that needs such a file, mktables will abort the
2305 # compilation, and tell you where to get the necessary file(s), and what
2306 # name(s) to use to store them as.
2307 # In the case of specifying an alternate file, the array must contain two
2308 # further elements:
2309 #
2310 # [1] is the name of the property that will be generated by this file.
2311 # The class automatically takes the input file and excludes any code
2312 # points in it that were not assigned in the Unicode version being
2313 # compiled. It then uses this result to define the property in the given
2314 # version. Since the property doesn't actually exist in the Unicode
2315 # version being compiled, this should be a name accessible only by core
2316 # perl. If it is the same name as the regular property, the constructor
2317 # will mark the output table as a $PLACEHOLDER so that it doesn't actually
2318 # get output, and so will be unusable by non-core code. Otherwise it gets
2319 # marked as $INTERNAL_ONLY.
2320 #
2321 # [2] is a property value to assign (only when compiling Unicode 1.1.5) to
2322 # the Hangul syllables in that release (which were ripped out in version
2323 # 2) for the given property . (Hence it is ignored except when compiling
2324 # version 1. You only get one value that applies to all of them, which
2325 # may not be the actual reality, but probably nobody cares anyway for
2326 # these obsolete characters.)
2327 #
3d36fe12
KW
2328 # [3] if present is the default value for the property to assign for code
2329 # points not given in the input. If not present, the default from the
2330 # normal property is used
2331 #
4ed2b786
KW
2332 # [-1] If there is an extra final element that is the string 'ONLY_EARLY'.
2333 # it means to not add the name in [1] as an alias to the property name
2334 # used for these. Normally, when compiling Unicode versions that don't
2335 # invoke the early handling, the name is added as a synonym.
cbdbe9d4 2336 #
a0204bb1
KW
2337 # Not all files can be handled in the above way, and so the code ref
2338 # alternative is available. It can do whatever it needs to. The other
2339 # array elements are optional in this case, and the code is free to use or
2340 # ignore them if they are present.
2341 #
2342 # Internally, the constructor unshifts a 0 or 1 onto this array to
2343 # indicate if an early alternative is actually being used or not. This
2344 # makes for easier testing later on.
2345 main::set_access('early', \%early, 'c');
2346
4ed2b786
KW
2347 my %only_early;
2348 main::set_access('only_early', \%only_early, 'c');
2349
0dee21d6
KW
2350 my %required_even_in_debug_skip;
2351 # debug_skip is used to speed up compilation during debugging by skipping
2352 # processing files that are not needed for the task at hand. However,
2353 # some files pretty much can never be skipped, and this is used to specify
2354 # that this is one of them. In order to skip this file, the call to the
2355 # constructor must be edited to comment out this parameter.
2356 main::set_access('required_even_in_debug_skip',
2357 \%required_even_in_debug_skip, 'c');
2358
d8062573
KW
2359 my %withdrawn;
2360 # Some files get removed from the Unicode DB. This is a version object
2361 # giving the first release without this file.
2362 main::set_access('withdrawn', \%withdrawn, 'c');
2363
0dee21d6
KW
2364 my %in_this_release;
2365 # Calculated value from %first_released and %withdrawn. Are we compiling
2366 # a Unicode release which includes this file?
2367 main::set_access('in_this_release', \%in_this_release);
2368
74cd47d0
KW
2369 sub _next_line;
2370 sub _next_line_with_remapped_range;
2371
99870f4d
KW
2372 sub new {
2373 my $class = shift;
2374
2375 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2376 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2377
2378 # Set defaults
2379 $handler{$addr} = \&main::process_generic_property_file;
4e058f44 2380 $retain_trailing_comments{$addr} = 0;
99870f4d 2381 $non_skip{$addr} = 0;
0dee21d6 2382 $skip{$addr} = undef;
99870f4d
KW
2383 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2384 $handle{$addr} = undef;
2385 $added_lines{$addr} = [ ];
74cd47d0 2386 $remapped_lines{$addr} = [ ];
99870f4d 2387 $each_line_handler{$addr} = [ ];
3848e5b7 2388 $eof_handler{$addr} = [ ];
99870f4d
KW
2389 $errors{$addr} = { };
2390 $missings{$addr} = [ ];
a0204bb1 2391 $early{$addr} = [ ];
3274d896 2392 $optional{$addr} = [ ];
99870f4d
KW
2393
2394 # Two positional parameters.
99f78760 2395 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2396 $file{$addr} = main::internal_file_to_platform(shift);
2397 $first_released{$addr} = shift;
2398
2399 # The rest of the arguments are key => value pairs
2400 # %constructor_fields has been set up earlier to list all possible
2401 # ones. Either set or push, depending on how the default has been set
2402 # up just above.
2403 my %args = @_;
2404 foreach my $key (keys %args) {
2405 my $argument = $args{$key};
2406
2407 # Note that the fields are the lower case of the constructor keys
2408 my $hash = $constructor_fields{lc $key};
2409 if (! defined $hash) {
2410 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2411 next;
2412 }
2413 if (ref $hash->{$addr} eq 'ARRAY') {
2414 if (ref $argument eq 'ARRAY') {
2415 foreach my $argument (@{$argument}) {
2416 next if ! defined $argument;
2417 push @{$hash->{$addr}}, $argument;
2418 }
2419 }
2420 else {
2421 push @{$hash->{$addr}}, $argument if defined $argument;
2422 }
2423 }
2424 else {
2425 $hash->{$addr} = $argument;
2426 }
2427 delete $args{$key};
2428 };
2429
0dee21d6
KW
2430 $non_skip{$addr} = 1 if $required_even_in_debug_skip{$addr};
2431
2432 # Convert 0 (meaning don't skip) to undef
2433 undef $skip{$addr} unless $skip{$addr};
2434
3274d896
KW
2435 # Handle the case where this file is optional
2436 my $pod_message_for_non_existent_optional = "";
2437 if ($optional{$addr}->@*) {
2438
2439 # First element is the pod message
2440 $pod_message_for_non_existent_optional
2441 = shift $optional{$addr}->@*;
2442 # Convert a 0 'Optional' argument to an empty list to make later
2443 # code more concise.
2444 if ( $optional{$addr}->@*
2445 && $optional{$addr}->@* == 1
2446 && $optional{$addr}[0] ne ""
2447 && $optional{$addr}[0] !~ /\D/
2448 && $optional{$addr}[0] == 0)
2449 {
2450 $optional{$addr} = [ ];
2451 }
2452 else { # But if the only element doesn't evaluate to 0, make sure
2453 # that this file is indeed considered optional below.
2454 unshift $optional{$addr}->@*, 1;
2455 }
2456 }
2457
0dee21d6 2458 my $progress;
a0204bb1
KW
2459 my $function_instead_of_file = 0;
2460
4ed2b786
KW
2461 if ($early{$addr}->@* && $early{$addr}[-1] eq 'ONLY_EARLY') {
2462 $only_early{$addr} = 1;
2463 pop $early{$addr}->@*;
cbdbe9d4
KW
2464 }
2465
a0204bb1
KW
2466 # If we are compiling a Unicode release earlier than the file became
2467 # available, the constructor may have supplied a substitute
2468 if ($first_released{$addr} gt $v_version && $early{$addr}->@*) {
2469
2470 # Yes, we have a substitute, that we will use; mark it so
2471 unshift $early{$addr}->@*, 1;
2472
2473 # See the definition of %early for what the array elements mean.
4ed2b786
KW
2474 # Note that we have just unshifted onto the array, so the numbers
2475 # below are +1 of those in the %early description.
a0204bb1
KW
2476 # If we have a property this defines, create a table and default
2477 # map for it now (at essentially compile time), so that it will be
2478 # available for the whole of run time. (We will want to add this
2479 # name as an alias when we are using the official property name;
2480 # but this must be deferred until run(), because at construction
2481 # time the official names have yet to be defined.)
2482 if ($early{$addr}[2]) {
2483 my $fate = ($property{$addr}
2484 && $property{$addr} eq $early{$addr}[2])
2485 ? $PLACEHOLDER
2486 : $INTERNAL_ONLY;
2487 my $prop_object = Property->new($early{$addr}[2],
2488 Fate => $fate,
2489 Perl_Extension => 1,
2490 );
2491
3d36fe12
KW
2492 # If not specified by the constructor, use the default mapping
2493 # for the regular property for this substitute one.
4ed2b786
KW
2494 if ($early{$addr}[4]) {
2495 $prop_object->set_default_map($early{$addr}[4]);
3d36fe12
KW
2496 }
2497 elsif ( defined $property{$addr}
2498 && defined $default_mapping{$property{$addr}})
a0204bb1
KW
2499 {
2500 $prop_object
2501 ->set_default_map($default_mapping{$property{$addr}});
2502 }
2503 }
2504
2505 if (ref $early{$addr}[1] eq 'CODE') {
2506 $function_instead_of_file = 1;
2507
2508 # If the first element of the array is a code ref, the others
2509 # are optional.
2510 $handler{$addr} = $early{$addr}[1];
2511 $property{$addr} = $early{$addr}[2]
2512 if defined $early{$addr}[2];
2513 $progress = "substitute $file{$addr}";
2514
2515 undef $file{$addr};
2516 }
2517 else { # Specifying a substitute file
2518
2519 if (! main::file_exists($early{$addr}[1])) {
2520
2521 # If we don't see the substitute file, generate an error
2522 # message giving the needed things, and add it to the list
2523 # of such to output before actual processing happens
2524 # (hence the user finds out all of them in one run).
ce8d64d9
KW
2525 # Instead of creating a general method for NameAliases,
2526 # hard-code it here, as there is unlikely to ever be a
2527 # second one which needs special handling.
2528 my $string_version = ($file{$addr} eq "NameAliases.txt")
2529 ? 'at least 6.1 (the later, the better)'
2530 : sprintf "%vd", $first_released{$addr};
a0204bb1
KW
2531 push @missing_early_files, <<END;
2532'$file{$addr}' version $string_version should be copied to '$early{$addr}[1]'.
2533END
2534 ;
2535 return;
2536 }
2537 $progress = $early{$addr}[1];
2538 $progress .= ", substituting for $file{$addr}" if $file{$addr};
2539 $file{$addr} = $early{$addr}[1];
2540 $property{$addr} = $early{$addr}[2];
0dee21d6 2541
a0204bb1
KW
2542 # Ignore code points not in the version being compiled
2543 push $each_line_handler{$addr}->@*, \&_exclude_unassigned;
2544
2545 if ( $v_version lt v2.0 # Hanguls in this release ...
4ed2b786 2546 && defined $early{$addr}[3]) # ... need special treatment
a0204bb1
KW
2547 {
2548 push $eof_handler{$addr}->@*, \&_fixup_obsolete_hanguls;
2549 }
2550 }
2551
2552 # And this substitute is valid for all releases.
2553 $first_released{$addr} = v0;
2554 }
2555 else { # Normal behavior
0dee21d6 2556 $progress = $file{$addr};
a0204bb1 2557 unshift $early{$addr}->@*, 0; # No substitute
99870f4d
KW
2558 }
2559
0dee21d6
KW
2560 my $file = $file{$addr};
2561 $progress_message{$addr} = "Processing $progress"
2562 unless $progress_message{$addr};
2563
d8062573
KW
2564 # A file should be there if it is within the window of versions for
2565 # which Unicode supplies it
2566 if ($withdrawn{$addr} && $withdrawn{$addr} le $v_version) {
2567 $in_this_release{$addr} = 0;
2568 $skip{$addr} = "";
2569 }
2570 else {
0dee21d6
KW
2571 $in_this_release{$addr} = $first_released{$addr} le $v_version;
2572
a0204bb1
KW
2573 # Check that the file for this object (possibly using a substitute
2574 # for early releases) exists or we have a function alternative
2575 if ( ! $function_instead_of_file
2576 && ! main::file_exists($file))
0dee21d6
KW
2577 {
2578 # Here there is nothing available for this release. This is
2579 # fine if we aren't expecting anything in this release.
2580 if (! $in_this_release{$addr}) {
2581 $skip{$addr} = ""; # Don't remark since we expected
2582 # nothing and got nothing
2583 }
3274d896 2584 elsif ($optional{$addr}->@*) {
0dee21d6 2585
3274d896
KW
2586 # Here the file is optional in this release; Use the
2587 # passed in text to document this case in the pod.
2588 $skip{$addr} = $pod_message_for_non_existent_optional;
0dee21d6
KW
2589 }
2590 elsif ( $in_this_release{$addr}
2591 && ! defined $skip{$addr}
2592 && defined $file)
2593 { # Doesn't exist but should.
2594 $skip{$addr} = "'$file' not found. Possibly Big problems";
2595 Carp::my_carp($skip{$addr});
2596 }
2597 }
2598 elsif ($debug_skip && ! defined $skip{$addr} && ! $non_skip{$addr})
2599 {
2600
2601 # The file exists; if not skipped for another reason, and we are
2602 # skipping most everything during debugging builds, use that as
2603 # the skip reason.
2604 $skip{$addr} = '$debug_skip is on'
2605 }
d8062573 2606 }
0dee21d6
KW
2607
2608 if ( ! $debug_skip
2609 && $non_skip{$addr}
2610 && ! $required_even_in_debug_skip{$addr}
2611 && $verbosity)
2612 {
2613 print "Warning: " . __PACKAGE__ . " constructor for $file has useless 'non_skip' in it\n";
a3a8c5f0 2614 }
99870f4d 2615
0dee21d6 2616 # Here, we have figured out if we will be skipping this file or not.
3274d896
KW
2617 # If so, we add any single property it defines to any passed in
2618 # optional property list. These will be dealt with at run time.
0dee21d6 2619 if (defined $skip{$addr}) {
3274d896
KW
2620 if ($property{$addr}) {
2621 push $optional{$addr}->@*, $property{$addr};
2622 }
2623 } # Otherwise, are going to process the file.
0dee21d6
KW
2624 elsif ($property{$addr}) {
2625
2626 # If the file has a property defined in the constructor for it, it
2627 # means that the property is not listed in the file's entries. So
2628 # add a handler (to the list of line handlers) to insert the
2629 # property name into the lines, to provide a uniform interface to
2630 # the final processing subroutine.
2631 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
09ca89ce 2632 }
83b68635
KW
2633 elsif ($properties{$addr}) {
2634
9be22ca5
KW
2635 # Similarly, there may be more than one property represented on
2636 # each line, with no clue but the constructor input what those
2637 # might be. Add a handler for each line in the input so that it
2638 # creates a separate input line for each property in those input
2639 # lines, thus making them suitable to handle generically.
83b68635
KW
2640
2641 push @{$each_line_handler{$addr}},
2642 sub {
2643 my $file = shift;
2644 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2645
2646 my @fields = split /\s*;\s*/, $_, -1;
2647
2648 if (@fields - 1 > @{$properties{$addr}}) {
2649 $file->carp_bad_line('Extra fields');
2650 $_ = "";
2651 return;
2652 }
2653 my $range = shift @fields; # 0th element is always the
2654 # range
2655
2656 # The next fields in the input line correspond
2657 # respectively to the stored properties.
2658 for my $i (0 .. @{$properties{$addr}} - 1) {
2659 my $property_name = $properties{$addr}[$i];
2660 next if $property_name eq '<ignored>';
2661 $file->insert_adjusted_lines(
2662 "$range; $property_name; $fields[$i]");
2663 }
2664 $_ = "";
2665
2666 return;
2667 };
2668 }
37e2e78e 2669
9be22ca5 2670 { # On non-ascii platforms, we use a special pre-handler
74cd47d0
KW
2671 no strict;
2672 no warnings 'once';
2673 *next_line = (main::NON_ASCII_PLATFORM)
2674 ? *_next_line_with_remapped_range
2675 : *_next_line;
2676 }
2677
40045c2d
KW
2678 &{$construction_time_handler{$addr}}($self)
2679 if $construction_time_handler{$addr};
2680
99870f4d 2681 return $self;
d73e5302
JH
2682 }
2683
cf25bb62 2684
99870f4d
KW
2685 use overload
2686 fallback => 0,
2687 qw("") => "_operator_stringify",
2688 "." => \&main::_operator_dot,
1285127e 2689 ".=" => \&main::_operator_dot_equal,
99870f4d 2690 ;
cf25bb62 2691
99870f4d
KW
2692 sub _operator_stringify {
2693 my $self = shift;
cf25bb62 2694
99870f4d 2695 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2696 }
d73e5302 2697
99870f4d
KW
2698 sub run {
2699 # Process the input object $self. This opens and closes the file and
2700 # calls all the handlers for it. Currently, this can only be called
3848e5b7 2701 # once per file, as it destroy's the EOF handlers
d73e5302 2702
0dee21d6 2703 # flag to make sure extracted files are processed early
f8b69004 2704 state $seen_non_extracted = 0;
0dee21d6 2705
99870f4d
KW
2706 my $self = shift;
2707 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2708
ffe43484 2709 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2710
99870f4d 2711 my $file = $file{$addr};
d73e5302 2712
0dee21d6
KW
2713 if (! $file) {
2714 $handle{$addr} = 'pretend_is_open';
99870f4d
KW
2715 }
2716 else {
f8b69004 2717 if ($seen_non_extracted) {
9be22ca5
KW
2718 if ($file =~ /$EXTRACTED/i) # Some platforms may change the
2719 # case of the file's name
2720 {
1675ea0d 2721 Carp::my_carp_bug(main::join_lines(<<END
99f78760 2722$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2723anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2724have subtle problems
2725END
2726 ));
2727 }
2728 }
2729 elsif ($EXTRACTED_DIR
0dee21d6
KW
2730
2731 # We only do this check for generic property files
2732 && $handler{$addr} == \&main::process_generic_property_file
2733
f8b69004 2734 && $file !~ /$EXTRACTED/i)
99870f4d
KW
2735 {
2736 # We don't set this (by the 'if' above) if we have no
2737 # extracted directory, so if running on an early version,
2738 # this test won't work. Not worth worrying about.
f8b69004 2739 $seen_non_extracted = 1;
99870f4d
KW
2740 }
2741
9be22ca5 2742 # Mark the file as having being processed, and warn if it
99870f4d
KW
2743 # isn't a file we are expecting. As we process the files,
2744 # they are deleted from the hash, so any that remain at the
2745 # end of the program are files that we didn't process.
517956bf 2746 my $fkey = File::Spec->rel2abs($file);
0dee21d6 2747 my $exists = delete $potential_files{lc($fkey)};
faf3cf6b 2748
0dee21d6
KW
2749 Carp::my_carp("Was not expecting '$file'.")
2750 if $exists && ! $in_this_release{$addr};
2751
a0204bb1
KW
2752 # If there is special handling for compiling Unicode releases
2753 # earlier than the first one in which Unicode defines this
2754 # property ...
2755 if ($early{$addr}->@* > 1) {
2756
2757 # Mark as processed any substitute file that would be used in
2758 # such a release
2759 $fkey = File::Spec->rel2abs($early{$addr}[1]);
2760 delete $potential_files{lc($fkey)};
2761
2762 # As commented in the constructor code, when using the
2763 # official property, we still have to allow the publicly
2764 # inaccessible early name so that the core code which uses it
2765 # will work regardless.
4ed2b786
KW
2766 if ( ! $only_early{$addr}
2767 && ! $early{$addr}[0]
2768 && $early{$addr}->@* > 2)
2769 {
a0204bb1 2770 my $early_property_name = $early{$addr}[2];
4ed2b786 2771 if ($property{$addr} ne $early_property_name) {
a0204bb1
KW
2772 main::property_ref($property{$addr})
2773 ->add_alias($early_property_name);
2774 }
2775 }
2776 }
2777
0dee21d6
KW
2778 # We may be skipping this file ...
2779 if (defined $skip{$addr}) {
2780
2781 # If the file isn't supposed to be in this release, there is
2782 # nothing to do
2783 if ($in_this_release{$addr}) {
2784
2785 # But otherwise, we may print a message
2786 if ($debug_skip) {
2787 print STDERR "Skipping input file '$file'",
2788 " because '$skip{$addr}'\n";
2789 }
2790
2791 # And add it to the list of skipped files, which is later
2792 # used to make the pod
2793 $skipped_files{$file} = $skip{$addr};
3274d896
KW
2794
2795 # The 'optional' list contains properties that are also to
2796 # be skipped along with the file. (There may also be
2797 # digits which are just placeholders to make sure it isn't
2798 # an empty list
2799 foreach my $property ($optional{$addr}->@*) {
2800 next unless $property =~ /\D/;
2801 my $prop_object = main::property_ref($property);
2802 next unless defined $prop_object;
2803 $prop_object->set_fate($SUPPRESSED, $skip{$addr});
2804 }
0dee21d6 2805 }
99870f4d 2806
37e2e78e
KW
2807 return;
2808 }
2809
9be22ca5
KW
2810 # Here, we are going to process the file. Open it, converting the
2811 # slashes used in this program into the proper form for the OS
99870f4d
KW
2812 my $file_handle;
2813 if (not open $file_handle, "<", $file) {
2814 Carp::my_carp("Can't open $file. Skipping: $!");
d3664eb9 2815 return;
99870f4d
KW
2816 }
2817 $handle{$addr} = $file_handle; # Cache the open file handle
9e65c3f4 2818
0dee21d6
KW
2819 # If possible, make sure that the file is the correct version.
2820 # (This data isn't available on early Unicode releases or in
2821 # UnicodeData.txt.) We don't do this check if we are using a
2822 # substitute file instead of the official one (though the code
2823 # could be extended to do so).
2824 if ($in_this_release{$addr}
a0204bb1 2825 && ! $early{$addr}[0]
0dee21d6
KW
2826 && lc($file) ne 'unicodedata.txt')
2827 {
96f226dc 2828 if ($file !~ /^Unihan/i) {
0dee21d6
KW
2829
2830 # The non-Unihan files started getting version numbers in
2831 # 3.2, but some files in 4.0 are unchanged from 3.2, and
2832 # marked as 3.2. 4.0.1 is the first version where there
2833 # are no files marked as being from less than 4.0, though
2834 # some are marked as 4.0. In versions after that, the
2835 # numbers are correct.
2836 if ($v_version ge v4.0.1) {
2837 $_ = <$file_handle>; # The version number is in the
2838 # very first line
2839 if ($_ !~ / - $string_version \. /x) {
2840 chomp;
2841 $_ =~ s/^#\s*//;
2842
2843 # 4.0.1 had some valid files that weren't updated.
2844 if (! ($v_version eq v4.0.1 && $_ =~ /4\.0\.0/)) {
2845 die Carp::my_carp("File '$file' is version "
2846 . "'$_'. It should be "
2847 . "version $string_version");
2848 }
2849 }
cafe9cf0 2850 }
96f226dc 2851 }
0dee21d6
KW
2852 elsif ($v_version ge v6.0.0) { # Unihan
2853
2854 # Unihan files didn't get accurate version numbers until
2855 # 6.0. The version is somewhere in the first comment
2856 # block
96f226dc
KW
2857 while (<$file_handle>) {
2858 if ($_ !~ /^#/) {
0dee21d6
KW
2859 Carp::my_carp_bug("Could not find the expected "
2860 . "version info in file '$file'");
96f226dc
KW
2861 last;
2862 }
2863 chomp;
2864 $_ =~ s/^#\s*//;
2865 next if $_ !~ / version: /x;
2866 last if $_ =~ /$string_version/;
0dee21d6
KW
2867 die Carp::my_carp("File '$file' is version "
2868 . "'$_'. It should be "
2869 . "version $string_version");
96f226dc
KW
2870 }
2871 }
9e65c3f4 2872 }
99870f4d
KW
2873 }
2874
0dee21d6 2875 print "$progress_message{$addr}\n" if $verbosity >= $PROGRESS;
99870f4d
KW
2876
2877 # Call any special handler for before the file.
2878 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2879
2880 # Then the main handler
2881 &{$handler{$addr}}($self);
2882
2883 # Then any special post-file handler.
2884 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2885
2886 # If any errors have been accumulated, output the counts (as the first
2887 # error message in each class was output when it was encountered).
2888 if ($errors{$addr}) {
2889 my $total = 0;
2890 my $types = 0;
2891 foreach my $error (keys %{$errors{$addr}}) {
2892 $total += $errors{$addr}->{$error};
2893 delete $errors{$addr}->{$error};
2894 $types++;
2895 }
2896 if ($total > 1) {
2897 my $message
2898 = "A total of $total lines had errors in $file. ";
2899
2900 $message .= ($types == 1)
2901 ? '(Only the first one was displayed.)'
2902 : '(Only the first of each type was displayed.)';
2903 Carp::my_carp($message);
2904 }
2905 }
2906
2907 if (@{$missings{$addr}}) {
2908 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2909 }
2910
2911 # If a real file handle, close it.
2912 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2913 ref $handle{$addr};
2914 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2915 # the file, as opposed to undef
2916 return;
2917 }
2918
74cd47d0 2919 sub _next_line {
99870f4d
KW
2920 # Sets $_ to be the next logical input line, if any. Returns non-zero
2921 # if such a line exists. 'logical' means that any lines that have
2922 # been added via insert_lines() will be returned in $_ before the file
2923 # is read again.
2924
2925 my $self = shift;
2926 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2927
ffe43484 2928 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2929
2930 # Here the file is open (or if the handle is not a ref, is an open
2931 # 'virtual' file). Get the next line; any inserted lines get priority
2932 # over the file itself.
2933 my $adjusted;
2934
2935 LINE:
2936 while (1) { # Loop until find non-comment, non-empty line
2937 #local $to_trace = 1 if main::DEBUG;
2938 my $inserted_ref = shift @{$added_lines{$addr}};
2939 if (defined $inserted_ref) {
2940 ($adjusted, $_) = @{$inserted_ref};
2941 trace $adjusted, $_ if main::DEBUG && $to_trace;
2942 return 1 if $adjusted;
2943 }
2944 else {
2945 last if ! ref $handle{$addr}; # Don't read unless is real file
2946 last if ! defined ($_ = readline $handle{$addr});
2947 }
2948 chomp;
2949 trace $_ if main::DEBUG && $to_trace;
2950
2951 # See if this line is the comment line that defines what property
2952 # value that code points that are not listed in the file should
2953 # have. The format or existence of these lines is not guaranteed
2954 # by Unicode since they are comments, but the documentation says
2955 # that this was added for machine-readability, so probably won't
2956 # change. This works starting in Unicode Version 5.0. They look
2957 # like:
2958 #
2959 # @missing: 0000..10FFFF; Not_Reordered
2960 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2961 # @missing: 0000..10FFFF; ; NaN
2962 #
2963 # Save the line for a later get_missings() call.
2964 if (/$missing_defaults_prefix/) {
2965 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2966 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2967 }
2968 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2969 my @defaults = split /\s* ; \s*/x, $_;
2970
2971 # The first field is the @missing, which ends in a
2972 # semi-colon, so can safely shift.
2973 shift @defaults;
2974
2975 # Some of these lines may have empty field placeholders
2976 # which get in the way. An example is:
2977 # @missing: 0000..10FFFF; ; NaN
2978 # Remove them. Process starting from the top so the
2979 # splice doesn't affect things still to be looked at.
2980 for (my $i = @defaults - 1; $i >= 0; $i--) {
2981 next if $defaults[$i] ne "";
2982 splice @defaults, $i, 1;
2983 }
2984
2985 # What's left should be just the property (maybe) and the
2986 # default. Having only one element means it doesn't have
2987 # the property.
2988 my $default;
2989 my $property;
2990 if (@defaults >= 1) {
2991 if (@defaults == 1) {
2992 $default = $defaults[0];
2993 }
2994 else {
2995 $property = $defaults[0];
2996 $default = $defaults[1];
2997 }
2998 }
2999
3000 if (@defaults < 1
3001 || @defaults > 2
3002 || ($default =~ /^</
3003 && $default !~ /^<code *point>$/i
09f8d0ac
KW
3004 && $default !~ /^<none>$/i
3005 && $default !~ /^<script>$/i))
99870f4d
KW
3006 {
3007 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
3008 }
3009 else {
3010
3011 # If the property is missing from the line, it should
3012 # be the one for the whole file
3013 $property = $property{$addr} if ! defined $property;
3014
3015 # Change <none> to the null string, which is what it
3016 # really means. If the default is the code point
3017 # itself, set it to <code point>, which is what
3018 # Unicode uses (but sometimes they've forgotten the
3019 # space)
3020 if ($default =~ /^<none>$/i) {
3021 $default = "";
3022 }
3023 elsif ($default =~ /^<code *point>$/i) {
3024 $default = $CODE_POINT;
3025 }
09f8d0ac
KW
3026 elsif ($default =~ /^<script>$/i) {
3027
3028 # Special case this one. Currently is from
3029 # ScriptExtensions.txt, and means for all unlisted
3030 # code points, use their Script property values.
3031 # For the code points not listed in that file, the
3032 # default value is 'Unknown'.
3033 $default = "Unknown";
3034 }
99870f4d
KW
3035
3036 # Store them as a sub-arrays with both components.
3037 push @{$missings{$addr}}, [ $default, $property ];
3038 }
3039 }
3040
3041 # There is nothing for the caller to process on this comment
3042 # line.
3043 next;
3044 }
3045
4e058f44
KW
3046 # Unless to keep, remove comments. If to keep, ignore
3047 # comment-only lines
3048 if ($retain_trailing_comments{$addr}) {
3049 next if / ^ \s* \# /x;
3050
3051 # But escape any single quotes (done in both the comment and
3052 # non-comment portion; this could be a bug someday, but not
3053 # likely)
3054 s/'/\\'/g;
3055 }
3056 else {
3057 s/#.*//;
3058 }
3059
3060 # Remove trailing space, and skip this line if the result is empty
99870f4d
KW
3061 s/\s+$//;
3062 next if /^$/;
3063
3064 # Call any handlers for this line, and skip further processing of
3065 # the line if the handler sets the line to null.
3066 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
3067 &{$sub_ref}($self);
3068 next LINE if /^$/;
3069 }
3070
3071 # Here the line is ok. return success.
3072 return 1;
3073 } # End of looping through lines.
3074
3848e5b7 3075 # If there are EOF handlers, call each (only once) and if it generates
99870f4d 3076 # more lines to process go back in the loop to handle them.
3848e5b7
KW
3077 while ($eof_handler{$addr}->@*) {
3078 &{$eof_handler{$addr}[0]}($self);
3079 shift $eof_handler{$addr}->@*; # Currently only get one shot at it.
99870f4d
KW
3080 goto LINE if $added_lines{$addr};
3081 }
3082
3083 # Return failure -- no more lines.
3084 return 0;
3085
3086 }
3087
74cd47d0
KW
3088 sub _next_line_with_remapped_range {
3089 my $self = shift;
3090 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3091
3092 # like _next_line(), but for use on non-ASCII platforms. It sets $_
3093 # to be the next logical input line, if any. Returns non-zero if such
3094 # a line exists. 'logical' means that any lines that have been added
3095 # via insert_lines() will be returned in $_ before the file is read
3096 # again.
3097 #
3098 # The difference from _next_line() is that this remaps the Unicode
3099 # code points in the input to those of the native platform. Each
3100 # input line contains a single code point, or a single contiguous
3101 # range of them This routine splits each range into its individual
3102 # code points and caches them. It returns the cached values,
3103 # translated into their native equivalents, one at a time, for each
3104 # call, before reading the next line. Since native values can only be
3105 # a single byte wide, no translation is needed for code points above
3106 # 0xFF, and ranges that are entirely above that number are not split.
3107 # If an input line contains the range 254-1000, it would be split into
3108 # three elements: 254, 255, and 256-1000. (The downstream table
3109 # insertion code will sort and coalesce the individual code points
3110 # into appropriate ranges.)
3111
3112 my $addr = do { no overloading; pack 'J', $self; };
3113
3114 while (1) {
3115
3116 # Look in cache before reading the next line. Return any cached
3117 # value, translated
3118 my $inserted = shift @{$remapped_lines{$addr}};
3119 if (defined $inserted) {
3120 trace $inserted if main::DEBUG && $to_trace;
3121 $_ = $inserted =~ s/^ ( \d+ ) /sprintf("%04X", utf8::unicode_to_native($1))/xer;
3122 trace $_ if main::DEBUG && $to_trace;
3123 return 1;
3124 }
3125
3126 # Get the next line.
3127 return 0 unless _next_line($self);
3128
3129 # If there is a special handler for it, return the line,
3130 # untranslated. This should happen only for files that are
3131 # special, not being code-point related, such as property names.
3132 return 1 if $handler{$addr}
3133 != \&main::process_generic_property_file;
3134
3135 my ($range, $property_name, $map, @remainder)
3136 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3137
3138 if (@remainder
3139 || ! defined $property_name
3140 || $range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3141 {
3142 Carp::my_carp_bug("Unrecognized input line '$_'. Ignored");
3143 }
3144
3145 my $low = hex $1;
3146 my $high = (defined $2) ? hex $2 : $low;
3147
3148 # If the input maps the range to another code point, remap the
3149 # target if it is between 0 and 255.
3150 my $tail;
3151 if (defined $map) {
3152 $map =~ s/\b 00 ( [0-9A-F]{2} ) \b/sprintf("%04X", utf8::unicode_to_native(hex $1))/gxe;
3153 $tail = "$property_name; $map";
3154 $_ = "$range; $tail";
3155 }
3156 else {
3157 $tail = $property_name;
3158 }
3159
3160 # If entire range is above 255, just return it, unchanged (except
3161 # any mapped-to code point, already changed above)
3162 return 1 if $low > 255;
3163
3164 # Cache an entry for every code point < 255. For those in the
3165 # range above 255, return a dummy entry for just that portion of
3166 # the range. Note that this will be out-of-order, but that is not
3167 # a problem.
3168 foreach my $code_point ($low .. $high) {
3169 if ($code_point > 255) {
3170 $_ = sprintf "%04X..%04X; $tail", $code_point, $high;
3171 return 1;
3172 }
3173 push @{$remapped_lines{$addr}}, "$code_point; $tail";
3174 }
3175 } # End of looping through lines.
3176
3177 # NOTREACHED
3178 }
3179
99870f4d
KW
3180# Not currently used, not fully tested.
3181# sub peek {
f67a5002 3182# # Non-destructive lookahead one non-adjusted, non-comment, non-blank
99870f4d
KW
3183# # record. Not callable from an each_line_handler(), nor does it call
3184# # an each_line_handler() on the line.
3185#
3186# my $self = shift;
ffe43484 3187# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3188#
3189# foreach my $inserted_ref (@{$added_lines{$addr}}) {
3190# my ($adjusted, $line) = @{$inserted_ref};
3191# next if $adjusted;
3192#
3193# # Remove comments and trailing space, and return a non-empty
3194# # resulting line
3195# $line =~ s/#.*//;
3196# $line =~ s/\s+$//;
3197# return $line if $line ne "";
3198# }
3199#
3200# return if ! ref $handle{$addr}; # Don't read unless is real file
3201# while (1) { # Loop until find non-comment, non-empty line
3202# local $to_trace = 1 if main::DEBUG;
3203# trace $_ if main::DEBUG && $to_trace;
3204# return if ! defined (my $line = readline $handle{$addr});
3205# chomp $line;
3206# push @{$added_lines{$addr}}, [ 0, $line ];
3207#
3208# $line =~ s/#.*//;
3209# $line =~ s/\s+$//;
3210# return $line if $line ne "";
3211# }
3212#
3213# return;
3214# }
3215
3216
3217 sub insert_lines {
3218 # Lines can be inserted so that it looks like they were in the input
3219 # file at the place it was when this routine is called. See also
3220 # insert_adjusted_lines(). Lines inserted via this routine go through
3221 # any each_line_handler()
3222
3223 my $self = shift;
3224
3225 # Each inserted line is an array, with the first element being 0 to
3226 # indicate that this line hasn't been adjusted, and needs to be
3227 # processed.
f998e60c 3228 no overloading;
051df77b 3229 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
3230 return;
3231 }
3232
3233 sub insert_adjusted_lines {
3234 # Lines can be inserted so that it looks like they were in the input
3235 # file at the place it was when this routine is called. See also
3236 # insert_lines(). Lines inserted via this routine are already fully
3237 # adjusted, ready to be processed; each_line_handler()s handlers will
3238 # not be called. This means this is not a completely general
3239 # facility, as only the last each_line_handler on the stack should
3240 # call this. It could be made more general, by passing to each of the
3241 # line_handlers their position on the stack, which they would pass on
3242 # to this routine, and that would replace the boolean first element in
3243 # the anonymous array pushed here, so that the next_line routine could
3244 # use that to call only those handlers whose index is after it on the
3245 # stack. But this is overkill for what is needed now.
3246
3247 my $self = shift;
3248 trace $_[0] if main::DEBUG && $to_trace;
3249
3250 # Each inserted line is an array, with the first element being 1 to
3251 # indicate that this line has been adjusted
f998e60c 3252 no overloading;
051df77b 3253 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
3254 return;
3255 }
3256
3257 sub get_missings {
3258 # Returns the stored up @missings lines' values, and clears the list.
3259 # The values are in an array, consisting of the default in the first
3260 # element, and the property in the 2nd. However, since these lines
3261 # can be stacked up, the return is an array of all these arrays.
3262
3263 my $self = shift;
3264 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3265
ffe43484 3266 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3267
3268 # If not accepting a list return, just return the first one.
3269 return shift @{$missings{$addr}} unless wantarray;
3270
3271 my @return = @{$missings{$addr}};
3272 undef @{$missings{$addr}};
3273 return @return;
3274 }
3275
a0204bb1
KW
3276 sub _exclude_unassigned {
3277
3278 # Takes the range in $_ and excludes code points that aren't assigned
3279 # in this release
3280
3281 state $skip_inserted_count = 0;
3282
3283 # Ignore recursive calls.
3284 if ($skip_inserted_count) {
3285 $skip_inserted_count--;
3286 return;
3287 }
3288
3289 # Find what code points are assigned in this release
3290 main::calculate_Assigned() if ! defined $Assigned;
3291
3292 my $self = shift;
3293 my $addr = do { no overloading; pack 'J', $self; };
3294 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3295
3296 my ($range, @remainder)
3297 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3298
3299 # Examine the range.
3300 if ($range =~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3301 {
3302 my $low = hex $1;
3303 my $high = (defined $2) ? hex $2 : $low;
3304
3305 # Split the range into subranges of just those code points in it
3306 # that are assigned.
3307 my @ranges = (Range_List->new(Initialize
3308 => Range->new($low, $high)) & $Assigned)->ranges;
3309
3310 # Do nothing if nothing in the original range is assigned in this
3311 # release; handle normally if everything is in this release.
3312 if (! @ranges) {
3313 $_ = "";
3314 }
3315 elsif (@ranges != 1) {
3316
3317 # Here, some code points in the original range aren't in this
3318 # release; @ranges gives the ones that are. Create fake input
3319 # lines for each of the ranges, and set things up so that when
3320 # this routine is called on that fake input, it will do
3321 # nothing.
3322 $skip_inserted_count = @ranges;
3323 my $remainder = join ";", @remainder;
3324 for my $range (@ranges) {
3325 $self->insert_lines(sprintf("%04X..%04X;%s",
3326 $range->start, $range->end, $remainder));
3327 }
3328 $_ = ""; # The original range is now defunct.
3329 }
3330 }
3331
3332 return;
3333 }
3334
3335 sub _fixup_obsolete_hanguls {
3336
3337 # This is called only when compiling Unicode version 1. All Unicode
3338 # data for subsequent releases assumes that the code points that were
3339 # Hangul syllables in this release only are something else, so if
3340 # using such data, we have to override it
3341
3342 my $self = shift;
3343 my $addr = do { no overloading; pack 'J', $self; };
3344 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3345
3346 my $object = main::property_ref($property{$addr});
53adf6a2
KW
3347 $object->add_map($FIRST_REMOVED_HANGUL_SYLLABLE,
3348 $FINAL_REMOVED_HANGUL_SYLLABLE,
a0204bb1
KW
3349 $early{$addr}[3], # Passed-in value for these
3350 Replace => $UNCONDITIONALLY);
3351 }
3352
99870f4d
KW
3353 sub _insert_property_into_line {
3354 # Add a property field to $_, if this file requires it.
3355
f998e60c 3356 my $self = shift;
ffe43484 3357 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 3358 my $property = $property{$addr};
99870f4d
KW
3359 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3360
3361 $_ =~ s/(;|$)/; $property$1/;
3362 return;
3363 }
3364
3365 sub carp_bad_line {
3366 # Output consistent error messages, using either a generic one, or the
3367 # one given by the optional parameter. To avoid gazillions of the
3368 # same message in case the syntax of a file is way off, this routine
3369 # only outputs the first instance of each message, incrementing a
3370 # count so the totals can be output at the end of the file.
3371
3372 my $self = shift;
3373 my $message = shift;
3374 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3375
ffe43484 3376 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3377
3378 $message = 'Unexpected line' unless $message;
3379
3380 # No trailing punctuation so as to fit with our addenda.
3381 $message =~ s/[.:;,]$//;
3382
3383 # If haven't seen this exact message before, output it now. Otherwise
3384 # increment the count of how many times it has occurred
3385 unless ($errors{$addr}->{$message}) {
3386 Carp::my_carp("$message in '$_' in "
f998e60c 3387 . $file{$addr}
99870f4d
KW
3388 . " at line $.. Skipping this line;");
3389 $errors{$addr}->{$message} = 1;
3390 }
3391 else {
3392 $errors{$addr}->{$message}++;
3393 }
3394
3395 # Clear the line to prevent any further (meaningful) processing of it.
3396 $_ = "";
3397
3398 return;
3399 }
3400} # End closure
3401
3402package Multi_Default;
3403
3404# Certain properties in early versions of Unicode had more than one possible
3405# default for code points missing from the files. In these cases, one
3406# default applies to everything left over after all the others are applied,
3407# and for each of the others, there is a description of which class of code
3408# points applies to it. This object helps implement this by storing the
3409# defaults, and for all but that final default, an eval string that generates
3410# the class that it applies to.
3411
3412
3413{ # Closure
3414
3415 main::setup_package();
3416
3417 my %class_defaults;
3418 # The defaults structure for the classes
3419 main::set_access('class_defaults', \%class_defaults);
3420
3421 my %other_default;
3422 # The default that applies to everything left over.
3423 main::set_access('other_default', \%other_default, 'r');
3424
3425
3426 sub new {
3427 # The constructor is called with default => eval pairs, terminated by
3428 # the left-over default. e.g.
3429 # Multi_Default->new(
3430 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
3431 # - 0x200D',
3432 # 'R' => 'some other expression that evaluates to code points',
3433 # .
3434 # .
3435 # .
3436 # 'U'));
68c3923f
KW
3437 # It is best to leave the final value be the one that matches the
3438 # above-Unicode code points.
99870f4d
KW
3439
3440 my $class = shift;
3441
3442 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 3443 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3444
3445 while (@_ > 1) {
3446 my $default = shift;
3447 my $eval = shift;
3448 $class_defaults{$addr}->{$default} = $eval;
3449 }
3450
3451 $other_default{$addr} = shift;
3452
3453 return $self;
3454 }
3455
3456 sub get_next_defaults {
3457 # Iterates and returns the next class of defaults.
3458 my $self = shift;
3459 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3460
ffe43484 3461 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3462
3463 return each %{$class_defaults{$addr}};
3464 }
3465}
3466
3467package Alias;
3468
3469# An alias is one of the names that a table goes by. This class defines them
3470# including some attributes. Everything is currently setup in the
3471# constructor.
3472
3473
3474{ # Closure
3475
3476 main::setup_package();
3477
3478 my %name;
3479 main::set_access('name', \%name, 'r');
3480
3481 my %loose_match;
c12f2655 3482 # Should this name match loosely or not.
99870f4d
KW
3483 main::set_access('loose_match', \%loose_match, 'r');
3484
33e96e72
KW
3485 my %make_re_pod_entry;
3486 # Some aliases should not get their own entries in the re section of the
3487 # pod, because they are covered by a wild-card, and some we want to
3488 # discourage use of. Binary
f82fe4ba 3489 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r', 's');
99870f4d 3490
fd1e3e84
KW
3491 my %ucd;
3492 # Is this documented to be accessible via Unicode::UCD
3493 main::set_access('ucd', \%ucd, 'r', 's');
3494
99870f4d
KW
3495 my %status;
3496 # Aliases have a status, like deprecated, or even suppressed (which means
3497 # they don't appear in documentation). Enum
3498 main::set_access('status', \%status, 'r');
3499
0eac1e20 3500 my %ok_as_filename;
99870f4d
KW
3501 # Similarly, some aliases should not be considered as usable ones for
3502 # external use, such as file names, or we don't want documentation to
3503 # recommend them. Boolean
0eac1e20 3504 main::set_access('ok_as_filename', \%ok_as_filename, 'r');
99870f4d
KW
3505
3506 sub new {
3507 my $class = shift;
3508
3509 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3510 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3511
3512 $name{$addr} = shift;
3513 $loose_match{$addr} = shift;
33e96e72 3514 $make_re_pod_entry{$addr} = shift;
0eac1e20 3515 $ok_as_filename{$addr} = shift;
99870f4d 3516 $status{$addr} = shift;
fd1e3e84 3517 $ucd{$addr} = shift;
99870f4d
KW
3518
3519 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3520
3521 # Null names are never ok externally
0eac1e20 3522 $ok_as_filename{$addr} = 0 if $name{$addr} eq "";
99870f4d
KW
3523
3524 return $self;
3525 }
3526}
3527
3528package Range;
3529
3530# A range is the basic unit for storing code points, and is described in the
3531# comments at the beginning of the program. Each range has a starting code
3532# point; an ending code point (not less than the starting one); a value
3533# that applies to every code point in between the two end-points, inclusive;
3534# and an enum type that applies to the value. The type is for the user's
3535# convenience, and has no meaning here, except that a non-zero type is
3536# considered to not obey the normal Unicode rules for having standard forms.
3537#
3538# The same structure is used for both map and match tables, even though in the
3539# latter, the value (and hence type) is irrelevant and could be used as a
3540# comment. In map tables, the value is what all the code points in the range
3541# map to. Type 0 values have the standardized version of the value stored as
3542# well, so as to not have to recalculate it a lot.
3543
3544sub trace { return main::trace(@_); }
3545
3546{ # Closure
3547
3548 main::setup_package();
3549
3550 my %start;
3551 main::set_access('start', \%start, 'r', 's');
3552
3553 my %end;
3554 main::set_access('end', \%end, 'r', 's');
3555
3556 my %value;
eeeafd9c 3557 main::set_access('value', \%value, 'r', 's');
99870f4d
KW
3558
3559 my %type;
3560 main::set_access('type', \%type, 'r');
3561
3562 my %standard_form;
3563 # The value in internal standard form. Defined only if the type is 0.
3564 main::set_access('standard_form', \%standard_form);
3565
3566 # Note that if these fields change, the dump() method should as well
3567
3568 sub new {
3569 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
3570 my $class = shift;
3571
3572 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3573 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3574
3575 $start{$addr} = shift;
3576 $end{$addr} = shift;
3577
3578 my %args = @_;
3579
3580 my $value = delete $args{'Value'}; # Can be 0
3581 $value = "" unless defined $value;
3582 $value{$addr} = $value;
3583
3584 $type{$addr} = delete $args{'Type'} || 0;
3585
3586 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3587
99870f4d
KW
3588 return $self;
3589 }
3590
3591 use overload
3592 fallback => 0,
3593 qw("") => "_operator_stringify",
3594 "." => \&main::_operator_dot,
1285127e 3595 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3596 ;
3597
3598 sub _operator_stringify {
3599 my $self = shift;
ffe43484 3600 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3601
3602 # Output it like '0041..0065 (value)'
3603 my $return = sprintf("%04X", $start{$addr})
3604 . '..'
3605 . sprintf("%04X", $end{$addr});
3606 my $value = $value{$addr};
3607 my $type = $type{$addr};
3608 $return .= ' (';
3609 $return .= "$value";
3610 $return .= ", Type=$type" if $type != 0;
3611 $return .= ')';
3612
3613 return $return;
3614 }
3615
3616 sub standard_form {
c292d35a
NC
3617 # Calculate the standard form only if needed, and cache the result.
3618 # The standard form is the value itself if the type is special.
3619 # This represents a considerable CPU and memory saving - at the time
3620 # of writing there are 368676 non-special objects, but the standard
3621 # form is only requested for 22047 of them - ie about 6%.
99870f4d
KW
3622
3623 my $self = shift;
3624 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3625
ffe43484 3626 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3627
3628 return $standard_form{$addr} if defined $standard_form{$addr};
c292d35a
NC
3629
3630 my $value = $value{$addr};
3631 return $value if $type{$addr};
3632 return $standard_form{$addr} = main::standardize($value);
99870f4d
KW
3633 }
3634
3635 sub dump {
3636 # Human, not machine readable. For machine readable, comment out this
3637 # entire routine and let the standard one take effect.
3638 my $self = shift;
3639 my $indent = shift;
3640 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3641
ffe43484 3642 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3643
3644 my $return = $indent
3645 . sprintf("%04X", $start{$addr})
3646 . '..'
3647 . sprintf("%04X", $end{$addr})
3648 . " '$value{$addr}';";
3649 if (! defined $standard_form{$addr}) {
3650 $return .= "(type=$type{$addr})";
3651 }
3652 elsif ($standard_form{$addr} ne $value{$addr}) {
3653 $return .= "(standard '$standard_form{$addr}')";
3654 }
3655 return $return;
3656 }
3657} # End closure
3658
3659package _Range_List_Base;
3660
3661# Base class for range lists. A range list is simply an ordered list of
3662# ranges, so that the ranges with the lowest starting numbers are first in it.
3663#
3664# When a new range is added that is adjacent to an existing range that has the
3665# same value and type, it merges with it to form a larger range.
3666#
3667# Ranges generally do not overlap, except that there can be multiple entries
3668# of single code point ranges. This is because of NameAliases.txt.
3669#
3670# In this program, there is a standard value such that if tw