This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
mktables: Move code to common functions
[perl5.git] / lib / unicore / mktables
CommitLineData
d73e5302 1#!/usr/bin/perl -w
99870f4d
KW
2
3# !!!!!!!!!!!!!! IF YOU MODIFY THIS FILE !!!!!!!!!!!!!!!!!!!!!!!!!
4# Any files created or read by this program should be listed in 'mktables.lst'
5# Use -makelist to regenerate it.
6
9be22ca5
KW
7# There was an attempt when this was first rewritten to make it 5.8
8# compatible, but that has now been abandoned, and newer constructs are used
9# as convenient.
6c68572b 10
cdcef19a 11my $start_time;
98dc9551 12BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
13 # get it as close as possible
14 $start_time= time;
15}
16
23e33b60 17require 5.010_001;
d73e5302 18use strict;
99870f4d 19use warnings;
cf25bb62 20use Carp;
bd9ebcfd 21use Config;
99870f4d
KW
22use File::Find;
23use File::Path;
d07a55ed 24use File::Spec;
99870f4d 25use Text::Tabs;
6b64c11c 26use re "/aa";
0dee21d6 27use feature 'state';
99870f4d
KW
28
29sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 30my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d 31
74cd47d0
KW
32sub NON_ASCII_PLATFORM { ord("A") != 65 }
33
99870f4d
KW
34##########################################################################
35#
36# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
37# from the Unicode database files (lib/unicore/.../*.txt), It also generates
232ed87f 38# a pod file and .t files, depending on option parameters.
99870f4d
KW
39#
40# The structure of this file is:
41# First these introductory comments; then
42# code needed for everywhere, such as debugging stuff; then
43# code to handle input parameters; then
44# data structures likely to be of external interest (some of which depend on
45# the input parameters, so follows them; then
46# more data structures and subroutine and package (class) definitions; then
47# the small actual loop to process the input files and finish up; then
48# a __DATA__ section, for the .t tests
49#
232ed87f
KW
50# This program works on all releases of Unicode so far. The outputs have been
51# scrutinized most intently for release 5.1. The others have been checked for
52# somewhat more than just sanity. It can handle all non-provisional Unicode
53# character properties in those releases.
99870f4d 54#
99870f4d
KW
55# This program is mostly about Unicode character (or code point) properties.
56# A property describes some attribute or quality of a code point, like if it
57# is lowercase or not, its name, what version of Unicode it was first defined
58# in, or what its uppercase equivalent is. Unicode deals with these disparate
59# possibilities by making all properties into mappings from each code point
60# into some corresponding value. In the case of it being lowercase or not,
61# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
62# property maps each Unicode code point to a single value, called a "property
232ed87f
KW
63# value". (Some more recently defined properties, map a code point to a set
64# of values.)
99870f4d
KW
65#
66# When using a property in a regular expression, what is desired isn't the
67# mapping of the code point to its property's value, but the reverse (or the
68# mathematical "inverse relation"): starting with the property value, "Does a
69# code point map to it?" These are written in a "compound" form:
70# \p{property=value}, e.g., \p{category=punctuation}. This program generates
71# files containing the lists of code points that map to each such regular
72# expression property value, one file per list
73#
74# There is also a single form shortcut that Perl adds for many of the commonly
75# used properties. This happens for all binary properties, plus script,
76# general_category, and block properties.
77#
78# Thus the outputs of this program are files. There are map files, mostly in
79# the 'To' directory; and there are list files for use in regular expression
80# matching, all in subdirectories of the 'lib' directory, with each
81# subdirectory being named for the property that the lists in it are for.
82# Bookkeeping, test, and documentation files are also generated.
83
84my $matches_directory = 'lib'; # Where match (\p{}) files go.
85my $map_directory = 'To'; # Where map files go.
86
87# DATA STRUCTURES
88#
89# The major data structures of this program are Property, of course, but also
90# Table. There are two kinds of tables, very similar to each other.
91# "Match_Table" is the data structure giving the list of code points that have
92# a particular property value, mentioned above. There is also a "Map_Table"
93# data structure which gives the property's mapping from code point to value.
94# There are two structures because the match tables need to be combined in
95# various ways, such as constructing unions, intersections, complements, etc.,
96# and the map ones don't. And there would be problems, perhaps subtle, if
97# a map table were inadvertently operated on in some of those ways.
98# The use of separate classes with operations defined on one but not the other
99# prevents accidentally confusing the two.
100#
101# At the heart of each table's data structure is a "Range_List", which is just
102# an ordered list of "Ranges", plus ancillary information, and methods to
103# operate on them. A Range is a compact way to store property information.
104# Each range has a starting code point, an ending code point, and a value that
105# is meant to apply to all the code points between the two end points,
106# inclusive. For a map table, this value is the property value for those
107# code points. Two such ranges could be written like this:
108# 0x41 .. 0x5A, 'Upper',
109# 0x61 .. 0x7A, 'Lower'
110#
111# Each range also has a type used as a convenience to classify the values.
112# Most ranges in this program will be Type 0, or normal, but there are some
113# ranges that have a non-zero type. These are used only in map tables, and
114# are for mappings that don't fit into the normal scheme of things. Mappings
115# that require a hash entry to communicate with utf8.c are one example;
116# another example is mappings for charnames.pm to use which indicate a name
232ed87f 117# that is algorithmically determinable from its code point (and the reverse).
99870f4d
KW
118# These are used to significantly compact these tables, instead of listing
119# each one of the tens of thousands individually.
120#
121# In a match table, the value of a range is irrelevant (and hence the type as
122# well, which will always be 0), and arbitrarily set to the null string.
123# Using the example above, there would be two match tables for those two
124# entries, one named Upper would contain the 0x41..0x5A range, and the other
125# named Lower would contain 0x61..0x7A.
126#
127# Actually, there are two types of range lists, "Range_Map" is the one
128# associated with map tables, and "Range_List" with match tables.
232ed87f
KW
129# Again, this is so that methods can be defined on one and not the others so
130# as to prevent operating on them in incorrect ways.
99870f4d
KW
131#
132# Eventually, most tables are written out to files to be read by utf8_heavy.pl
133# in the perl core. All tables could in theory be written, but some are
134# suppressed because there is no current practical use for them. It is easy
135# to change which get written by changing various lists that are near the top
136# of the actual code in this file. The table data structures contain enough
137# ancillary information to allow them to be treated as separate entities for
138# writing, such as the path to each one's file. There is a heading in each
139# map table that gives the format of its entries, and what the map is for all
140# the code points missing from it. (This allows tables to be more compact.)
678f13d5 141#
99870f4d
KW
142# The Property data structure contains one or more tables. All properties
143# contain a map table (except the $perl property which is a
144# pseudo-property containing only match tables), and any properties that
145# are usable in regular expression matches also contain various matching
146# tables, one for each value the property can have. A binary property can
147# have two values, True and False (or Y and N, which are preferred by Unicode
148# terminology). Thus each of these properties will have a map table that
149# takes every code point and maps it to Y or N (but having ranges cuts the
150# number of entries in that table way down), and two match tables, one
151# which has a list of all the code points that map to Y, and one for all the
232ed87f 152# code points that map to N. (For each binary property, a third table is also
99870f4d 153# generated for the pseudo Perl property. It contains the identical code
232ed87f
KW
154# points as the Y table, but can be written in regular expressions, not in the
155# compound form, but in a "single" form like \p{IsUppercase}.) Many
156# properties are binary, but some properties have several possible values,
157# some have many, and properties like Name have a different value for every
158# named code point. Those will not, unless the controlling lists are changed,
159# have their match tables written out. But all the ones which can be used in
160# regular expression \p{} and \P{} constructs will. Prior to 5.14, generally
161# a property would have either its map table or its match tables written but
162# not both. Again, what gets written is controlled by lists which can easily
163# be changed. Starting in 5.14, advantage was taken of this, and all the map
164# tables needed to reconstruct the Unicode db are now written out, while
165# suppressing the Unicode .txt files that contain the data. Our tables are
166# much more compact than the .txt files, so a significant space savings was
167# achieved. Also, tables are not written out that are trivially derivable
168# from tables that do get written. So, there typically is no file containing
169# the code points not matched by a binary property (the table for \P{} versus
170# lowercase \p{}), since you just need to invert the True table to get the
171# False table.
172
173# Properties have a 'Type', like 'binary', or 'string', or 'enum' depending on
174# how many match tables there are and the content of the maps. This 'Type' is
c12f2655
KW
175# different than a range 'Type', so don't get confused by the two concepts
176# having the same name.
678f13d5 177#
99870f4d
KW
178# For information about the Unicode properties, see Unicode's UAX44 document:
179
180my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
181
182# As stated earlier, this program will work on any release of Unicode so far.
183# Most obvious problems in earlier data have NOT been corrected except when
be864b6c 184# necessary to make Perl or this program work reasonably, and to keep out
232ed87f
KW
185# potential security issues. For example, no folding information was given in
186# early releases, so this program substitutes lower case instead, just so that
187# a regular expression with the /i option will do something that actually
188# gives the right results in many cases. There are also a couple other
189# corrections for version 1.1.5, commented at the point they are made. As an
190# example of corrections that weren't made (but could be) is this statement
191# from DerivedAge.txt: "The supplementary private use code points and the
192# non-character code points were assigned in version 2.0, but not specifically
193# listed in the UCD until versions 3.0 and 3.1 respectively." (To be precise
194# it was 3.0.1 not 3.0.0) More information on Unicode version glitches is
195# further down in these introductory comments.
99870f4d 196#
232ed87f
KW
197# This program works on all non-provisional properties as of the current
198# Unicode release, though the files for some are suppressed for various
199# reasons. You can change which are output by changing lists in this program.
678f13d5 200#
dc85bd38 201# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
202# loose matchings rules (from Unicode TR18):
203#
204# The recommended names for UCD properties and property values are in
205# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
206# [PropValue]. There are both abbreviated names and longer, more
207# descriptive names. It is strongly recommended that both names be
208# recognized, and that loose matching of property names be used,
209# whereby the case distinctions, whitespace, hyphens, and underbar
210# are ignored.
232ed87f 211#
99870f4d
KW
212# The program still allows Fuzzy to override its determination of if loose
213# matching should be used, but it isn't currently used, as it is no longer
214# needed; the calculations it makes are good enough.
678f13d5 215#
99870f4d
KW
216# SUMMARY OF HOW IT WORKS:
217#
218# Process arguments
219#
220# A list is constructed containing each input file that is to be processed
221#
222# Each file on the list is processed in a loop, using the associated handler
223# code for each:
224# The PropertyAliases.txt and PropValueAliases.txt files are processed
225# first. These files name the properties and property values.
226# Objects are created of all the property and property value names
227# that the rest of the input should expect, including all synonyms.
228# The other input files give mappings from properties to property
229# values. That is, they list code points and say what the mapping
230# is under the given property. Some files give the mappings for
231# just one property; and some for many. This program goes through
232ed87f
KW
232# each file and populates the properties and their map tables from
233# them. Some properties are listed in more than one file, and
234# Unicode has set up a precedence as to which has priority if there
235# is a conflict. Thus the order of processing matters, and this
236# program handles the conflict possibility by processing the
237# overriding input files last, so that if necessary they replace
238# earlier values.
99870f4d
KW
239# After this is all done, the program creates the property mappings not
240# furnished by Unicode, but derivable from what it does give.
241# The tables of code points that match each property value in each
242# property that is accessible by regular expressions are created.
243# The Perl-defined properties are created and populated. Many of these
244# require data determined from the earlier steps
245# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 246# and Unicode are reconciled and warned about.
99870f4d
KW
247# All the properties are written to files
248# Any other files are written, and final warnings issued.
678f13d5 249#
99870f4d
KW
250# For clarity, a number of operators have been overloaded to work on tables:
251# ~ means invert (take all characters not in the set). The more
252# conventional '!' is not used because of the possibility of confusing
253# it with the actual boolean operation.
254# + means union
255# - means subtraction
256# & means intersection
257# The precedence of these is the order listed. Parentheses should be
258# copiously used. These are not a general scheme. The operations aren't
259# defined for a number of things, deliberately, to avoid getting into trouble.
260# Operations are done on references and affect the underlying structures, so
261# that the copy constructors for them have been overloaded to not return a new
262# clone, but the input object itself.
678f13d5 263#
99870f4d
KW
264# The bool operator is deliberately not overloaded to avoid confusion with
265# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
266#
267# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
268#
269# This program needs to be able to run under miniperl. Therefore, it uses a
270# minimum of other modules, and hence implements some things itself that could
271# be gotten from CPAN
272#
273# This program uses inputs published by the Unicode Consortium. These can
274# change incompatibly between releases without the Perl maintainers realizing
275# it. Therefore this program is now designed to try to flag these. It looks
276# at the directories where the inputs are, and flags any unrecognized files.
277# It keeps track of all the properties in the files it handles, and flags any
278# that it doesn't know how to handle. It also flags any input lines that
279# don't match the expected syntax, among other checks.
280#
281# It is also designed so if a new input file matches one of the known
282# templates, one hopefully just needs to add it to a list to have it
283# processed.
284#
285# As mentioned earlier, some properties are given in more than one file. In
286# particular, the files in the extracted directory are supposedly just
287# reformattings of the others. But they contain information not easily
9be22ca5
KW
288# derivable from the other files, including results for Unihan (which isn't
289# usually available to this program) and for unassigned code points. They
678f13d5
KW
290# also have historically had errors or been incomplete. In an attempt to
291# create the best possible data, this program thus processes them first to
292# glean information missing from the other files; then processes those other
293# files to override any errors in the extracted ones. Much of the design was
294# driven by this need to store things and then possibly override them.
295#
296# It tries to keep fatal errors to a minimum, to generate something usable for
297# testing purposes. It always looks for files that could be inputs, and will
298# warn about any that it doesn't know how to handle (the -q option suppresses
299# the warning).
99870f4d 300#
678f13d5
KW
301# Why is there more than one type of range?
302# This simplified things. There are some very specialized code points that
303# have to be handled specially for output, such as Hangul syllable names.
304# By creating a range type (done late in the development process), it
305# allowed this to be stored with the range, and overridden by other input.
306# Originally these were stored in another data structure, and it became a
307# mess trying to decide if a second file that was for the same property was
308# overriding the earlier one or not.
309#
310# Why are there two kinds of tables, match and map?
311# (And there is a base class shared by the two as well.) As stated above,
312# they actually are for different things. Development proceeded much more
313# smoothly when I (khw) realized the distinction. Map tables are used to
314# give the property value for every code point (actually every code point
315# that doesn't map to a default value). Match tables are used for regular
316# expression matches, and are essentially the inverse mapping. Separating
317# the two allows more specialized methods, and error checks so that one
318# can't just take the intersection of two map tables, for example, as that
319# is nonsensical.
99870f4d 320#
232ed87f
KW
321# What about 'fate' and 'status'. The concept of a table's fate was created
322# late when it became clear that something more was needed. The difference
323# between this and 'status' is unclean, and could be improved if someone
324# wanted to spend the effort.
325#
23e33b60
KW
326# DEBUGGING
327#
678f13d5
KW
328# This program is written so it will run under miniperl. Occasionally changes
329# will cause an error where the backtrace doesn't work well under miniperl.
330# To diagnose the problem, you can instead run it under regular perl, if you
331# have one compiled.
332#
333# There is a good trace facility. To enable it, first sub DEBUG must be set
334# to return true. Then a line like
335#
336# local $to_trace = 1 if main::DEBUG;
337#
232ed87f
KW
338# can be added to enable tracing in its lexical scope (plus dynamic) or until
339# you insert another line:
678f13d5
KW
340#
341# local $to_trace = 0 if main::DEBUG;
342#
232ed87f 343# To actually trace, use a line like "trace $a, @b, %c, ...;
678f13d5
KW
344#
345# Some of the more complex subroutines already have trace statements in them.
346# Permanent trace statements should be like:
347#
348# trace ... if main::DEBUG && $to_trace;
349#
350# If there is just one or a few files that you're debugging, you can easily
351# cause most everything else to be skipped. Change the line
352#
353# my $debug_skip = 0;
354#
355# to 1, and every file whose object is in @input_file_objects and doesn't have
232ed87f
KW
356# a, 'non_skip => 1,' in its constructor will be skipped. However, skipping
357# Jamo.txt or UnicodeData.txt will likely cause fatal errors.
678f13d5 358#
b4a0206c 359# To compare the output tables, it may be useful to specify the -annotate
97a8a595
KW
360# flag. (As of this writing, this can't be done on a clean workspace, due to
361# requirements in Text::Tabs used in this option; so first run mktables
362# without this option.) This option adds comment lines to each table, one for
363# each non-algorithmically named character giving, currently its code point,
364# name, and graphic representation if printable (and you have a font that
365# knows about it). This makes it easier to see what the particular code
366# points are in each output table. Non-named code points are annotated with a
367# description of their status, and contiguous ones with the same description
368# will be output as a range rather than individually. Algorithmically named
369# characters are also output as ranges, except when there are just a few
370# contiguous ones.
c4019d52 371#
99870f4d
KW
372# FUTURE ISSUES
373#
374# The program would break if Unicode were to change its names so that
375# interior white space, underscores, or dashes differences were significant
376# within property and property value names.
377#
378# It might be easier to use the xml versions of the UCD if this program ever
379# would need heavy revision, and the ability to handle old versions was not
380# required.
381#
382# There is the potential for name collisions, in that Perl has chosen names
383# that Unicode could decide it also likes. There have been such collisions in
384# the past, with mostly Perl deciding to adopt the Unicode definition of the
385# name. However in the 5.2 Unicode beta testing, there were a number of such
386# collisions, which were withdrawn before the final release, because of Perl's
387# and other's protests. These all involved new properties which began with
388# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
389# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
390# Unicode document, so they are unlikely to be used by Unicode for another
391# purpose. However, they might try something beginning with 'In', or use any
392# of the other Perl-defined properties. This program will warn you of name
393# collisions, and refuse to generate tables with them, but manual intervention
394# will be required in this event. One scheme that could be implemented, if
395# necessary, would be to have this program generate another file, or add a
396# field to mktables.lst that gives the date of first definition of a property.
397# Each new release of Unicode would use that file as a basis for the next
398# iteration. And the Perl synonym addition code could sort based on the age
399# of the property, so older properties get priority, and newer ones that clash
400# would be refused; hence existing code would not be impacted, and some other
401# synonym would have to be used for the new property. This is ugly, and
402# manual intervention would certainly be easier to do in the short run; lets
403# hope it never comes to this.
678f13d5 404#
99870f4d
KW
405# A NOTE ON UNIHAN
406#
b9e57660
KW
407# This program can generate tables from the Unihan database. But that db
408# isn't normally available, so it is marked as optional. Prior to version
409# 5.2, this database was in a single file, Unihan.txt. In 5.2 the database
410# was split into 8 different files, all beginning with the letters 'Unihan'.
411# If you plunk those files down into the directory mktables ($0) is in, this
412# program will read them and automatically create tables for the properties
413# from it that are listed in PropertyAliases.txt and PropValueAliases.txt,
414# plus any you add to the @cjk_properties array and the @cjk_property_values
415# array, being sure to add necessary '# @missings' lines to the latter. For
416# Unicode versions earlier than 5.2, most of the Unihan properties are not
417# listed at all in PropertyAliases nor PropValueAliases. This program assumes
418# for these early releases that you want the properties that are specified in
419# the 5.2 release.
99870f4d
KW
420#
421# You may need to adjust the entries to suit your purposes. setup_unihan(),
422# and filter_unihan_line() are the functions where this is done. This program
423# already does some adjusting to make the lines look more like the rest of the
424# Unicode DB; You can see what that is in filter_unihan_line()
425#
426# There is a bug in the 3.2 data file in which some values for the
427# kPrimaryNumeric property have commas and an unexpected comment. A filter
9be22ca5
KW
428# could be added to correct these; or for a particular installation, the
429# Unihan.txt file could be edited to fix them.
99870f4d 430#
678f13d5
KW
431# HOW TO ADD A FILE TO BE PROCESSED
432#
433# A new file from Unicode needs to have an object constructed for it in
434# @input_file_objects, probably at the end or at the end of the extracted
435# ones. The program should warn you if its name will clash with others on
436# restrictive file systems, like DOS. If so, figure out a better name, and
437# add lines to the README.perl file giving that. If the file is a character
232ed87f 438# property, it should be in the format that Unicode has implicitly
678f13d5
KW
439# standardized for such files for the more recently introduced ones.
440# If so, the Input_file constructor for @input_file_objects can just be the
441# file name and release it first appeared in. If not, then it should be
442# possible to construct an each_line_handler() to massage the line into the
443# standardized form.
444#
445# For non-character properties, more code will be needed. You can look at
446# the existing entries for clues.
447#
448# UNICODE VERSIONS NOTES
449#
450# The Unicode UCD has had a number of errors in it over the versions. And
451# these remain, by policy, in the standard for that version. Therefore it is
452# risky to correct them, because code may be expecting the error. So this
453# program doesn't generally make changes, unless the error breaks the Perl
454# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
455# for U+1105, which causes real problems for the algorithms for Jamo
456# calculations, so it is changed here.
457#
458# But it isn't so clear cut as to what to do about concepts that are
459# introduced in a later release; should they extend back to earlier releases
460# where the concept just didn't exist? It was easier to do this than to not,
461# so that's what was done. For example, the default value for code points not
462# in the files for various properties was probably undefined until changed by
463# some version. No_Block for blocks is such an example. This program will
464# assign No_Block even in Unicode versions that didn't have it. This has the
465# benefit that code being written doesn't have to special case earlier
466# versions; and the detriment that it doesn't match the Standard precisely for
467# the affected versions.
468#
469# Here are some observations about some of the issues in early versions:
470#
232ed87f
KW
471# Prior to version 3.0, there were 3 character decompositions. These are not
472# handled by Unicode::Normalize, nor will it compile when presented a version
473# that has them. However, you can trivially get it to compile by simply
474# ignoring those decompositions, by changing the croak to a carp. At the time
9be22ca5
KW
475# of this writing, the line (in cpan/Unicode-Normalize/Normalize.pm or
476# cpan/Unicode-Normalize/mkheader) reads
232ed87f
KW
477#
478# croak("Weird Canonical Decomposition of U+$h");
479#
28807e1d 480# Simply comment it out. It will compile, but will not know about any three
9be22ca5 481# character decompositions.
232ed87f
KW
482
483# The number of code points in \p{alpha=True} halved in 2.1.9. It turns out
484# that the reason is that the CJK block starting at 4E00 was removed from
485# PropList, and was not put back in until 3.1.0. The Perl extension (the
486# single property name \p{alpha}) has the correct values. But the compound
487# form is simply not generated until 3.1, as it can be argued that prior to
488# this release, this was not an official property. The comments for
489# filter_old_style_proplist() give more details.
678f13d5
KW
490#
491# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
492# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
493# reason is that 3.2 introduced U+205F=medium math space, which was not
494# classed as white space, but Perl figured out that it should have been. 4.0
495# reclassified it correctly.
496#
497# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
232ed87f
KW
498# this was erroneously a synonym for 202 (it should be 200). In 4.0, ATB
499# became 202, and ATBL was left with no code points, as all the ones that
500# mapped to 202 stayed mapped to 202. Thus if your program used the numeric
501# name for the class, it would not have been affected, but if it used the
502# mnemonic, it would have been.
678f13d5 503#
9be22ca5 504# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that, code
678f13d5
KW
505# points which eventually came to have this script property value, instead
506# mapped to "Unknown". But in the next release all these code points were
507# moved to \p{sc=common} instead.
9be22ca5
KW
508
509# The tests furnished by Unicode for testing WordBreak and SentenceBreak
510# generate errors in 5.0 and earlier.
99870f4d
KW
511#
512# The default for missing code points for BidiClass is complicated. Starting
513# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
514# tries to do the best it can for earlier releases. It is done in
515# process_PropertyAliases()
516#
232ed87f
KW
517# In version 2.1.2, the entry in UnicodeData.txt:
518# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;;019F;
519# should instead be
520# 0275;LATIN SMALL LETTER BARRED O;Ll;0;L;;;;;N;;;019F;;019F
521# Without this change, there are casing problems for this character.
522#
7803ad2d
KW
523# Search for $string_compare_versions to see how to compare changes to
524# properties between Unicode versions
525#
99870f4d
KW
526##############################################################################
527
528my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
529 # and errors
530my $MAX_LINE_WIDTH = 78;
531
532# Debugging aid to skip most files so as to not be distracted by them when
533# concentrating on the ones being debugged. Add
534# non_skip => 1,
535# to the constructor for those files you want processed when you set this.
536# Files with a first version number of 0 are special: they are always
c12f2655
KW
537# processed regardless of the state of this flag. Generally, Jamo.txt and
538# UnicodeData.txt must not be skipped if you want this program to not die
539# before normal completion.
99870f4d
KW
540my $debug_skip = 0;
541
e9c4b4f8
KW
542
543# Normally these are suppressed.
544my $write_Unicode_deprecated_tables = 0;
545
99870f4d
KW
546# Set to 1 to enable tracing.
547our $to_trace = 0;
548
549{ # Closure for trace: debugging aid
550 my $print_caller = 1; # ? Include calling subroutine name
551 my $main_with_colon = 'main::';
552 my $main_colon_length = length($main_with_colon);
553
554 sub trace {
555 return unless $to_trace; # Do nothing if global flag not set
556
557 my @input = @_;
558
559 local $DB::trace = 0;
560 $DB::trace = 0; # Quiet 'used only once' message
561
562 my $line_number;
563
564 # Loop looking up the stack to get the first non-trace caller
565 my $caller_line;
566 my $caller_name;
567 my $i = 0;
568 do {
569 $line_number = $caller_line;
570 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
571 $caller = $main_with_colon unless defined $caller;
572
573 $caller_name = $caller;
574
575 # get rid of pkg
576 $caller_name =~ s/.*:://;
577 if (substr($caller_name, 0, $main_colon_length)
578 eq $main_with_colon)
579 {
580 $caller_name = substr($caller_name, $main_colon_length);
581 }
582
583 } until ($caller_name ne 'trace');
584
585 # If the stack was empty, we were called from the top level
586 $caller_name = 'main' if ($caller_name eq ""
587 || $caller_name eq 'trace');
588
589 my $output = "";
9346f59a 590 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
99870f4d 591 foreach my $string (@input) {
99870f4d
KW
592 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
593 $output .= simple_dumper($string);
594 }
595 else {
596 $string = "$string" if ref $string;
597 $string = $UNDEF unless defined $string;
598 chomp $string;
599 $string = '""' if $string eq "";
600 $output .= " " if $output ne ""
601 && $string ne ""
602 && substr($output, -1, 1) ne " "
603 && substr($string, 0, 1) ne " ";
604 $output .= $string;
605 }
606 }
607
99f78760
KW
608 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
609 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
610 print STDERR $output, "\n";
611 return;
612 }
613}
614
615# This is for a rarely used development feature that allows you to compare two
616# versions of the Unicode standard without having to deal with changes caused
9be22ca5
KW
617# by the code points introduced in the later version. You probably also want
618# to use the -annotate option when using this. Change the 0 to a string
619# containing a SINGLE dotted Unicode release number (e.g. "2.1"). Only code
620# points introduced in that release and earlier will be used; later ones are
621# thrown away. You use the version number of the earliest one you want to
c12f2655
KW
622# compare; then run this program on directory structures containing each
623# release, and compare the outputs. These outputs will therefore include only
624# the code points common to both releases, and you can see the changes caused
625# just by the underlying release semantic changes. For versions earlier than
626# 3.2, you must copy a version of DAge.txt into the directory.
627my $string_compare_versions = DEBUG && 0; # e.g., "2.1";
99870f4d
KW
628my $compare_versions = DEBUG
629 && $string_compare_versions
630 && pack "C*", split /\./, $string_compare_versions;
631
632sub uniques {
633 # Returns non-duplicated input values. From "Perl Best Practices:
634 # Encapsulated Cleverness". p. 455 in first edition.
635
636 my %seen;
0e407844
NC
637 # Arguably this breaks encapsulation, if the goal is to permit multiple
638 # distinct objects to stringify to the same value, and be interchangeable.
639 # However, for this program, no two objects stringify identically, and all
640 # lists passed to this function are either objects or strings. So this
641 # doesn't affect correctness, but it does give a couple of percent speedup.
642 no overloading;
99870f4d
KW
643 return grep { ! $seen{$_}++ } @_;
644}
645
646$0 = File::Spec->canonpath($0);
647
648my $make_test_script = 0; # ? Should we output a test script
6b5ab373 649my $make_norm_test_script = 0; # ? Should we output a normalization test script
99870f4d
KW
650my $write_unchanged_files = 0; # ? Should we update the output files even if
651 # we don't think they have changed
652my $use_directory = ""; # ? Should we chdir somewhere.
653my $pod_directory; # input directory to store the pod file.
654my $pod_file = 'perluniprops';
655my $t_path; # Path to the .t test file
656my $file_list = 'mktables.lst'; # File to store input and output file names.
657 # This is used to speed up the build, by not
658 # executing the main body of the program if
659 # nothing on the list has changed since the
660 # previous build
661my $make_list = 1; # ? Should we write $file_list. Set to always
662 # make a list so that when the pumpking is
663 # preparing a release, s/he won't have to do
664 # special things
665my $glob_list = 0; # ? Should we try to include unknown .txt files
666 # in the input.
bd9ebcfd
KW
667my $output_range_counts = $debugging_build; # ? Should we include the number
668 # of code points in ranges in
669 # the output
558712cf 670my $annotate = 0; # ? Should character names be in the output
9ef2b94f 671
99870f4d
KW
672# Verbosity levels; 0 is quiet
673my $NORMAL_VERBOSITY = 1;
674my $PROGRESS = 2;
675my $VERBOSE = 3;
676
677my $verbosity = $NORMAL_VERBOSITY;
678
0458fbc1
KW
679# Stored in mktables.lst so that if this program is called with different
680# options, will regenerate even if the files otherwise look like they're
681# up-to-date.
682my $command_line_arguments = join " ", @ARGV;
683
99870f4d
KW
684# Process arguments
685while (@ARGV) {
cf25bb62
JH
686 my $arg = shift @ARGV;
687 if ($arg eq '-v') {
99870f4d
KW
688 $verbosity = $VERBOSE;
689 }
690 elsif ($arg eq '-p') {
691 $verbosity = $PROGRESS;
692 $| = 1; # Flush buffers as we go.
693 }
694 elsif ($arg eq '-q') {
695 $verbosity = 0;
696 }
697 elsif ($arg eq '-w') {
698 $write_unchanged_files = 1; # update the files even if havent changed
699 }
700 elsif ($arg eq '-check') {
6ae7e459
YO
701 my $this = shift @ARGV;
702 my $ok = shift @ARGV;
703 if ($this ne $ok) {
704 print "Skipping as check params are not the same.\n";
705 exit(0);
706 }
00a8df5c 707 }
99870f4d
KW
708 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
709 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
710 }
3df51b85
KW
711 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
712 {
99870f4d 713 $make_test_script = 1;
99870f4d 714 }
6b5ab373
KW
715 elsif ($arg eq '-makenormtest')
716 {
717 $make_norm_test_script = 1;
718 }
99870f4d
KW
719 elsif ($arg eq '-makelist') {
720 $make_list = 1;
721 }
722 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
723 -d $use_directory or croak "Unknown directory '$use_directory'";
724 }
725 elsif ($arg eq '-L') {
726
727 # Existence not tested until have chdir'd
728 $file_list = shift;
729 }
730 elsif ($arg eq '-globlist') {
731 $glob_list = 1;
732 }
733 elsif ($arg eq '-c') {
734 $output_range_counts = ! $output_range_counts
735 }
b4a0206c 736 elsif ($arg eq '-annotate') {
558712cf 737 $annotate = 1;
bd9ebcfd
KW
738 $debugging_build = 1;
739 $output_range_counts = 1;
9ef2b94f 740 }
99870f4d
KW
741 else {
742 my $with_c = 'with';
743 $with_c .= 'out' if $output_range_counts; # Complements the state
744 croak <<END;
745usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
746 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
747 [-check A B ]
748 -c : Output comments $with_c number of code points in ranges
749 -q : Quiet Mode: Only output serious warnings.
750 -p : Set verbosity level to normal plus show progress.
751 -v : Set Verbosity level high: Show progress and non-serious
752 warnings
753 -w : Write files regardless
754 -C dir : Change to this directory before proceeding. All relative paths
755 except those specified by the -P and -T options will be done
756 with respect to this directory.
757 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 758 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
759 -L filelist : Use alternate 'filelist' instead of standard one
760 -globlist : Take as input all non-Test *.txt files in current and sub
761 directories
3df51b85
KW
762 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
763 overrides -T
99870f4d 764 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 765 -annotate : Output an annotation for each character in the table files;
97a8a595
KW
766 useful for debugging mktables, looking at diffs; but is slow
767 and memory intensive
99870f4d
KW
768 -check A B : Executes $0 only if A and B are the same
769END
770 }
771}
772
773# Stores the most-recently changed file. If none have changed, can skip the
774# build
aeab6150 775my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
776
777# Change directories now, because need to read 'version' early.
778if ($use_directory) {
3df51b85 779 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
780 $pod_directory = File::Spec->rel2abs($pod_directory);
781 }
3df51b85 782 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 783 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 784 }
99870f4d 785 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 786 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 787 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 788 }
3df51b85 789 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 790 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 791 }
00a8df5c
YO
792}
793
99870f4d
KW
794# Get Unicode version into regular and v-string. This is done now because
795# various tables below get populated based on it. These tables are populated
796# here to be near the top of the file, and so easily seeable by those needing
797# to modify things.
798open my $VERSION, "<", "version"
799 or croak "$0: can't open required file 'version': $!\n";
800my $string_version = <$VERSION>;
801close $VERSION;
802chomp $string_version;
803my $v_version = pack "C*", split /\./, $string_version; # v string
804
805# The following are the complete names of properties with property values that
806# are known to not match any code points in some versions of Unicode, but that
807# may change in the future so they should be matchable, hence an empty file is
808# generated for them.
a9c9e371
KW
809my @tables_that_may_be_empty;
810push @tables_that_may_be_empty, 'Joining_Type=Left_Joining'
811 if $v_version lt v6.3.0;
99870f4d
KW
812push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
813push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
814push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
815 if $v_version ge v4.1.0;
82aed44a
KW
816push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
817 if $v_version ge v6.0.0;
f583b44c
KW
818push @tables_that_may_be_empty, 'Grapheme_Cluster_Break=Prepend'
819 if $v_version ge v6.1.0;
1e958ea9
KW
820push @tables_that_may_be_empty, 'Canonical_Combining_Class=CCC133'
821 if $v_version ge v6.2.0;
99870f4d
KW
822
823# The lists below are hashes, so the key is the item in the list, and the
824# value is the reason why it is in the list. This makes generation of
825# documentation easier.
826
827my %why_suppressed; # No file generated for these.
828
829# Files aren't generated for empty extraneous properties. This is arguable.
830# Extraneous properties generally come about because a property is no longer
831# used in a newer version of Unicode. If we generated a file without code
832# points, programs that used to work on that property will still execute
833# without errors. It just won't ever match (or will always match, with \P{}).
834# This means that the logic is now likely wrong. I (khw) think its better to
835# find this out by getting an error message. Just move them to the table
836# above to change this behavior
837my %why_suppress_if_empty_warn_if_not = (
838
839 # It is the only property that has ever officially been removed from the
840 # Standard. The database never contained any code points for it.
841 'Special_Case_Condition' => 'Obsolete',
842
843 # Apparently never official, but there were code points in some versions of
844 # old-style PropList.txt
845 'Non_Break' => 'Obsolete',
846);
847
848# These would normally go in the warn table just above, but they were changed
849# a long time before this program was written, so warnings about them are
850# moot.
851if ($v_version gt v3.2.0) {
852 push @tables_that_may_be_empty,
853 'Canonical_Combining_Class=Attached_Below_Left'
854}
855
9be22ca5
KW
856# Enum values for to_output_map() method in the Map_Table package. (0 is don't
857# output)
272501f6
KW
858my $EXTERNAL_MAP = 1;
859my $INTERNAL_MAP = 2;
ce712c88 860my $OUTPUT_ADJUSTED = 3;
272501f6 861
fcf1973c
KW
862# To override computed values for writing the map tables for these properties.
863# The default for enum map tables is to write them out, so that the Unicode
864# .txt files can be removed, but all the data to compute any property value
865# for any code point is available in a more compact form.
866my %global_to_output_map = (
867 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
868 # get stuck supporting it if things change. Since it is a STRING
869 # property, it normally would be listed in the pod, but INTERNAL_MAP
870 # suppresses that.
fcf1973c
KW
871 Unicode_1_Name => $INTERNAL_MAP,
872
873 Present_In => 0, # Suppress, as easily computed from Age
74cd47d0
KW
874 Block => (NON_ASCII_PLATFORM) ? 1 : 0, # Suppress, as Blocks.txt is
875 # retained, but needed for
876 # non-ASCII
53d34b6c
KW
877
878 # Suppress, as mapping can be found instead from the
879 # Perl_Decomposition_Mapping file
880 Decomposition_Type => 0,
fcf1973c
KW
881);
882
99870f4d
KW
883# There are several types of obsolete properties defined by Unicode. These
884# must be hand-edited for every new Unicode release.
885my %why_deprecated; # Generates a deprecated warning message if used.
886my %why_stabilized; # Documentation only
887my %why_obsolete; # Documentation only
888
889{ # Closure
8364f9eb 890 my $simple = 'Perl uses the more complete version';
99870f4d
KW
891 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
892
893 my $other_properties = 'other properties';
894 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 895 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
896
897 %why_deprecated = (
5f7264c7 898 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
899 'Jamo_Short_Name' => $contributory,
900 'Line_Break=Surrogate' => 'Deprecated by Unicode because surrogates should never appear in well-formed text, and therefore shouldn\'t be the basis for line breaking',
901 'Other_Alphabetic' => $contributory,
902 'Other_Default_Ignorable_Code_Point' => $contributory,
903 'Other_Grapheme_Extend' => $contributory,
904 'Other_ID_Continue' => $contributory,
905 'Other_ID_Start' => $contributory,
906 'Other_Lowercase' => $contributory,
907 'Other_Math' => $contributory,
908 'Other_Uppercase' => $contributory,
e22aaf5c
KW
909 'Expands_On_NFC' => $why_no_expand,
910 'Expands_On_NFD' => $why_no_expand,
911 'Expands_On_NFKC' => $why_no_expand,
912 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
913 );
914
915 %why_suppressed = (
5f7264c7 916 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
917 # contains the same information, but without the algorithmically
918 # determinable Hangul syllables'. This file is not published, so it's
919 # existence is not noted in the comment.
12fee290 920 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize or prop_invmap() or charprop() in Unicode::UCD::',
99870f4d 921
5f8d1a89
KW
922 # Don't suppress ISO_Comment, as otherwise special handling is needed
923 # to differentiate between it and gc=c, which can be written as 'isc',
924 # which is the same characters as ISO_Comment's short name.
99870f4d 925
12fee290 926 'Name' => "Accessible via \\N{...} or 'use charnames;' or charprop() or prop_invmap() in Unicode::UCD::",
e0b29447 927
12fee290
KW
928 'Simple_Case_Folding' => "$simple. Can access this through casefold(), charprop(), or prop_invmap() in Unicode::UCD",
929 'Simple_Lowercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
930 'Simple_Titlecase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
931 'Simple_Uppercase_Mapping' => "$simple. Can access this through charinfo(), charprop(), or prop_invmap() in Unicode::UCD",
99870f4d 932
dac6f618 933 FC_NFKC_Closure => 'Deprecated by Unicode, and supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
934 );
935
1704a0ea
KW
936 foreach my $property (
937
938 # The following are suppressed because they were made contributory
939 # or deprecated by Unicode before Perl ever thought about
940 # supporting them.
941 'Jamo_Short_Name',
942 'Grapheme_Link',
943 'Expands_On_NFC',
944 'Expands_On_NFD',
945 'Expands_On_NFKC',
946 'Expands_On_NFKD',
947
948 # The following are suppressed because they have been marked
949 # as deprecated for a sufficient amount of time
950 'Other_Alphabetic',
951 'Other_Default_Ignorable_Code_Point',
952 'Other_Grapheme_Extend',
953 'Other_ID_Continue',
954 'Other_ID_Start',
955 'Other_Lowercase',
956 'Other_Math',
957 'Other_Uppercase',
e22aaf5c 958 ) {
99870f4d
KW
959 $why_suppressed{$property} = $why_deprecated{$property};
960 }
cf25bb62 961
99870f4d
KW
962 # Customize the message for all the 'Other_' properties
963 foreach my $property (keys %why_deprecated) {
964 next if (my $main_property = $property) !~ s/^Other_//;
965 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
966 }
967}
968
e9c4b4f8
KW
969if ($write_Unicode_deprecated_tables) {
970 foreach my $property (keys %why_suppressed) {
971 delete $why_suppressed{$property} if $property =~
972 / ^ Other | Grapheme /x;
973 }
974}
975
99870f4d
KW
976if ($v_version ge 4.0.0) {
977 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
978 if ($v_version ge 6.0.0) {
979 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
980 }
99870f4d 981}
5f7264c7 982if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 983 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 984 if ($v_version ge 6.0.0) {
63f74647 985 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 986 }
99870f4d
KW
987}
988
989# Probably obsolete forever
990if ($v_version ge v4.1.0) {
82aed44a
KW
991 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
992}
993if ($v_version ge v6.0.0) {
caa75395 994 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana" (or both)';
2b352efd 995 $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
996}
997
998# This program can create files for enumerated-like properties, such as
999# 'Numeric_Type'. This file would be the same format as for a string
1000# property, with a mapping from code point to its value, so you could look up,
1001# for example, the script a code point is in. But no one so far wants this
1002# mapping, or they have found another way to get it since this is a new
1003# feature. So no file is generated except if it is in this list.
1004my @output_mapped_properties = split "\n", <<END;
1005END
1006
b9e57660
KW
1007# If you want more Unihan properties than the default, you need to add them to
1008# these arrays. Depending on the property type, @missing lines might have to
1009# be added to the second array. A sample entry would be (including the '#'):
1010# @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
99870f4d 1011my @cjk_properties = split "\n", <<'END';
99870f4d 1012END
99870f4d 1013my @cjk_property_values = split "\n", <<'END';
99870f4d
KW
1014END
1015
1016# The input files don't list every code point. Those not listed are to be
1017# defaulted to some value. Below are hard-coded what those values are for
1018# non-binary properties as of 5.1. Starting in 5.0, there are
caa75395 1019# machine-parsable comment lines in the files that give the defaults; so this
99870f4d
KW
1020# list shouldn't have to be extended. The claim is that all missing entries
1021# for binary properties will default to 'N'. Unicode tried to change that in
1022# 5.2, but the beta period produced enough protest that they backed off.
1023#
1024# The defaults for the fields that appear in UnicodeData.txt in this hash must
1025# be in the form that it expects. The others may be synonyms.
1026my $CODE_POINT = '<code point>';
1027my %default_mapping = (
1028 Age => "Unassigned",
1029 # Bidi_Class => Complicated; set in code
1030 Bidi_Mirroring_Glyph => "",
1031 Block => 'No_Block',
1032 Canonical_Combining_Class => 0,
1033 Case_Folding => $CODE_POINT,
1034 Decomposition_Mapping => $CODE_POINT,
1035 Decomposition_Type => 'None',
1036 East_Asian_Width => "Neutral",
1037 FC_NFKC_Closure => $CODE_POINT,
ec2cd995 1038 General_Category => ($v_version le 6.3.0) ? 'Cn' : 'Unassigned',
99870f4d
KW
1039 Grapheme_Cluster_Break => 'Other',
1040 Hangul_Syllable_Type => 'NA',
1041 ISO_Comment => "",
1042 Jamo_Short_Name => "",
1043 Joining_Group => "No_Joining_Group",
1044 # Joining_Type => Complicated; set in code
1045 kIICore => 'N', # Is converted to binary
1046 #Line_Break => Complicated; set in code
1047 Lowercase_Mapping => $CODE_POINT,
1048 Name => "",
1049 Name_Alias => "",
1050 NFC_QC => 'Yes',
1051 NFD_QC => 'Yes',
1052 NFKC_QC => 'Yes',
1053 NFKD_QC => 'Yes',
1054 Numeric_Type => 'None',
1055 Numeric_Value => 'NaN',
1056 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1057 Sentence_Break => 'Other',
1058 Simple_Case_Folding => $CODE_POINT,
1059 Simple_Lowercase_Mapping => $CODE_POINT,
1060 Simple_Titlecase_Mapping => $CODE_POINT,
1061 Simple_Uppercase_Mapping => $CODE_POINT,
1062 Titlecase_Mapping => $CODE_POINT,
1063 Unicode_1_Name => "",
1064 Unicode_Radical_Stroke => "",
1065 Uppercase_Mapping => $CODE_POINT,
1066 Word_Break => 'Other',
1067);
1068
678f13d5 1069### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1070
1071my $HEADER=<<"EOF";
1072# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85
KW
1073# This file is machine-generated by $0 from the Unicode
1074# database, Version $string_version. Any changes made here will be lost!
cf25bb62
JH
1075EOF
1076
126c3d4e 1077my $INTERNAL_ONLY_HEADER = <<"EOF";
99870f4d
KW
1078
1079# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1080# This file is for internal use by core Perl only. The format and even the
1081# name or existence of this file are subject to change without notice. Don't
d9ae3878
KW
1082# use it directly. Use Unicode::UCD to access the Unicode character data
1083# base.
99870f4d
KW
1084EOF
1085
1086my $DEVELOPMENT_ONLY=<<"EOF";
1087# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1088# This file contains information artificially constrained to code points
1089# present in Unicode release $string_compare_versions.
1090# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1091# not be used for production.
b6922eda
KW
1092
1093EOF
1094
346a20cf
KW
1095my $MAX_UNICODE_CODEPOINT_STRING = ($v_version ge v2.0.0)
1096 ? "10FFFF"
1097 : "FFFF";
6189eadc
KW
1098my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1099my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d 1100
2d88a86a
KW
1101# We work with above-Unicode code points, up to UV_MAX. But when you get
1102# that high, above IV_MAX, some operations don't work, and you can easily get
1103# overflow. Therefore for internal use, we use a much smaller number,
1104# translating it to UV_MAX only for output. The exact number is immaterial
1105# (all Unicode code points are treated exactly the same), but the algorithm
1106# requires it to be at least 2 * $MAX_UNICODE_CODEPOINTS + 1;
1107my $MAX_WORKING_CODEPOINTS= $MAX_UNICODE_CODEPOINT * 8;
1108my $MAX_WORKING_CODEPOINT = $MAX_WORKING_CODEPOINTS - 1;
1109my $MAX_WORKING_CODEPOINT_STRING = sprintf("%X", $MAX_WORKING_CODEPOINT);
1110
1111my $MAX_PLATFORM_CODEPOINT = ~0;
1112
99870f4d
KW
1113# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1114# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1115# to decrease backtracking. The first regex allows the code point to be at
1116# the end of a word, but to work properly, the word shouldn't end with a valid
1117# hex character. The second one won't match a code point at the end of a
1118# word, and doesn't have the run-on issue
8c32d378
KW
1119my $run_on_code_point_re =
1120 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1121my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d
KW
1122
1123# This matches the beginning of the line in the Unicode db files that give the
1124# defaults for code points not listed (i.e., missing) in the file. The code
1125# depends on this ending with a semi-colon, so it can assume it is a valid
1126# field when the line is split() by semi-colons
346a20cf 1127my $missing_defaults_prefix = qr/^#\s+\@missing:\s+0000\.\.10FFFF\s*;/;
99870f4d
KW
1128
1129# Property types. Unicode has more types, but these are sufficient for our
1130# purposes.
1131my $UNKNOWN = -1; # initialized to illegal value
1132my $NON_STRING = 1; # Either binary or enum
1133my $BINARY = 2;
06f26c45
KW
1134my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1135 # tables, additional true and false tables are
1136 # generated so that false is anything matching the
1137 # default value, and true is everything else.
1138my $ENUM = 4; # Include catalog
1139my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1140
1141# Some input files have lines that give default values for code points not
1142# contained in the file. Sometimes these should be ignored.
1143my $NO_DEFAULTS = 0; # Must evaluate to false
1144my $NOT_IGNORED = 1;
1145my $IGNORED = 2;
1146
1147# Range types. Each range has a type. Most ranges are type 0, for normal,
1148# and will appear in the main body of the tables in the output files, but
1149# there are other types of ranges as well, listed below, that are specially
1150# handled. There are pseudo-types as well that will never be stored as a
1151# type, but will affect the calculation of the type.
1152
1153# 0 is for normal, non-specials
1154my $MULTI_CP = 1; # Sequence of more than code point
1155my $HANGUL_SYLLABLE = 2;
1156my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1157my $NULL = 4; # The map is to the null string; utf8.c can't
1158 # handle these, nor is there an accepted syntax
1159 # for them in \p{} constructs
f86864ac 1160my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1161 # otherwise be $MULTI_CP type are instead type 0
1162
1163# process_generic_property_file() can accept certain overrides in its input.
1164# Each of these must begin AND end with $CMD_DELIM.
1165my $CMD_DELIM = "\a";
1166my $REPLACE_CMD = 'replace'; # Override the Replace
1167my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1168
1169my $NO = 0;
1170my $YES = 1;
1171
1172# Values for the Replace argument to add_range.
1173# $NO # Don't replace; add only the code points not
1174 # already present.
1175my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1176 # the comments at the subroutine definition.
1177my $UNCONDITIONALLY = 2; # Replace without conditions.
9470941f 1178my $MULTIPLE_BEFORE = 4; # Don't replace, but add a duplicate record if
99870f4d 1179 # already there
7f4b1e25
KW
1180my $MULTIPLE_AFTER = 5; # Don't replace, but add a duplicate record if
1181 # already there
1182my $CROAK = 6; # Die with an error if is already there
99870f4d
KW
1183
1184# Flags to give property statuses. The phrases are to remind maintainers that
1185# if the flag is changed, the indefinite article referring to it in the
1186# documentation may need to be as well.
1187my $NORMAL = "";
99870f4d
KW
1188my $DEPRECATED = 'D';
1189my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1190my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1191my $DISCOURAGED = 'X';
1192my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1193my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1194my $STRICTER = 'T';
1195my $a_bold_stricter = "a 'B<$STRICTER>'";
1196my $A_bold_stricter = "A 'B<$STRICTER>'";
1197my $STABILIZED = 'S';
1198my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1199my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1200my $OBSOLETE = 'O';
1201my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1202my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1203
5c6d9947
KW
1204# Aliases can also have an extra status:
1205my $INTERNAL_ALIAS = 'P';
1206
99870f4d
KW
1207my %status_past_participles = (
1208 $DISCOURAGED => 'discouraged',
99870f4d
KW
1209 $STABILIZED => 'stabilized',
1210 $OBSOLETE => 'obsolete',
37e2e78e 1211 $DEPRECATED => 'deprecated',
5c6d9947 1212 $INTERNAL_ALIAS => 'reserved for Perl core internal use only',
99870f4d
KW
1213);
1214
395dfc19
KW
1215# Table fates. These are somewhat ordered, so that fates < $MAP_PROXIED should be
1216# externally documented.
301ba948 1217my $ORDINARY = 0; # The normal fate.
395dfc19
KW
1218my $MAP_PROXIED = 1; # The map table for the property isn't written out,
1219 # but there is a file written that can be used to
1220 # reconstruct this table
3cdaf629 1221my $INTERNAL_ONLY = 2; # The file for this table is written out, but it is
301ba948 1222 # for Perl's internal use only
277b7b16
KW
1223my $LEGACY_ONLY = 3; # Like $INTERNAL_ONLY, but not actually used by Perl.
1224 # Is for backwards compatibility for applications that
1225 # read the file directly, so it's format is
1226 # unchangeable.
1227my $SUPPRESSED = 4; # The file for this table is not written out, and as a
3cdaf629
KW
1228 # result, we don't bother to do many computations on
1229 # it.
277b7b16 1230my $PLACEHOLDER = 5; # Like $SUPPRESSED, but we go through all the
3cdaf629
KW
1231 # computations anyway, as the values are needed for
1232 # things to work. This happens when we have Perl
1233 # extensions that depend on Unicode tables that
1234 # wouldn't normally be in a given Unicode version.
301ba948 1235
f5817e0a
KW
1236# The format of the values of the tables:
1237my $EMPTY_FORMAT = "";
99870f4d
KW
1238my $BINARY_FORMAT = 'b';
1239my $DECIMAL_FORMAT = 'd';
1240my $FLOAT_FORMAT = 'f';
1241my $INTEGER_FORMAT = 'i';
1242my $HEX_FORMAT = 'x';
1243my $RATIONAL_FORMAT = 'r';
1244my $STRING_FORMAT = 's';
d11155ec 1245my $ADJUST_FORMAT = 'a';
24303724 1246my $HEX_ADJUST_FORMAT = 'ax';
a14f3cb1 1247my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1248my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1249
1250my %map_table_formats = (
1251 $BINARY_FORMAT => 'binary',
1252 $DECIMAL_FORMAT => 'single decimal digit',
1253 $FLOAT_FORMAT => 'floating point number',
1254 $INTEGER_FORMAT => 'integer',
add63c13 1255 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1256 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1257 $STRING_FORMAT => 'string',
d11155ec 1258 $ADJUST_FORMAT => 'some entries need adjustment',
24303724 1259 $HEX_ADJUST_FORMAT => 'mapped value in hex; some entries need adjustment',
92f9d56c 1260 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1261 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1262);
1263
1264# Unicode didn't put such derived files in a separate directory at first.
1265my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1266my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1267my $AUXILIARY = 'auxiliary';
1268
3854b4b8
KW
1269# Hashes and arrays that will eventually go into Heavy.pl for the use of
1270# utf8_heavy.pl and into UCD.pl for the use of UCD.pm
99870f4d
KW
1271my %loose_to_file_of; # loosely maps table names to their respective
1272 # files
1273my %stricter_to_file_of; # same; but for stricter mapping.
315bfd4e 1274my %loose_property_to_file_of; # Maps a loose property name to its map file
e1dc048d 1275my %strict_property_to_file_of; # Same, but strict
3854b4b8
KW
1276my @inline_definitions = "V0"; # Each element gives a definition of a unique
1277 # inversion list. When a definition is inlined,
1278 # its value in the hash it's in (one of the two
1279 # defined just above) will include an index into
1280 # this array. The 0th element is initialized to
9be22ca5 1281 # the definition for a zero length inversion list
89cf10cc
KW
1282my %file_to_swash_name; # Maps the file name to its corresponding key name
1283 # in the hash %utf8::SwashInfo
99870f4d
KW
1284my %nv_floating_to_rational; # maps numeric values floating point numbers to
1285 # their rational equivalent
c12f2655
KW
1286my %loose_property_name_of; # Loosely maps (non_string) property names to
1287 # standard form
e1dc048d
KW
1288my %strict_property_name_of; # Strictly maps (non_string) property names to
1289 # standard form
86a52d1e 1290my %string_property_loose_to_name; # Same, for string properties.
c15fda25
KW
1291my %loose_defaults; # keys are of form "prop=value", where 'prop' is
1292 # the property name in standard loose form, and
1293 # 'value' is the default value for that property,
1294 # also in standard loose form.
9e4a1e86
KW
1295my %loose_to_standard_value; # loosely maps table names to the canonical
1296 # alias for them
2df7880f
KW
1297my %ambiguous_names; # keys are alias names (in standard form) that
1298 # have more than one possible meaning.
d963b40d
KW
1299my %combination_property; # keys are alias names (in standard form) that
1300 # have both a map table, and a binary one that
1301 # yields true for all non-null maps.
5d1df013
KW
1302my %prop_aliases; # Keys are standard property name; values are each
1303 # one's aliases
1e863613
KW
1304my %prop_value_aliases; # Keys of top level are standard property name;
1305 # values are keys to another hash, Each one is
1306 # one of the property's values, in standard form.
1307 # The values are that prop-val's aliases.
b9e57660 1308my %skipped_files; # List of files that we skip
2df7880f 1309my %ucd_pod; # Holds entries that will go into the UCD section of the pod
99870f4d 1310
d867ccfb
KW
1311# Most properties are immune to caseless matching, otherwise you would get
1312# nonsensical results, as properties are a function of a code point, not
1313# everything that is caselessly equivalent to that code point. For example,
1314# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1315# be true because 's' and 'S' are equivalent caselessly. However,
1316# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1317# extend that concept to those very few properties that are like this. Each
1318# such property will match the full range caselessly. They are hard-coded in
1319# the program; it's not worth trying to make it general as it's extremely
1320# unlikely that they will ever change.
1321my %caseless_equivalent_to;
1322
53adf6a2
KW
1323# This is the range of characters that were in Release 1 of Unicode, and
1324# removed in Release 2 (replaced with the current Hangul syllables starting at
1325# U+AC00). The range was reused starting in Release 3 for other purposes.
1326my $FIRST_REMOVED_HANGUL_SYLLABLE = 0x3400;
1327my $FINAL_REMOVED_HANGUL_SYLLABLE = 0x4DFF;
1328
99870f4d
KW
1329# These constants names and values were taken from the Unicode standard,
1330# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1331# syllables. The '_string' versions are so generated tables can retain the
1332# hex format, which is the more familiar value
1333my $SBase_string = "0xAC00";
1334my $SBase = CORE::hex $SBase_string;
1335my $LBase_string = "0x1100";
1336my $LBase = CORE::hex $LBase_string;
1337my $VBase_string = "0x1161";
1338my $VBase = CORE::hex $VBase_string;
1339my $TBase_string = "0x11A7";
1340my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1341my $SCount = 11172;
1342my $LCount = 19;
1343my $VCount = 21;
1344my $TCount = 28;
1345my $NCount = $VCount * $TCount;
1346
1347# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1348# with the above published constants.
1349my %Jamo;
1350my %Jamo_L; # Leading consonants
1351my %Jamo_V; # Vowels
1352my %Jamo_T; # Trailing consonants
1353
bb1dd3da
KW
1354# For code points whose name contains its ordinal as a '-ABCD' suffix.
1355# The key is the base name of the code point, and the value is an
1356# array giving all the ranges that use this base name. Each range
1357# is actually a hash giving the 'low' and 'high' values of it.
1358my %names_ending_in_code_point;
1359my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1360 # removed from the names
1361# Inverse mapping. The list of ranges that have these kinds of
1362# names. Each element contains the low, high, and base names in an
1363# anonymous hash.
1364my @code_points_ending_in_code_point;
1365
6b5ab373
KW
1366# To hold Unicode's normalization test suite
1367my @normalization_tests;
1368
bb1dd3da
KW
1369# Boolean: does this Unicode version have the hangul syllables, and are we
1370# writing out a table for them?
1371my $has_hangul_syllables = 0;
1372
1373# Does this Unicode version have code points whose names end in their
1374# respective code points, and are we writing out a table for them? 0 for no;
1375# otherwise points to first property that a table is needed for them, so that
1376# if multiple tables are needed, we don't create duplicates
1377my $needing_code_points_ending_in_code_point = 0;
1378
37e2e78e 1379my @backslash_X_tests; # List of tests read in for testing \X
06ae2722 1380my @SB_tests; # List of tests read in for testing \b{sb}
ae3bb8ea 1381my @WB_tests; # List of tests read in for testing \b{wb}
99870f4d
KW
1382my @unhandled_properties; # Will contain a list of properties found in
1383 # the input that we didn't process.
f86864ac 1384my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1385 # listed in the pod
1386my @map_properties; # Properties that get map files written
1387my @named_sequences; # NamedSequences.txt contents.
1388my %potential_files; # Generated list of all .txt files in the directory
1389 # structure so we can warn if something is being
1390 # ignored.
a0204bb1
KW
1391my @missing_early_files; # Generated list of absent files that we need to
1392 # proceed in compiling this early Unicode version
99870f4d
KW
1393my @files_actually_output; # List of files we generated.
1394my @more_Names; # Some code point names are compound; this is used
1395 # to store the extra components of them.
1396my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1397 # the minimum before we consider it equivalent to a
1398 # candidate rational
1399my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1400
1401# These store references to certain commonly used property objects
3c88a801 1402my $ccc;
99870f4d
KW
1403my $gc;
1404my $perl;
1405my $block;
3e20195b
KW
1406my $perl_charname;
1407my $print;
ac7dbdb6 1408my $All;
bc0c431d 1409my $Assigned; # All assigned characters in this Unicode release
e47e66b9
KW
1410my $DI; # Default_Ignorable_Code_Point property
1411my $NChar; # Noncharacter_Code_Point property
359523e2 1412my $script;
99870f4d
KW
1413
1414# Are there conflicting names because of beginning with 'In_', or 'Is_'
1415my $has_In_conflicts = 0;
1416my $has_Is_conflicts = 0;
1417
1418sub internal_file_to_platform ($) {
1419 # Convert our file paths which have '/' separators to those of the
1420 # platform.
1421
1422 my $file = shift;
1423 return undef unless defined $file;
1424
1425 return File::Spec->join(split '/', $file);
d07a55ed 1426}
5beb625e 1427
99870f4d
KW
1428sub file_exists ($) { # platform independent '-e'. This program internally
1429 # uses slash as a path separator.
1430 my $file = shift;
1431 return 0 if ! defined $file;
1432 return -e internal_file_to_platform($file);
1433}
5beb625e 1434
99870f4d 1435sub objaddr($) {
23e33b60
KW
1436 # Returns the address of the blessed input object.
1437 # It doesn't check for blessedness because that would do a string eval
1438 # every call, and the program is structured so that this is never called
1439 # for a non-blessed object.
99870f4d 1440
23e33b60 1441 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1442
1443 # Numifying a ref gives its address.
051df77b 1444 return pack 'J', $_[0];
99870f4d
KW
1445}
1446
558712cf 1447# These are used only if $annotate is true.
c4019d52
KW
1448# The entire range of Unicode characters is examined to populate these
1449# after all the input has been processed. But most can be skipped, as they
1450# have the same descriptive phrases, such as being unassigned
1451my @viacode; # Contains the 1 million character names
4b06b8de 1452my @age; # And their ages ("" if none)
c4019d52
KW
1453my @printable; # boolean: And are those characters printable?
1454my @annotate_char_type; # Contains a type of those characters, specifically
1455 # for the purposes of annotation.
1456my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1457 # name for the purposes of annotation. They map to the
c4019d52
KW
1458 # upper edge of the range, so that the end point can
1459 # be immediately found. This is used to skip ahead to
1460 # the end of a range, and avoid processing each
1461 # individual code point in it.
1462my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1463 # characters, but excluding those which are
1464 # also noncharacter code points
1465
1466# The annotation types are an extension of the regular range types, though
1467# some of the latter are folded into one. Make the new types negative to
1468# avoid conflicting with the regular types
1469my $SURROGATE_TYPE = -1;
1470my $UNASSIGNED_TYPE = -2;
1471my $PRIVATE_USE_TYPE = -3;
1472my $NONCHARACTER_TYPE = -4;
1473my $CONTROL_TYPE = -5;
2d88a86a
KW
1474my $ABOVE_UNICODE_TYPE = -6;
1475my $UNKNOWN_TYPE = -7; # Used only if there is a bug in this program
c4019d52
KW
1476
1477sub populate_char_info ($) {
558712cf 1478 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1479 # input code point's info that are needed for outputting more detailed
1480 # comments. If calling context wants a return, it is the end point of
1481 # any contiguous range of characters that share essentially the same info
1482
1483 my $i = shift;
1484 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1485
1486 $viacode[$i] = $perl_charname->value_of($i) || "";
1487
1488 # A character is generally printable if Unicode says it is,
1489 # but below we make sure that most Unicode general category 'C' types
1490 # aren't.
1491 $printable[$i] = $print->contains($i);
1492
1493 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1494
1495 # Only these two regular types are treated specially for annotations
1496 # purposes
1497 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1498 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1499
1500 # Give a generic name to all code points that don't have a real name.
1501 # We output ranges, if applicable, for these. Also calculate the end
1502 # point of the range.
1503 my $end;
1504 if (! $viacode[$i]) {
2d88a86a
KW
1505 if ($i > $MAX_UNICODE_CODEPOINT) {
1506 $viacode[$i] = 'Above-Unicode';
1507 $annotate_char_type[$i] = $ABOVE_UNICODE_TYPE;
1508 $printable[$i] = 0;
1509 $end = $MAX_WORKING_CODEPOINT;
4b06b8de 1510 $age[$i] = "";
2d88a86a
KW
1511 }
1512 elsif ($gc-> table('Private_use')->contains($i)) {
c4019d52
KW
1513 $viacode[$i] = 'Private Use';
1514 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1515 $printable[$i] = 0;
1516 $end = $gc->table('Private_Use')->containing_range($i)->end;
4b06b8de 1517 $age[$i] = property_ref("Age")->value_of($i);
c4019d52 1518 }
e47e66b9 1519 elsif ($NChar->contains($i)) {
c4019d52
KW
1520 $viacode[$i] = 'Noncharacter';
1521 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1522 $printable[$i] = 0;
e47e66b9 1523 $end = $NChar->containing_range($i)->end;
4b06b8de 1524 $age[$i] = property_ref("Age")->value_of($i);
c4019d52
KW
1525 }
1526 elsif ($gc-> table('Control')->contains($i)) {
c71dea7f 1527 $viacode[$i] = property_ref('Name_Alias')->value_of($i) || 'Control';
c4019d52
KW
1528 $annotate_char_type[$i] = $CONTROL_TYPE;
1529 $printable[$i] = 0;
4b06b8de 1530 $age[$i] = property_ref("Age")->value_of($i);
c4019d52
KW
1531 }
1532 elsif ($gc-> table('Unassigned')->contains($i)) {
c4019d52
KW
1533 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1534 $printable[$i] = 0;
349edad4 1535 if (defined $block) { # No blocks in earliest releases
1d025d66
KW
1536 $viacode[$i] = 'Unassigned';
1537 $end = $gc-> table('Unassigned')->containing_range($i)->end;
1538 }
1539 else {
1540 $viacode[$i] = 'Unassigned, block=' . $block-> value_of($i);
c4019d52 1541
bf06c733
KW
1542 # Because we name the unassigned by the blocks they are in, it
1543 # can't go past the end of that block, and it also can't go
1544 # past the unassigned range it is in. The special table makes
1545 # sure that the non-characters, which are unassigned, are
1546 # separated out.
1547 $end = min($block->containing_range($i)->end,
1548 $unassigned_sans_noncharacters->
1549 containing_range($i)->end);
1d025d66 1550 }
4b06b8de 1551 $age[$i] = property_ref("Age")->value_of($i);
1d025d66 1552 }
35e2b416 1553 elsif ($perl->table('_Perl_Surrogate')->contains($i)) {
1d025d66
KW
1554 $viacode[$i] = 'Surrogate';
1555 $annotate_char_type[$i] = $SURROGATE_TYPE;
1556 $printable[$i] = 0;
1557 $end = $gc->table('Surrogate')->containing_range($i)->end;
4b06b8de 1558 $age[$i] = property_ref("Age")->value_of($i);
13ca76ff
KW
1559 }
1560 else {
1561 Carp::my_carp_bug("Can't figure out how to annotate "
1562 . sprintf("U+%04X", $i)
1563 . ". Proceeding anyway.");
c4019d52
KW
1564 $viacode[$i] = 'UNKNOWN';
1565 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1566 $printable[$i] = 0;
1567 }
1568 }
1569
1570 # Here, has a name, but if it's one in which the code point number is
1571 # appended to the name, do that.
1572 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1573 $viacode[$i] .= sprintf("-%04X", $i);
4b06b8de
KW
1574
1575 # Do all these as groups of the same age, instead of individually,
1576 # because their names are so meaningless, and there are typically
1577 # large quantities of them.
1578 my $Age = property_ref("Age");
1579 $age[$i] = $Age->value_of($i);
1580 my $limit = $perl_charname->containing_range($i)->end;
1581 $end = $i + 1;
1582 while ($end <= $limit && $Age->value_of($end) == $age[$i]) {
1583 $end++;
1584 }
1585 $end--;
c4019d52
KW
1586 }
1587
1588 # And here, has a name, but if it's a hangul syllable one, replace it with
1589 # the correct name from the Unicode algorithm
1590 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1591 use integer;
1592 my $SIndex = $i - $SBase;
1593 my $L = $LBase + $SIndex / $NCount;
1594 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1595 my $T = $TBase + $SIndex % $TCount;
1596 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1597 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
4b06b8de 1598 $age[$i] = property_ref("Age")->value_of($i);
c4019d52
KW
1599 $end = $perl_charname->containing_range($i)->end;
1600 }
4b06b8de
KW
1601 else {
1602 $age[$i] = property_ref("Age")->value_of($i);
1603 }
c4019d52
KW
1604
1605 return if ! defined wantarray;
1606 return $i if ! defined $end; # If not a range, return the input
1607
1608 # Save this whole range so can find the end point quickly
1609 $annotate_ranges->add_map($i, $end, $end);
1610
1611 return $end;
1612}
1613
23e33b60
KW
1614# Commented code below should work on Perl 5.8.
1615## This 'require' doesn't necessarily work in miniperl, and even if it does,
1616## the native perl version of it (which is what would operate under miniperl)
1617## is extremely slow, as it does a string eval every call.
7e017d6d 1618#my $has_fast_scalar_util = $^X !~ /miniperl/
23e33b60
KW
1619# && defined eval "require Scalar::Util";
1620#
1621#sub objaddr($) {
1622# # Returns the address of the blessed input object. Uses the XS version if
1623# # available. It doesn't check for blessedness because that would do a
1624# # string eval every call, and the program is structured so that this is
1625# # never called for a non-blessed object.
1626#
1627# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1628#
1629# # Check at least that is a ref.
1630# my $pkg = ref($_[0]) or return undef;
1631#
1632# # Change to a fake package to defeat any overloaded stringify
1633# bless $_[0], 'main::Fake';
1634#
1635# # Numifying a ref gives its address.
051df77b 1636# my $addr = pack 'J', $_[0];
23e33b60
KW
1637#
1638# # Return to original class
1639# bless $_[0], $pkg;
1640# return $addr;
1641#}
1642
99870f4d
KW
1643sub max ($$) {
1644 my $a = shift;
1645 my $b = shift;
1646 return $a if $a >= $b;
1647 return $b;
1648}
1649
1650sub min ($$) {
1651 my $a = shift;
1652 my $b = shift;
1653 return $a if $a <= $b;
1654 return $b;
1655}
1656
1657sub clarify_number ($) {
1658 # This returns the input number with underscores inserted every 3 digits
1659 # in large (5 digits or more) numbers. Input must be entirely digits, not
1660 # checked.
1661
1662 my $number = shift;
1663 my $pos = length($number) - 3;
1664 return $number if $pos <= 1;
1665 while ($pos > 0) {
1666 substr($number, $pos, 0) = '_';
1667 $pos -= 3;
5beb625e 1668 }
99870f4d 1669 return $number;
99598c8c
JH
1670}
1671
731cb813
KW
1672sub clarify_code_point_count ($) {
1673 # This is like clarify_number(), but the input is assumed to be a count of
1674 # code points, rather than a generic number.
1675
2d88a86a
KW
1676 my $append = "";
1677
1678 my $number = shift;
1679 if ($number > $MAX_UNICODE_CODEPOINTS) {
1680 $number -= ($MAX_WORKING_CODEPOINTS - $MAX_UNICODE_CODEPOINTS);
1681 return "All above-Unicode code points" if $number == 0;
1682 $append = " + all above-Unicode code points";
1683 }
1684 return clarify_number($number) . $append;
731cb813 1685}
12ac2576 1686
99870f4d 1687package Carp;
7ebf06b3 1688
99870f4d
KW
1689# These routines give a uniform treatment of messages in this program. They
1690# are placed in the Carp package to cause the stack trace to not include them,
1691# although an alternative would be to use another package and set @CARP_NOT
1692# for it.
12ac2576 1693
99870f4d 1694our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1695
99f78760
KW
1696# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1697# and overload trying to load Scalar:Util under miniperl. See
1698# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1699undef $overload::VERSION;
1700
99870f4d
KW
1701sub my_carp {
1702 my $message = shift || "";
1703 my $nofold = shift || 0;
7ebf06b3 1704
99870f4d
KW
1705 if ($message) {
1706 $message = main::join_lines($message);
1707 $message =~ s/^$0: *//; # Remove initial program name
1708 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1709 $message = "\n$0: $message;";
12ac2576 1710
99870f4d
KW
1711 # Fold the message with program name, semi-colon end punctuation
1712 # (which looks good with the message that carp appends to it), and a
1713 # hanging indent for continuation lines.
1714 $message = main::simple_fold($message, "", 4) unless $nofold;
1715 $message =~ s/\n$//; # Remove the trailing nl so what carp
1716 # appends is to the same line
1717 }
12ac2576 1718
99870f4d 1719 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1720
99870f4d
KW
1721 carp $message;
1722 return;
1723}
7ebf06b3 1724
99870f4d
KW
1725sub my_carp_bug {
1726 # This is called when it is clear that the problem is caused by a bug in
1727 # this program.
7ebf06b3 1728
99870f4d
KW
1729 my $message = shift;
1730 $message =~ s/^$0: *//;
1731 $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");
1732 carp $message;
1733 return;
1734}
7ebf06b3 1735
99870f4d
KW
1736sub carp_too_few_args {
1737 if (@_ != 2) {
1738 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1739 return;
12ac2576 1740 }
7ebf06b3 1741
99870f4d
KW
1742 my $args_ref = shift;
1743 my $count = shift;
7ebf06b3 1744
99870f4d
KW
1745 my_carp_bug("Need at least $count arguments to "
1746 . (caller 1)[3]
1747 . ". Instead got: '"
1748 . join ', ', @$args_ref
1749 . "'. No action taken.");
1750 return;
12ac2576
JP
1751}
1752
99870f4d
KW
1753sub carp_extra_args {
1754 my $args_ref = shift;
1755 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1756
99870f4d
KW
1757 unless (ref $args_ref) {
1758 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1759 return;
1760 }
1761 my ($package, $file, $line) = caller;
1762 my $subroutine = (caller 1)[3];
cf25bb62 1763
99870f4d
KW
1764 my $list;
1765 if (ref $args_ref eq 'HASH') {
1766 foreach my $key (keys %$args_ref) {
1767 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1768 }
99870f4d 1769 $list = join ', ', each %{$args_ref};
cf25bb62 1770 }
99870f4d
KW
1771 elsif (ref $args_ref eq 'ARRAY') {
1772 foreach my $arg (@$args_ref) {
1773 $arg = $UNDEF unless defined $arg;
1774 }
1775 $list = join ', ', @$args_ref;
1776 }
1777 else {
1778 my_carp_bug("Can't cope with ref "
1779 . ref($args_ref)
1780 . " . argument to 'carp_extra_args'. Not checking arguments.");
1781 return;
1782 }
1783
1784 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1785 return;
d73e5302
JH
1786}
1787
99870f4d
KW
1788package main;
1789
1790{ # Closure
1791
1792 # This program uses the inside-out method for objects, as recommended in
1b6b3fa9
KW
1793 # "Perl Best Practices". (This is the best solution still, since this has
1794 # to run under miniperl.) This closure aids in generating those. There
99870f4d
KW
1795 # are two routines. setup_package() is called once per package to set
1796 # things up, and then set_access() is called for each hash representing a
1797 # field in the object. These routines arrange for the object to be
1798 # properly destroyed when no longer used, and for standard accessor
1799 # functions to be generated. If you need more complex accessors, just
1800 # write your own and leave those accesses out of the call to set_access().
1801 # More details below.
1802
1803 my %constructor_fields; # fields that are to be used in constructors; see
1804 # below
1805
1806 # The values of this hash will be the package names as keys to other
1807 # hashes containing the name of each field in the package as keys, and
1808 # references to their respective hashes as values.
1809 my %package_fields;
1810
1811 sub setup_package {
1812 # Sets up the package, creating standard DESTROY and dump methods
1813 # (unless already defined). The dump method is used in debugging by
1814 # simple_dumper().
1815 # The optional parameters are:
1816 # a) a reference to a hash, that gets populated by later
1817 # set_access() calls with one of the accesses being
1818 # 'constructor'. The caller can then refer to this, but it is
1819 # not otherwise used by these two routines.
1820 # b) a reference to a callback routine to call during destruction
1821 # of the object, before any fields are actually destroyed
1822
1823 my %args = @_;
1824 my $constructor_ref = delete $args{'Constructor_Fields'};
1825 my $destroy_callback = delete $args{'Destroy_Callback'};
1826 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1827
1828 my %fields;
1829 my $package = (caller)[0];
1830
1831 $package_fields{$package} = \%fields;
1832 $constructor_fields{$package} = $constructor_ref;
1833
1834 unless ($package->can('DESTROY')) {
1835 my $destroy_name = "${package}::DESTROY";
1836 no strict "refs";
1837
1838 # Use typeglob to give the anonymous subroutine the name we want
1839 *$destroy_name = sub {
1840 my $self = shift;
ffe43484 1841 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1842
1843 $self->$destroy_callback if $destroy_callback;
1844 foreach my $field (keys %{$package_fields{$package}}) {
1845 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1846 delete $package_fields{$package}{$field}{$addr};
1847 }
1848 return;
1849 }
1850 }
1851
1852 unless ($package->can('dump')) {
1853 my $dump_name = "${package}::dump";
1854 no strict "refs";
1855 *$dump_name = sub {
1856 my $self = shift;
1857 return dump_inside_out($self, $package_fields{$package}, @_);
1858 }
1859 }
1860 return;
1861 }
1862
1863 sub set_access {
1864 # Arrange for the input field to be garbage collected when no longer
1865 # needed. Also, creates standard accessor functions for the field
1866 # based on the optional parameters-- none if none of these parameters:
1867 # 'addable' creates an 'add_NAME()' accessor function.
1868 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1869 # function.
1870 # 'settable' creates a 'set_NAME()' accessor function.
1871 # 'constructor' doesn't create an accessor function, but adds the
1872 # field to the hash that was previously passed to
1873 # setup_package();
1874 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1875 # 'add' etc. all mean 'addable'.
1876 # The read accessor function will work on both array and scalar
1877 # values. If another accessor in the parameter list is 'a', the read
1878 # access assumes an array. You can also force it to be array access
1879 # by specifying 'readable_array' instead of 'readable'
1880 #
1881 # A sort-of 'protected' access can be set-up by preceding the addable,
1882 # readable or settable with some initial portion of 'protected_' (but,
1883 # the underscore is required), like 'p_a', 'pro_set', etc. The
1884 # "protection" is only by convention. All that happens is that the
1885 # accessor functions' names begin with an underscore. So instead of
1886 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1887 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1888 # end of each package, and then storing the __LINE__ ranges and
1889 # checking them on every accessor. But that is way overkill.)
1890
1891 # We create anonymous subroutines as the accessors and then use
1892 # typeglobs to assign them to the proper package and name
1893
1894 my $name = shift; # Name of the field
1895 my $field = shift; # Reference to the inside-out hash containing the
1896 # field
1897
1898 my $package = (caller)[0];
1899
1900 if (! exists $package_fields{$package}) {
1901 croak "$0: Must call 'setup_package' before 'set_access'";
1902 }
d73e5302 1903
99870f4d
KW
1904 # Stash the field so DESTROY can get it.
1905 $package_fields{$package}{$name} = $field;
cf25bb62 1906
99870f4d
KW
1907 # Remaining arguments are the accessors. For each...
1908 foreach my $access (@_) {
1909 my $access = lc $access;
cf25bb62 1910
99870f4d 1911 my $protected = "";
cf25bb62 1912
99870f4d
KW
1913 # Match the input as far as it goes.
1914 if ($access =~ /^(p[^_]*)_/) {
1915 $protected = $1;
1916 if (substr('protected_', 0, length $protected)
1917 eq $protected)
1918 {
1919
1920 # Add 1 for the underscore not included in $protected
1921 $access = substr($access, length($protected) + 1);
1922 $protected = '_';
1923 }
1924 else {
1925 $protected = "";
1926 }
1927 }
1928
1929 if (substr('addable', 0, length $access) eq $access) {
1930 my $subname = "${package}::${protected}add_$name";
1931 no strict "refs";
1932
1933 # add_ accessor. Don't add if already there, which we
1934 # determine using 'eq' for scalars and '==' otherwise.
1935 *$subname = sub {
1936 use strict "refs";
1937 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1938 my $self = shift;
1939 my $value = shift;
ffe43484 1940 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1941 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1942 if (ref $value) {
f998e60c 1943 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1944 }
1945 else {
f998e60c 1946 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1947 }
f998e60c 1948 push @{$field->{$addr}}, $value;
99870f4d
KW
1949 return;
1950 }
1951 }
1952 elsif (substr('constructor', 0, length $access) eq $access) {
1953 if ($protected) {
1954 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1955 }
1956 else {
1957 $constructor_fields{$package}{$name} = $field;
1958 }
1959 }
1960 elsif (substr('readable_array', 0, length $access) eq $access) {
1961
1962 # Here has read access. If one of the other parameters for
1963 # access is array, or this one specifies array (by being more
1964 # than just 'readable_'), then create a subroutine that
1965 # assumes the data is an array. Otherwise just a scalar
1966 my $subname = "${package}::${protected}$name";
1967 if (grep { /^a/i } @_
1968 or length($access) > length('readable_'))
1969 {
1970 no strict "refs";
1971 *$subname = sub {
1972 use strict "refs";
23e33b60 1973 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 1974 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
1975 if (ref $field->{$addr} ne 'ARRAY') {
1976 my $type = ref $field->{$addr};
1977 $type = 'scalar' unless $type;
1978 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
1979 return;
1980 }
1981 return scalar @{$field->{$addr}} unless wantarray;
1982
1983 # Make a copy; had problems with caller modifying the
1984 # original otherwise
1985 my @return = @{$field->{$addr}};
1986 return @return;
1987 }
1988 }
1989 else {
1990
1991 # Here not an array value, a simpler function.
1992 no strict "refs";
1993 *$subname = sub {
1994 use strict "refs";
23e33b60 1995 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 1996 no overloading;
051df77b 1997 return $field->{pack 'J', $_[0]};
99870f4d
KW
1998 }
1999 }
2000 }
2001 elsif (substr('settable', 0, length $access) eq $access) {
2002 my $subname = "${package}::${protected}set_$name";
2003 no strict "refs";
2004 *$subname = sub {
2005 use strict "refs";
23e33b60
KW
2006 if (main::DEBUG) {
2007 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
2008 Carp::carp_extra_args(\@_) if @_ > 2;
2009 }
2010 # $self is $_[0]; $value is $_[1]
f998e60c 2011 no overloading;
051df77b 2012 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
2013 return;
2014 }
2015 }
2016 else {
2017 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
2018 }
cf25bb62 2019 }
99870f4d 2020 return;
cf25bb62 2021 }
99870f4d
KW
2022}
2023
2024package Input_file;
2025
2026# All input files use this object, which stores various attributes about them,
2027# and provides for convenient, uniform handling. The run method wraps the
2028# processing. It handles all the bookkeeping of opening, reading, and closing
2029# the file, returning only significant input lines.
2030#
2031# Each object gets a handler which processes the body of the file, and is
74cd47d0
KW
2032# called by run(). All character property files must use the generic,
2033# default handler, which has code scrubbed to handle things you might not
2034# expect, including automatic EBCDIC handling. For files that don't deal with
2035# mapping code points to a property value, such as test files,
2036# PropertyAliases, PropValueAliases, and named sequences, you can override the
2037# handler to be a custom one. Such a handler should basically be a
2038# while(next_line()) {...} loop.
99870f4d
KW
2039#
2040# You can also set up handlers to
40045c2d 2041# 0) call during object construction time, after everything else is done
537124e4 2042# 1) call before the first line is read, for pre processing
83b68635
KW
2043# 2) call to adjust each line of the input before the main handler gets
2044# them. This can be automatically generated, if appropriately simple
2045# enough, by specifiying a Properties parameter in the constructor.
99870f4d 2046# 3) call upon EOF before the main handler exits its loop
537124e4 2047# 4) call at the end, for post processing
99870f4d
KW
2048#
2049# $_ is used to store the input line, and is to be filtered by the
2050# each_line_handler()s. So, if the format of the line is not in the desired
2051# format for the main handler, these are used to do that adjusting. They can
2052# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
3848e5b7
KW
2053# so the $_ output of one is used as the input to the next. The eof handler
2054# is also stackable, but none of the others are, but could easily be changed
2055# to be so.
99870f4d 2056#
a0204bb1
KW
2057# Some properties are used by the Perl core but aren't defined until later
2058# Unicode releases. The perl interpreter would have problems working when
2059# compiled with an earlier Unicode version that doesn't have them, so we need
2060# to define them somehow for those releases. The 'Early' constructor
2061# parameter can be used to automatically handle this. It is essentially
2062# ignored if the Unicode version being compiled has a data file for this
2063# property. Either code to execute or a file to read can be specified.
2064# Details are at the %early definition.
2065#
99870f4d
KW
2066# Most of the handlers can call insert_lines() or insert_adjusted_lines()
2067# which insert the parameters as lines to be processed before the next input
3848e5b7 2068# file line is read. This allows the EOF handler(s) to flush buffers, for
99870f4d
KW
2069# example. The difference between the two routines is that the lines inserted
2070# by insert_lines() are subjected to the each_line_handler()s. (So if you
9be22ca5
KW
2071# called it from such a handler, you would get infinite recursion without some
2072# mechanism to prevent that.) Lines inserted by insert_adjusted_lines() go
2073# directly to the main handler without any adjustments. If the
2074# post-processing handler calls any of these, there will be no effect. Some
2075# error checking for these conditions could be added, but it hasn't been done.
99870f4d
KW
2076#
2077# carp_bad_line() should be called to warn of bad input lines, which clears $_
2078# to prevent further processing of the line. This routine will output the
2079# message as a warning once, and then keep a count of the lines that have the
2080# same message, and output that count at the end of the file's processing.
2081# This keeps the number of messages down to a manageable amount.
2082#
2083# get_missings() should be called to retrieve any @missing input lines.
2084# Messages will be raised if this isn't done if the options aren't to ignore
2085# missings.
2086
2087sub trace { return main::trace(@_); }
2088
99870f4d
KW
2089{ # Closure
2090 # Keep track of fields that are to be put into the constructor.
2091 my %constructor_fields;
2092
2093 main::setup_package(Constructor_Fields => \%constructor_fields);
2094
2095 my %file; # Input file name, required
2096 main::set_access('file', \%file, qw{ c r });
2097
2098 my %first_released; # Unicode version file was first released in, required
2099 main::set_access('first_released', \%first_released, qw{ c r });
2100
2101 my %handler; # Subroutine to process the input file, defaults to
2102 # 'process_generic_property_file'
2103 main::set_access('handler', \%handler, qw{ c });
2104
2105 my %property;
2106 # name of property this file is for. defaults to none, meaning not
2107 # applicable, or is otherwise determinable, for example, from each line.
696609bf 2108 main::set_access('property', \%property, qw{ c r });
99870f4d
KW
2109
2110 my %optional;
3274d896
KW
2111 # This is either an unsigned number, or a list of property names. In the
2112 # former case, if it is non-zero, it means the file is optional, so if the
2113 # file is absent, no warning about that is output. In the latter case, it
2114 # is a list of properties that the file (exclusively) defines. If the
2115 # file is present, tables for those properties will be produced; if
2116 # absent, none will, even if they are listed elsewhere (namely
2117 # PropertyAliases.txt and PropValueAliases.txt) as being in this release,
2118 # and no warnings will be raised about them not being available. (And no
2119 # warning about the file itself will be raised.)
2120 main::set_access('optional', \%optional, qw{ c readable_array } );
99870f4d
KW
2121
2122 my %non_skip;
2123 # This is used for debugging, to skip processing of all but a few input
2124 # files. Add 'non_skip => 1' to the constructor for those files you want
2125 # processed when you set the $debug_skip global.
2126 main::set_access('non_skip', \%non_skip, 'c');
2127
37e2e78e 2128 my %skip;
0dee21d6
KW
2129 # This is used to skip processing of this input file (semi-) permanently.
2130 # The value should be the reason the file is being skipped. It is used
2131 # for files that we aren't planning to process anytime soon, but want to
2132 # allow to be in the directory and be checked for their names not
2133 # conflicting with any other files on a DOS 8.3 name filesystem, but to
2134 # not otherwise be processed, and to not raise a warning about not being
2135 # handled. In the constructor call, any value that evaluates to a numeric
2136 # 0 or undef means don't skip. Any other value is a string giving the
2137 # reason it is being skippped, and this will appear in generated pod.
2138 # However, an empty string reason will suppress the pod entry.
2139 # Internally, calls that evaluate to numeric 0 are changed into undef to
2140 # distinguish them from an empty string call.
2141 main::set_access('skip', \%skip, 'c', 'r');
37e2e78e 2142
99870f4d
KW
2143 my %each_line_handler;
2144 # list of subroutines to look at and filter each non-comment line in the
2145 # file. defaults to none. The subroutines are called in order, each is
2146 # to adjust $_ for the next one, and the final one adjusts it for
2147 # 'handler'
2148 main::set_access('each_line_handler', \%each_line_handler, 'c');
2149
83b68635
KW
2150 my %properties; # Optional ordered list of the properties that occur in each
2151 # meaningful line of the input file. If present, an appropriate
2152 # each_line_handler() is automatically generated and pushed onto the stack
2153 # of such handlers. This is useful when a file contains multiple
2154 # proerties per line, but no other special considerations are necessary.
2155 # The special value "<ignored>" means to discard the corresponding input
2156 # field.
2157 # Any @missing lines in the file should also match this syntax; no such
2158 # files exist as of 6.3. But if it happens in a future release, the code
2159 # could be expanded to properly parse them.
2160 main::set_access('properties', \%properties, qw{ c r });
2161
99870f4d
KW
2162 my %has_missings_defaults;
2163 # ? Are there lines in the file giving default values for code points
2164 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
2165 # the norm, but IGNORED means it has such lines, but the handler doesn't
2166 # use them. Having these three states allows us to catch changes to the
83b68635
KW
2167 # UCD that this program should track. XXX This could be expanded to
2168 # specify the syntax for such lines, like %properties above.
99870f4d
KW
2169 main::set_access('has_missings_defaults',
2170 \%has_missings_defaults, qw{ c r });
2171
40045c2d
KW
2172 my %construction_time_handler;
2173 # Subroutine to call at the end of the new method. If undef, no such
2174 # handler is called.
2175 main::set_access('construction_time_handler',
2176 \%construction_time_handler, qw{ c });
2177
99870f4d
KW
2178 my %pre_handler;
2179 # Subroutine to call before doing anything else in the file. If undef, no
2180 # such handler is called.
2181 main::set_access('pre_handler', \%pre_handler, qw{ c });
2182
2183 my %eof_handler;
3848e5b7 2184 # Subroutines to call upon getting an EOF on the input file, but before
99870f4d
KW
2185 # that is returned to the main handler. This is to allow buffers to be
2186 # flushed. The handler is expected to call insert_lines() or
2187 # insert_adjusted() with the buffered material
3848e5b7 2188 main::set_access('eof_handler', \%eof_handler, qw{ c });
99870f4d
KW
2189
2190 my %post_handler;
2191 # Subroutine to call after all the lines of the file are read in and
9be22ca5
KW
2192 # processed. If undef, no such handler is called. Note that this cannot
2193 # add lines to be processed; instead use eof_handler
99870f4d
KW
2194 main::set_access('post_handler', \%post_handler, qw{ c });
2195
2196 my %progress_message;
2197 # Message to print to display progress in lieu of the standard one
2198 main::set_access('progress_message', \%progress_message, qw{ c });
2199
2200 my %handle;
2201 # cache open file handle, internal. Is undef if file hasn't been
2202 # processed at all, empty if has;
2203 main::set_access('handle', \%handle);
2204
2205 my %added_lines;
2206 # cache of lines added virtually to the file, internal
2207 main::set_access('added_lines', \%added_lines);
2208
74cd47d0
KW
2209 my %remapped_lines;
2210 # cache of lines added virtually to the file, internal
2211 main::set_access('remapped_lines', \%remapped_lines);
2212
99870f4d
KW
2213 my %errors;
2214 # cache of errors found, internal
2215 main::set_access('errors', \%errors);
2216
2217 my %missings;
2218 # storage of '@missing' defaults lines
2219 main::set_access('missings', \%missings);
2220
a0204bb1
KW
2221 my %early;
2222 # Used for properties that must be defined (for Perl's purposes) on
2223 # versions of Unicode earlier than Unicode itself defines them. The
2224 # parameter is an array (it would be better to be a hash, but not worth
2225 # bothering about due to its rare use).
2226 #
2227 # The first element is either a code reference to call when in a release
2228 # earlier than the Unicode file is available in, or it is an alternate
2229 # file to use instead of the non-existent one. This file must have been
2230 # plunked down in the same directory as mktables. Should you be compiling
2231 # on a release that needs such a file, mktables will abort the
2232 # compilation, and tell you where to get the necessary file(s), and what
2233 # name(s) to use to store them as.
2234 # In the case of specifying an alternate file, the array must contain two
2235 # further elements:
2236 #
2237 # [1] is the name of the property that will be generated by this file.
2238 # The class automatically takes the input file and excludes any code
2239 # points in it that were not assigned in the Unicode version being
2240 # compiled. It then uses this result to define the property in the given
2241 # version. Since the property doesn't actually exist in the Unicode
2242 # version being compiled, this should be a name accessible only by core
2243 # perl. If it is the same name as the regular property, the constructor
2244 # will mark the output table as a $PLACEHOLDER so that it doesn't actually
2245 # get output, and so will be unusable by non-core code. Otherwise it gets
2246 # marked as $INTERNAL_ONLY.
2247 #
2248 # [2] is a property value to assign (only when compiling Unicode 1.1.5) to
2249 # the Hangul syllables in that release (which were ripped out in version
2250 # 2) for the given property . (Hence it is ignored except when compiling
2251 # version 1. You only get one value that applies to all of them, which
2252 # may not be the actual reality, but probably nobody cares anyway for
2253 # these obsolete characters.)
2254 #
2255 # Not all files can be handled in the above way, and so the code ref
2256 # alternative is available. It can do whatever it needs to. The other
2257 # array elements are optional in this case, and the code is free to use or
2258 # ignore them if they are present.
2259 #
2260 # Internally, the constructor unshifts a 0 or 1 onto this array to
2261 # indicate if an early alternative is actually being used or not. This
2262 # makes for easier testing later on.
2263 main::set_access('early', \%early, 'c');
2264
0dee21d6
KW
2265 my %required_even_in_debug_skip;
2266 # debug_skip is used to speed up compilation during debugging by skipping
2267 # processing files that are not needed for the task at hand. However,
2268 # some files pretty much can never be skipped, and this is used to specify
2269 # that this is one of them. In order to skip this file, the call to the
2270 # constructor must be edited to comment out this parameter.
2271 main::set_access('required_even_in_debug_skip',
2272 \%required_even_in_debug_skip, 'c');
2273
d8062573
KW
2274 my %withdrawn;
2275 # Some files get removed from the Unicode DB. This is a version object
2276 # giving the first release without this file.
2277 main::set_access('withdrawn', \%withdrawn, 'c');
2278
0dee21d6
KW
2279 my %in_this_release;
2280 # Calculated value from %first_released and %withdrawn. Are we compiling
2281 # a Unicode release which includes this file?
2282 main::set_access('in_this_release', \%in_this_release);
2283
74cd47d0
KW
2284 sub _next_line;
2285 sub _next_line_with_remapped_range;
2286
99870f4d
KW
2287 sub new {
2288 my $class = shift;
2289
2290 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2291 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2292
2293 # Set defaults
2294 $handler{$addr} = \&main::process_generic_property_file;
2295 $non_skip{$addr} = 0;
0dee21d6 2296 $skip{$addr} = undef;
99870f4d
KW
2297 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2298 $handle{$addr} = undef;
2299 $added_lines{$addr} = [ ];
74cd47d0 2300 $remapped_lines{$addr} = [ ];
99870f4d 2301 $each_line_handler{$addr} = [ ];
3848e5b7 2302 $eof_handler{$addr} = [ ];
99870f4d
KW
2303 $errors{$addr} = { };
2304 $missings{$addr} = [ ];
a0204bb1 2305 $early{$addr} = [ ];
3274d896 2306 $optional{$addr} = [ ];
99870f4d
KW
2307
2308 # Two positional parameters.
99f78760 2309 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2310 $file{$addr} = main::internal_file_to_platform(shift);
2311 $first_released{$addr} = shift;
2312
2313 # The rest of the arguments are key => value pairs
2314 # %constructor_fields has been set up earlier to list all possible
2315 # ones. Either set or push, depending on how the default has been set
2316 # up just above.
2317 my %args = @_;
2318 foreach my $key (keys %args) {
2319 my $argument = $args{$key};
2320
2321 # Note that the fields are the lower case of the constructor keys
2322 my $hash = $constructor_fields{lc $key};
2323 if (! defined $hash) {
2324 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2325 next;
2326 }
2327 if (ref $hash->{$addr} eq 'ARRAY') {
2328 if (ref $argument eq 'ARRAY') {
2329 foreach my $argument (@{$argument}) {
2330 next if ! defined $argument;
2331 push @{$hash->{$addr}}, $argument;
2332 }
2333 }
2334 else {
2335 push @{$hash->{$addr}}, $argument if defined $argument;
2336 }
2337 }
2338 else {
2339 $hash->{$addr} = $argument;
2340 }
2341 delete $args{$key};
2342 };
2343
0dee21d6
KW
2344 $non_skip{$addr} = 1 if $required_even_in_debug_skip{$addr};
2345
2346 # Convert 0 (meaning don't skip) to undef
2347 undef $skip{$addr} unless $skip{$addr};
2348
3274d896
KW
2349 # Handle the case where this file is optional
2350 my $pod_message_for_non_existent_optional = "";
2351 if ($optional{$addr}->@*) {
2352
2353 # First element is the pod message
2354 $pod_message_for_non_existent_optional
2355 = shift $optional{$addr}->@*;
2356 # Convert a 0 'Optional' argument to an empty list to make later
2357 # code more concise.
2358 if ( $optional{$addr}->@*
2359 && $optional{$addr}->@* == 1
2360 && $optional{$addr}[0] ne ""
2361 && $optional{$addr}[0] !~ /\D/
2362 && $optional{$addr}[0] == 0)
2363 {
2364 $optional{$addr} = [ ];
2365 }
2366 else { # But if the only element doesn't evaluate to 0, make sure
2367 # that this file is indeed considered optional below.
2368 unshift $optional{$addr}->@*, 1;
2369 }
2370 }
2371
0dee21d6 2372 my $progress;
a0204bb1
KW
2373 my $function_instead_of_file = 0;
2374
2375 # If we are compiling a Unicode release earlier than the file became
2376 # available, the constructor may have supplied a substitute
2377 if ($first_released{$addr} gt $v_version && $early{$addr}->@*) {
2378
2379 # Yes, we have a substitute, that we will use; mark it so
2380 unshift $early{$addr}->@*, 1;
2381
2382 # See the definition of %early for what the array elements mean.
2383 # If we have a property this defines, create a table and default
2384 # map for it now (at essentially compile time), so that it will be
2385 # available for the whole of run time. (We will want to add this
2386 # name as an alias when we are using the official property name;
2387 # but this must be deferred until run(), because at construction
2388 # time the official names have yet to be defined.)
2389 if ($early{$addr}[2]) {
2390 my $fate = ($property{$addr}
2391 && $property{$addr} eq $early{$addr}[2])
2392 ? $PLACEHOLDER
2393 : $INTERNAL_ONLY;
2394 my $prop_object = Property->new($early{$addr}[2],
2395 Fate => $fate,
2396 Perl_Extension => 1,
2397 );
2398
2399 # Use the default mapping for the regular property for this
2400 # substitute one.
2401 if ( defined $property{$addr}
2402 && defined $default_mapping{$property{$addr}})
2403 {
2404 $prop_object
2405 ->set_default_map($default_mapping{$property{$addr}});
2406 }
2407 }
2408
2409 if (ref $early{$addr}[1] eq 'CODE') {
2410 $function_instead_of_file = 1;
2411
2412 # If the first element of the array is a code ref, the others
2413 # are optional.
2414 $handler{$addr} = $early{$addr}[1];
2415 $property{$addr} = $early{$addr}[2]
2416 if defined $early{$addr}[2];
2417 $progress = "substitute $file{$addr}";
2418
2419 undef $file{$addr};
2420 }
2421 else { # Specifying a substitute file
2422
2423 if (! main::file_exists($early{$addr}[1])) {
2424
2425 # If we don't see the substitute file, generate an error
2426 # message giving the needed things, and add it to the list
2427 # of such to output before actual processing happens
2428 # (hence the user finds out all of them in one run).
ce8d64d9
KW
2429 # Instead of creating a general method for NameAliases,
2430 # hard-code it here, as there is unlikely to ever be a
2431 # second one which needs special handling.
2432 my $string_version = ($file{$addr} eq "NameAliases.txt")
2433 ? 'at least 6.1 (the later, the better)'
2434 : sprintf "%vd", $first_released{$addr};
a0204bb1
KW
2435 push @missing_early_files, <<END;
2436'$file{$addr}' version $string_version should be copied to '$early{$addr}[1]'.
2437END
2438 ;
2439 return;
2440 }
2441 $progress = $early{$addr}[1];
2442 $progress .= ", substituting for $file{$addr}" if $file{$addr};
2443 $file{$addr} = $early{$addr}[1];
2444 $property{$addr} = $early{$addr}[2];
0dee21d6 2445
a0204bb1
KW
2446 # Ignore code points not in the version being compiled
2447 push $each_line_handler{$addr}->@*, \&_exclude_unassigned;
2448
2449 if ( $v_version lt v2.0 # Hanguls in this release ...
2450 && defined $early{$addr}[3]) # ... need special treatment
2451 {
2452 push $eof_handler{$addr}->@*, \&_fixup_obsolete_hanguls;
2453 }
2454 }
2455
2456 # And this substitute is valid for all releases.
2457 $first_released{$addr} = v0;
2458 }
2459 else { # Normal behavior
0dee21d6 2460 $progress = $file{$addr};
a0204bb1 2461 unshift $early{$addr}->@*, 0; # No substitute
99870f4d
KW
2462 }
2463
0dee21d6
KW
2464 my $file = $file{$addr};
2465 $progress_message{$addr} = "Processing $progress"
2466 unless $progress_message{$addr};
2467
d8062573
KW
2468 # A file should be there if it is within the window of versions for
2469 # which Unicode supplies it
2470 if ($withdrawn{$addr} && $withdrawn{$addr} le $v_version) {
2471 $in_this_release{$addr} = 0;
2472 $skip{$addr} = "";
2473 }
2474 else {
0dee21d6
KW
2475 $in_this_release{$addr} = $first_released{$addr} le $v_version;
2476
a0204bb1
KW
2477 # Check that the file for this object (possibly using a substitute
2478 # for early releases) exists or we have a function alternative
2479 if ( ! $function_instead_of_file
2480 && ! main::file_exists($file))
0dee21d6
KW
2481 {
2482 # Here there is nothing available for this release. This is
2483 # fine if we aren't expecting anything in this release.
2484 if (! $in_this_release{$addr}) {
2485 $skip{$addr} = ""; # Don't remark since we expected
2486 # nothing and got nothing
2487 }
3274d896 2488 elsif ($optional{$addr}->@*) {
0dee21d6 2489
3274d896
KW
2490 # Here the file is optional in this release; Use the
2491 # passed in text to document this case in the pod.
2492 $skip{$addr} = $pod_message_for_non_existent_optional;
0dee21d6
KW
2493 }
2494 elsif ( $in_this_release{$addr}
2495 && ! defined $skip{$addr}
2496 && defined $file)
2497 { # Doesn't exist but should.
2498 $skip{$addr} = "'$file' not found. Possibly Big problems";
2499 Carp::my_carp($skip{$addr});
2500 }
2501 }
2502 elsif ($debug_skip && ! defined $skip{$addr} && ! $non_skip{$addr})
2503 {
2504
2505 # The file exists; if not skipped for another reason, and we are
2506 # skipping most everything during debugging builds, use that as
2507 # the skip reason.
2508 $skip{$addr} = '$debug_skip is on'
2509 }
d8062573 2510 }
0dee21d6
KW
2511
2512 if ( ! $debug_skip
2513 && $non_skip{$addr}
2514 && ! $required_even_in_debug_skip{$addr}
2515 && $verbosity)
2516 {
2517 print "Warning: " . __PACKAGE__ . " constructor for $file has useless 'non_skip' in it\n";
a3a8c5f0 2518 }
99870f4d 2519
0dee21d6 2520 # Here, we have figured out if we will be skipping this file or not.
3274d896
KW
2521 # If so, we add any single property it defines to any passed in
2522 # optional property list. These will be dealt with at run time.
0dee21d6 2523 if (defined $skip{$addr}) {
3274d896
KW
2524 if ($property{$addr}) {
2525 push $optional{$addr}->@*, $property{$addr};
2526 }
2527 } # Otherwise, are going to process the file.
0dee21d6
KW
2528 elsif ($property{$addr}) {
2529
2530 # If the file has a property defined in the constructor for it, it
2531 # means that the property is not listed in the file's entries. So
2532 # add a handler (to the list of line handlers) to insert the
2533 # property name into the lines, to provide a uniform interface to
2534 # the final processing subroutine.
2535 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
09ca89ce 2536 }
83b68635
KW
2537 elsif ($properties{$addr}) {
2538
9be22ca5
KW
2539 # Similarly, there may be more than one property represented on
2540 # each line, with no clue but the constructor input what those
2541 # might be. Add a handler for each line in the input so that it
2542 # creates a separate input line for each property in those input
2543 # lines, thus making them suitable to handle generically.
83b68635
KW
2544
2545 push @{$each_line_handler{$addr}},
2546 sub {
2547 my $file = shift;
2548 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2549
2550 my @fields = split /\s*;\s*/, $_, -1;
2551
2552 if (@fields - 1 > @{$properties{$addr}}) {
2553 $file->carp_bad_line('Extra fields');
2554 $_ = "";
2555 return;
2556 }
2557 my $range = shift @fields; # 0th element is always the
2558 # range
2559
2560 # The next fields in the input line correspond
2561 # respectively to the stored properties.
2562 for my $i (0 .. @{$properties{$addr}} - 1) {
2563 my $property_name = $properties{$addr}[$i];
2564 next if $property_name eq '<ignored>';
2565 $file->insert_adjusted_lines(
2566 "$range; $property_name; $fields[$i]");
2567 }
2568 $_ = "";
2569
2570 return;
2571 };
2572 }
37e2e78e 2573
9be22ca5 2574 { # On non-ascii platforms, we use a special pre-handler
74cd47d0
KW
2575 no strict;
2576 no warnings 'once';
2577 *next_line = (main::NON_ASCII_PLATFORM)
2578 ? *_next_line_with_remapped_range
2579 : *_next_line;
2580 }
2581
40045c2d
KW
2582 &{$construction_time_handler{$addr}}($self)
2583 if $construction_time_handler{$addr};
2584
99870f4d 2585 return $self;
d73e5302
JH
2586 }
2587
cf25bb62 2588
99870f4d
KW
2589 use overload
2590 fallback => 0,
2591 qw("") => "_operator_stringify",
2592 "." => \&main::_operator_dot,
1285127e 2593 ".=" => \&main::_operator_dot_equal,
99870f4d 2594 ;
cf25bb62 2595
99870f4d
KW
2596 sub _operator_stringify {
2597 my $self = shift;
cf25bb62 2598
99870f4d 2599 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2600 }
d73e5302 2601
99870f4d
KW
2602 sub run {
2603 # Process the input object $self. This opens and closes the file and
2604 # calls all the handlers for it. Currently, this can only be called
3848e5b7 2605 # once per file, as it destroy's the EOF handlers
d73e5302 2606
0dee21d6
KW
2607 # flag to make sure extracted files are processed early
2608 state $seen_non_extracted_non_age = 0;
2609
99870f4d
KW
2610 my $self = shift;
2611 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2612
ffe43484 2613 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2614
99870f4d 2615 my $file = $file{$addr};
d73e5302 2616
0dee21d6
KW
2617 if (! $file) {
2618 $handle{$addr} = 'pretend_is_open';
99870f4d
KW
2619 }
2620 else {
99870f4d 2621 if ($seen_non_extracted_non_age) {
9be22ca5
KW
2622 if ($file =~ /$EXTRACTED/i) # Some platforms may change the
2623 # case of the file's name
2624 {
1675ea0d 2625 Carp::my_carp_bug(main::join_lines(<<END
99f78760 2626$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2627anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2628have subtle problems
2629END
2630 ));
2631 }
2632 }
2633 elsif ($EXTRACTED_DIR
0dee21d6
KW
2634
2635 # We only do this check for generic property files
2636 && $handler{$addr} == \&main::process_generic_property_file
2637
517956bf
CB
2638 && $file !~ /$EXTRACTED/i
2639 && lc($file) ne 'dage.txt')
99870f4d
KW
2640 {
2641 # We don't set this (by the 'if' above) if we have no
2642 # extracted directory, so if running on an early version,
2643 # this test won't work. Not worth worrying about.
2644 $seen_non_extracted_non_age = 1;
2645 }
2646
9be22ca5 2647 # Mark the file as having being processed, and warn if it
99870f4d
KW
2648 # isn't a file we are expecting. As we process the files,
2649 # they are deleted from the hash, so any that remain at the
2650 # end of the program are files that we didn't process.
517956bf 2651 my $fkey = File::Spec->rel2abs($file);
0dee21d6 2652 my $exists = delete $potential_files{lc($fkey)};
faf3cf6b 2653
0dee21d6
KW
2654 Carp::my_carp("Was not expecting '$file'.")
2655 if $exists && ! $in_this_release{$addr};
2656
a0204bb1
KW
2657 # If there is special handling for compiling Unicode releases
2658 # earlier than the first one in which Unicode defines this
2659 # property ...
2660 if ($early{$addr}->@* > 1) {
2661
2662 # Mark as processed any substitute file that would be used in
2663 # such a release
2664 $fkey = File::Spec->rel2abs($early{$addr}[1]);
2665 delete $potential_files{lc($fkey)};
2666
2667 # As commented in the constructor code, when using the
2668 # official property, we still have to allow the publicly
2669 # inaccessible early name so that the core code which uses it
2670 # will work regardless.
2671 if (! $early{$addr}[0] && $early{$addr}->@* > 2) {
2672 my $early_property_name = $early{$addr}[2];
2673 if ($property{$addr} ne $early_property_name) {
2674 main::property_ref($property{$addr})
2675 ->add_alias($early_property_name);
2676 }
2677 }
2678 }
2679
0dee21d6
KW
2680 # We may be skipping this file ...
2681 if (defined $skip{$addr}) {
2682
2683 # If the file isn't supposed to be in this release, there is
2684 # nothing to do
2685 if ($in_this_release{$addr}) {
2686
2687 # But otherwise, we may print a message
2688 if ($debug_skip) {
2689 print STDERR "Skipping input file '$file'",
2690 " because '$skip{$addr}'\n";
2691 }
2692
2693 # And add it to the list of skipped files, which is later
2694 # used to make the pod
2695 $skipped_files{$file} = $skip{$addr};
3274d896
KW
2696
2697 # The 'optional' list contains properties that are also to
2698 # be skipped along with the file. (There may also be
2699 # digits which are just placeholders to make sure it isn't
2700 # an empty list
2701 foreach my $property ($optional{$addr}->@*) {
2702 next unless $property =~ /\D/;
2703 my $prop_object = main::property_ref($property);
2704 next unless defined $prop_object;
2705 $prop_object->set_fate($SUPPRESSED, $skip{$addr});
2706 }
0dee21d6 2707 }
99870f4d 2708
37e2e78e
KW
2709 return;
2710 }
2711
9be22ca5
KW
2712 # Here, we are going to process the file. Open it, converting the
2713 # slashes used in this program into the proper form for the OS
99870f4d
KW
2714 my $file_handle;
2715 if (not open $file_handle, "<", $file) {
2716 Carp::my_carp("Can't open $file. Skipping: $!");
d3664eb9 2717 return;
99870f4d
KW
2718 }
2719 $handle{$addr} = $file_handle; # Cache the open file handle
9e65c3f4 2720
0dee21d6
KW
2721 # If possible, make sure that the file is the correct version.
2722 # (This data isn't available on early Unicode releases or in
2723 # UnicodeData.txt.) We don't do this check if we are using a
2724 # substitute file instead of the official one (though the code
2725 # could be extended to do so).
2726 if ($in_this_release{$addr}
a0204bb1 2727 && ! $early{$addr}[0]
0dee21d6
KW
2728 && lc($file) ne 'unicodedata.txt')
2729 {
96f226dc 2730 if ($file !~ /^Unihan/i) {
0dee21d6
KW
2731
2732 # The non-Unihan files started getting version numbers in
2733 # 3.2, but some files in 4.0 are unchanged from 3.2, and
2734 # marked as 3.2. 4.0.1 is the first version where there
2735 # are no files marked as being from less than 4.0, though
2736 # some are marked as 4.0. In versions after that, the
2737 # numbers are correct.
2738 if ($v_version ge v4.0.1) {
2739 $_ = <$file_handle>; # The version number is in the
2740 # very first line
2741 if ($_ !~ / - $string_version \. /x) {
2742 chomp;
2743 $_ =~ s/^#\s*//;
2744
2745 # 4.0.1 had some valid files that weren't updated.
2746 if (! ($v_version eq v4.0.1 && $_ =~ /4\.0\.0/)) {
2747 die Carp::my_carp("File '$file' is version "
2748 . "'$_'. It should be "
2749 . "version $string_version");
2750 }
2751 }
cafe9cf0 2752 }
96f226dc 2753 }
0dee21d6
KW
2754 elsif ($v_version ge v6.0.0) { # Unihan
2755
2756 # Unihan files didn't get accurate version numbers until
2757 # 6.0. The version is somewhere in the first comment
2758 # block
96f226dc
KW
2759 while (<$file_handle>) {
2760 if ($_ !~ /^#/) {
0dee21d6
KW
2761 Carp::my_carp_bug("Could not find the expected "
2762 . "version info in file '$file'");
96f226dc
KW
2763 last;
2764 }
2765 chomp;
2766 $_ =~ s/^#\s*//;
2767 next if $_ !~ / version: /x;
2768 last if $_ =~ /$string_version/;
0dee21d6
KW
2769 die Carp::my_carp("File '$file' is version "
2770 . "'$_'. It should be "
2771 . "version $string_version");
96f226dc
KW
2772 }
2773 }
9e65c3f4 2774 }
99870f4d
KW
2775 }
2776
0dee21d6 2777 print "$progress_message{$addr}\n" if $verbosity >= $PROGRESS;
99870f4d
KW
2778
2779 # Call any special handler for before the file.
2780 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2781
2782 # Then the main handler
2783 &{$handler{$addr}}($self);
2784
2785 # Then any special post-file handler.
2786 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2787
2788 # If any errors have been accumulated, output the counts (as the first
2789 # error message in each class was output when it was encountered).
2790 if ($errors{$addr}) {
2791 my $total = 0;
2792 my $types = 0;
2793 foreach my $error (keys %{$errors{$addr}}) {
2794 $total += $errors{$addr}->{$error};
2795 delete $errors{$addr}->{$error};
2796 $types++;
2797 }
2798 if ($total > 1) {
2799 my $message
2800 = "A total of $total lines had errors in $file. ";
2801
2802 $message .= ($types == 1)
2803 ? '(Only the first one was displayed.)'
2804 : '(Only the first of each type was displayed.)';
2805 Carp::my_carp($message);
2806 }
2807 }
2808
2809 if (@{$missings{$addr}}) {
2810 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2811 }
2812
2813 # If a real file handle, close it.
2814 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2815 ref $handle{$addr};
2816 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2817 # the file, as opposed to undef
2818 return;
2819 }
2820
74cd47d0 2821 sub _next_line {
99870f4d
KW
2822 # Sets $_ to be the next logical input line, if any. Returns non-zero
2823 # if such a line exists. 'logical' means that any lines that have
2824 # been added via insert_lines() will be returned in $_ before the file
2825 # is read again.
2826
2827 my $self = shift;
2828 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2829
ffe43484 2830 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2831
2832 # Here the file is open (or if the handle is not a ref, is an open
2833 # 'virtual' file). Get the next line; any inserted lines get priority
2834 # over the file itself.
2835 my $adjusted;
2836
2837 LINE:
2838 while (1) { # Loop until find non-comment, non-empty line
2839 #local $to_trace = 1 if main::DEBUG;
2840 my $inserted_ref = shift @{$added_lines{$addr}};
2841 if (defined $inserted_ref) {
2842 ($adjusted, $_) = @{$inserted_ref};
2843 trace $adjusted, $_ if main::DEBUG && $to_trace;
2844 return 1 if $adjusted;
2845 }
2846 else {
2847 last if ! ref $handle{$addr}; # Don't read unless is real file
2848 last if ! defined ($_ = readline $handle{$addr});
2849 }
2850 chomp;
2851 trace $_ if main::DEBUG && $to_trace;
2852
2853 # See if this line is the comment line that defines what property
2854 # value that code points that are not listed in the file should
2855 # have. The format or existence of these lines is not guaranteed
2856 # by Unicode since they are comments, but the documentation says
2857 # that this was added for machine-readability, so probably won't
2858 # change. This works starting in Unicode Version 5.0. They look
2859 # like:
2860 #
2861 # @missing: 0000..10FFFF; Not_Reordered
2862 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2863 # @missing: 0000..10FFFF; ; NaN
2864 #
2865 # Save the line for a later get_missings() call.
2866 if (/$missing_defaults_prefix/) {
2867 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2868 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2869 }
2870 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2871 my @defaults = split /\s* ; \s*/x, $_;
2872
2873 # The first field is the @missing, which ends in a
2874 # semi-colon, so can safely shift.
2875 shift @defaults;
2876
2877 # Some of these lines may have empty field placeholders
2878 # which get in the way. An example is:
2879 # @missing: 0000..10FFFF; ; NaN
2880 # Remove them. Process starting from the top so the
2881 # splice doesn't affect things still to be looked at.
2882 for (my $i = @defaults - 1; $i >= 0; $i--) {
2883 next if $defaults[$i] ne "";
2884 splice @defaults, $i, 1;
2885 }
2886
2887 # What's left should be just the property (maybe) and the
2888 # default. Having only one element means it doesn't have
2889 # the property.
2890 my $default;
2891 my $property;
2892 if (@defaults >= 1) {
2893 if (@defaults == 1) {
2894 $default = $defaults[0];
2895 }
2896 else {
2897 $property = $defaults[0];
2898 $default = $defaults[1];
2899 }
2900 }
2901
2902 if (@defaults < 1
2903 || @defaults > 2
2904 || ($default =~ /^</
2905 && $default !~ /^<code *point>$/i
09f8d0ac
KW
2906 && $default !~ /^<none>$/i
2907 && $default !~ /^<script>$/i))
99870f4d
KW
2908 {
2909 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2910 }
2911 else {
2912
2913 # If the property is missing from the line, it should
2914 # be the one for the whole file
2915 $property = $property{$addr} if ! defined $property;
2916
2917 # Change <none> to the null string, which is what it
2918 # really means. If the default is the code point
2919 # itself, set it to <code point>, which is what
2920 # Unicode uses (but sometimes they've forgotten the
2921 # space)
2922 if ($default =~ /^<none>$/i) {
2923 $default = "";
2924 }
2925 elsif ($default =~ /^<code *point>$/i) {
2926 $default = $CODE_POINT;
2927 }
09f8d0ac
KW
2928 elsif ($default =~ /^<script>$/i) {
2929
2930 # Special case this one. Currently is from
2931 # ScriptExtensions.txt, and means for all unlisted
2932 # code points, use their Script property values.
2933 # For the code points not listed in that file, the
2934 # default value is 'Unknown'.
2935 $default = "Unknown";
2936 }
99870f4d
KW
2937
2938 # Store them as a sub-arrays with both components.
2939 push @{$missings{$addr}}, [ $default, $property ];
2940 }
2941 }
2942
2943 # There is nothing for the caller to process on this comment
2944 # line.
2945 next;
2946 }
2947
2948 # Remove comments and trailing space, and skip this line if the
2949 # result is empty
2950 s/#.*//;
2951 s/\s+$//;
2952 next if /^$/;
2953
2954 # Call any handlers for this line, and skip further processing of
2955 # the line if the handler sets the line to null.
2956 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2957 &{$sub_ref}($self);
2958 next LINE if /^$/;
2959 }
2960
2961 # Here the line is ok. return success.
2962 return 1;
2963 } # End of looping through lines.
2964
3848e5b7 2965 # If there are EOF handlers, call each (only once) and if it generates
99870f4d 2966 # more lines to process go back in the loop to handle them.
3848e5b7
KW
2967 while ($eof_handler{$addr}->@*) {
2968 &{$eof_handler{$addr}[0]}($self);
2969 shift $eof_handler{$addr}->@*; # Currently only get one shot at it.
99870f4d
KW
2970 goto LINE if $added_lines{$addr};
2971 }
2972
2973 # Return failure -- no more lines.
2974 return 0;
2975
2976 }
2977
74cd47d0
KW
2978 sub _next_line_with_remapped_range {
2979 my $self = shift;
2980 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2981
2982 # like _next_line(), but for use on non-ASCII platforms. It sets $_
2983 # to be the next logical input line, if any. Returns non-zero if such
2984 # a line exists. 'logical' means that any lines that have been added
2985 # via insert_lines() will be returned in $_ before the file is read
2986 # again.
2987 #
2988 # The difference from _next_line() is that this remaps the Unicode
2989 # code points in the input to those of the native platform. Each
2990 # input line contains a single code point, or a single contiguous
2991 # range of them This routine splits each range into its individual
2992 # code points and caches them. It returns the cached values,
2993 # translated into their native equivalents, one at a time, for each
2994 # call, before reading the next line. Since native values can only be
2995 # a single byte wide, no translation is needed for code points above
2996 # 0xFF, and ranges that are entirely above that number are not split.
2997 # If an input line contains the range 254-1000, it would be split into
2998 # three elements: 254, 255, and 256-1000. (The downstream table
2999 # insertion code will sort and coalesce the individual code points
3000 # into appropriate ranges.)
3001
3002 my $addr = do { no overloading; pack 'J', $self; };
3003
3004 while (1) {
3005
3006 # Look in cache before reading the next line. Return any cached
3007 # value, translated
3008 my $inserted = shift @{$remapped_lines{$addr}};
3009 if (defined $inserted) {
3010 trace $inserted if main::DEBUG && $to_trace;
3011 $_ = $inserted =~ s/^ ( \d+ ) /sprintf("%04X", utf8::unicode_to_native($1))/xer;
3012 trace $_ if main::DEBUG && $to_trace;
3013 return 1;
3014 }
3015
3016 # Get the next line.
3017 return 0 unless _next_line($self);
3018
3019 # If there is a special handler for it, return the line,
3020 # untranslated. This should happen only for files that are
3021 # special, not being code-point related, such as property names.
3022 return 1 if $handler{$addr}
3023 != \&main::process_generic_property_file;
3024
3025 my ($range, $property_name, $map, @remainder)
3026 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3027
3028 if (@remainder
3029 || ! defined $property_name
3030 || $range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3031 {
3032 Carp::my_carp_bug("Unrecognized input line '$_'. Ignored");
3033 }
3034
3035 my $low = hex $1;
3036 my $high = (defined $2) ? hex $2 : $low;
3037
3038 # If the input maps the range to another code point, remap the
3039 # target if it is between 0 and 255.
3040 my $tail;
3041 if (defined $map) {
3042 $map =~ s/\b 00 ( [0-9A-F]{2} ) \b/sprintf("%04X", utf8::unicode_to_native(hex $1))/gxe;
3043 $tail = "$property_name; $map";
3044 $_ = "$range; $tail";
3045 }
3046 else {
3047 $tail = $property_name;
3048 }
3049
3050 # If entire range is above 255, just return it, unchanged (except
3051 # any mapped-to code point, already changed above)
3052 return 1 if $low > 255;
3053
3054 # Cache an entry for every code point < 255. For those in the
3055 # range above 255, return a dummy entry for just that portion of
3056 # the range. Note that this will be out-of-order, but that is not
3057 # a problem.
3058 foreach my $code_point ($low .. $high) {
3059 if ($code_point > 255) {
3060 $_ = sprintf "%04X..%04X; $tail", $code_point, $high;
3061 return 1;
3062 }
3063 push @{$remapped_lines{$addr}}, "$code_point; $tail";
3064 }
3065 } # End of looping through lines.
3066
3067 # NOTREACHED
3068 }
3069
99870f4d
KW
3070# Not currently used, not fully tested.
3071# sub peek {
3072# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
3073# # record. Not callable from an each_line_handler(), nor does it call
3074# # an each_line_handler() on the line.
3075#
3076# my $self = shift;
ffe43484 3077# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3078#
3079# foreach my $inserted_ref (@{$added_lines{$addr}}) {
3080# my ($adjusted, $line) = @{$inserted_ref};
3081# next if $adjusted;
3082#
3083# # Remove comments and trailing space, and return a non-empty
3084# # resulting line
3085# $line =~ s/#.*//;
3086# $line =~ s/\s+$//;
3087# return $line if $line ne "";
3088# }
3089#
3090# return if ! ref $handle{$addr}; # Don't read unless is real file
3091# while (1) { # Loop until find non-comment, non-empty line
3092# local $to_trace = 1 if main::DEBUG;
3093# trace $_ if main::DEBUG && $to_trace;
3094# return if ! defined (my $line = readline $handle{$addr});
3095# chomp $line;
3096# push @{$added_lines{$addr}}, [ 0, $line ];
3097#
3098# $line =~ s/#.*//;
3099# $line =~ s/\s+$//;
3100# return $line if $line ne "";
3101# }
3102#
3103# return;
3104# }
3105
3106
3107 sub insert_lines {
3108 # Lines can be inserted so that it looks like they were in the input
3109 # file at the place it was when this routine is called. See also
3110 # insert_adjusted_lines(). Lines inserted via this routine go through
3111 # any each_line_handler()
3112
3113 my $self = shift;
3114
3115 # Each inserted line is an array, with the first element being 0 to
3116 # indicate that this line hasn't been adjusted, and needs to be
3117 # processed.
f998e60c 3118 no overloading;
051df77b 3119 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
3120 return;
3121 }
3122
3123 sub insert_adjusted_lines {
3124 # Lines can be inserted so that it looks like they were in the input
3125 # file at the place it was when this routine is called. See also
3126 # insert_lines(). Lines inserted via this routine are already fully
3127 # adjusted, ready to be processed; each_line_handler()s handlers will
3128 # not be called. This means this is not a completely general
3129 # facility, as only the last each_line_handler on the stack should
3130 # call this. It could be made more general, by passing to each of the
3131 # line_handlers their position on the stack, which they would pass on
3132 # to this routine, and that would replace the boolean first element in
3133 # the anonymous array pushed here, so that the next_line routine could
3134 # use that to call only those handlers whose index is after it on the
3135 # stack. But this is overkill for what is needed now.
3136
3137 my $self = shift;
3138 trace $_[0] if main::DEBUG && $to_trace;
3139
3140 # Each inserted line is an array, with the first element being 1 to
3141 # indicate that this line has been adjusted
f998e60c 3142 no overloading;
051df77b 3143 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
3144 return;
3145 }
3146
3147 sub get_missings {
3148 # Returns the stored up @missings lines' values, and clears the list.
3149 # The values are in an array, consisting of the default in the first
3150 # element, and the property in the 2nd. However, since these lines
3151 # can be stacked up, the return is an array of all these arrays.
3152
3153 my $self = shift;
3154 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3155
ffe43484 3156 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3157
3158 # If not accepting a list return, just return the first one.
3159 return shift @{$missings{$addr}} unless wantarray;
3160
3161 my @return = @{$missings{$addr}};
3162 undef @{$missings{$addr}};
3163 return @return;
3164 }
3165
a0204bb1
KW
3166 sub _exclude_unassigned {
3167
3168 # Takes the range in $_ and excludes code points that aren't assigned
3169 # in this release
3170
3171 state $skip_inserted_count = 0;
3172
3173 # Ignore recursive calls.
3174 if ($skip_inserted_count) {
3175 $skip_inserted_count--;
3176 return;
3177 }
3178
3179 # Find what code points are assigned in this release
3180 main::calculate_Assigned() if ! defined $Assigned;
3181
3182 my $self = shift;
3183 my $addr = do { no overloading; pack 'J', $self; };
3184 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3185
3186 my ($range, @remainder)
3187 = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
3188
3189 # Examine the range.
3190 if ($range =~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
3191 {
3192 my $low = hex $1;
3193 my $high = (defined $2) ? hex $2 : $low;
3194
3195 # Split the range into subranges of just those code points in it
3196 # that are assigned.
3197 my @ranges = (Range_List->new(Initialize
3198 => Range->new($low, $high)) & $Assigned)->ranges;
3199
3200 # Do nothing if nothing in the original range is assigned in this
3201 # release; handle normally if everything is in this release.
3202 if (! @ranges) {
3203 $_ = "";
3204 }
3205 elsif (@ranges != 1) {
3206
3207 # Here, some code points in the original range aren't in this
3208 # release; @ranges gives the ones that are. Create fake input
3209 # lines for each of the ranges, and set things up so that when
3210 # this routine is called on that fake input, it will do
3211 # nothing.
3212 $skip_inserted_count = @ranges;
3213 my $remainder = join ";", @remainder;
3214 for my $range (@ranges) {
3215 $self->insert_lines(sprintf("%04X..%04X;%s",
3216 $range->start, $range->end, $remainder));
3217 }
3218 $_ = ""; # The original range is now defunct.
3219 }
3220 }
3221
3222 return;
3223 }
3224
3225 sub _fixup_obsolete_hanguls {
3226
3227 # This is called only when compiling Unicode version 1. All Unicode
3228 # data for subsequent releases assumes that the code points that were
3229 # Hangul syllables in this release only are something else, so if
3230 # using such data, we have to override it
3231
3232 my $self = shift;
3233 my $addr = do { no overloading; pack 'J', $self; };
3234 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3235
3236 my $object = main::property_ref($property{$addr});
53adf6a2
KW
3237 $object->add_map($FIRST_REMOVED_HANGUL_SYLLABLE,
3238 $FINAL_REMOVED_HANGUL_SYLLABLE,
a0204bb1
KW
3239 $early{$addr}[3], # Passed-in value for these
3240 Replace => $UNCONDITIONALLY);
3241 }
3242
99870f4d
KW
3243 sub _insert_property_into_line {
3244 # Add a property field to $_, if this file requires it.
3245
f998e60c 3246 my $self = shift;
ffe43484 3247 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 3248 my $property = $property{$addr};
99870f4d
KW
3249 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3250
3251 $_ =~ s/(;|$)/; $property$1/;
3252 return;
3253 }
3254
3255 sub carp_bad_line {
3256 # Output consistent error messages, using either a generic one, or the
3257 # one given by the optional parameter. To avoid gazillions of the
3258 # same message in case the syntax of a file is way off, this routine
3259 # only outputs the first instance of each message, incrementing a
3260 # count so the totals can be output at the end of the file.
3261
3262 my $self = shift;
3263 my $message = shift;
3264 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3265
ffe43484 3266 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3267
3268 $message = 'Unexpected line' unless $message;
3269
3270 # No trailing punctuation so as to fit with our addenda.
3271 $message =~ s/[.:;,]$//;
3272
3273 # If haven't seen this exact message before, output it now. Otherwise
3274 # increment the count of how many times it has occurred
3275 unless ($errors{$addr}->{$message}) {
3276 Carp::my_carp("$message in '$_' in "
f998e60c 3277 . $file{$addr}
99870f4d
KW
3278 . " at line $.. Skipping this line;");
3279 $errors{$addr}->{$message} = 1;
3280 }
3281 else {
3282 $errors{$addr}->{$message}++;
3283 }
3284
3285 # Clear the line to prevent any further (meaningful) processing of it.
3286 $_ = "";
3287
3288 return;
3289 }
3290} # End closure
3291
3292package Multi_Default;
3293
3294# Certain properties in early versions of Unicode had more than one possible
3295# default for code points missing from the files. In these cases, one
3296# default applies to everything left over after all the others are applied,
3297# and for each of the others, there is a description of which class of code
3298# points applies to it. This object helps implement this by storing the
3299# defaults, and for all but that final default, an eval string that generates
3300# the class that it applies to.
3301
3302
3303{ # Closure
3304
3305 main::setup_package();
3306
3307 my %class_defaults;
3308 # The defaults structure for the classes
3309 main::set_access('class_defaults', \%class_defaults);
3310
3311 my %other_default;
3312 # The default that applies to everything left over.
3313 main::set_access('other_default', \%other_default, 'r');
3314
3315
3316 sub new {
3317 # The constructor is called with default => eval pairs, terminated by
3318 # the left-over default. e.g.
3319 # Multi_Default->new(
3320 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
3321 # - 0x200D',
3322 # 'R' => 'some other expression that evaluates to code points',
3323 # .
3324 # .
3325 # .
3326 # 'U'));
68c3923f
KW
3327 # It is best to leave the final value be the one that matches the
3328 # above-Unicode code points.
99870f4d
KW
3329
3330 my $class = shift;
3331
3332 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 3333 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3334
3335 while (@_ > 1) {
3336 my $default = shift;
3337 my $eval = shift;
3338 $class_defaults{$addr}->{$default} = $eval;
3339 }
3340
3341 $other_default{$addr} = shift;
3342
3343 return $self;
3344 }
3345
3346 sub get_next_defaults {
3347 # Iterates and returns the next class of defaults.
3348 my $self = shift;
3349 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3350
ffe43484 3351 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3352
3353 return each %{$class_defaults{$addr}};
3354 }
3355}
3356
3357package Alias;
3358
3359# An alias is one of the names that a table goes by. This class defines them
3360# including some attributes. Everything is currently setup in the
3361# constructor.
3362
3363
3364{ # Closure
3365
3366 main::setup_package();
3367
3368 my %name;
3369 main::set_access('name', \%name, 'r');
3370
3371 my %loose_match;
c12f2655 3372 # Should this name match loosely or not.
99870f4d
KW
3373 main::set_access('loose_match', \%loose_match, 'r');
3374
33e96e72
KW
3375 my %make_re_pod_entry;
3376 # Some aliases should not get their own entries in the re section of the
3377 # pod, because they are covered by a wild-card, and some we want to
3378 # discourage use of. Binary
f82fe4ba 3379 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r', 's');
99870f4d 3380
fd1e3e84
KW
3381 my %ucd;
3382 # Is this documented to be accessible via Unicode::UCD
3383 main::set_access('ucd', \%ucd, 'r', 's');
3384
99870f4d
KW
3385 my %status;
3386 # Aliases have a status, like deprecated, or even suppressed (which means
3387 # they don't appear in documentation). Enum
3388 main::set_access('status', \%status, 'r');
3389
0eac1e20 3390 my %ok_as_filename;
99870f4d
KW
3391 # Similarly, some aliases should not be considered as usable ones for
3392 # external use, such as file names, or we don't want documentation to
3393 # recommend them. Boolean
0eac1e20 3394 main::set_access('ok_as_filename', \%ok_as_filename, 'r');
99870f4d
KW
3395
3396 sub new {
3397 my $class = shift;
3398
3399 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3400 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3401
3402 $name{$addr} = shift;
3403 $loose_match{$addr} = shift;
33e96e72 3404 $make_re_pod_entry{$addr} = shift;
0eac1e20 3405 $ok_as_filename{$addr} = shift;
99870f4d 3406 $status{$addr} = shift;
fd1e3e84 3407 $ucd{$addr} = shift;
99870f4d
KW
3408
3409 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3410
3411 # Null names are never ok externally
0eac1e20 3412 $ok_as_filename{$addr} = 0 if $name{$addr} eq "";
99870f4d
KW
3413
3414 return $self;
3415 }
3416}
3417
3418package Range;
3419
3420# A range is the basic unit for storing code points, and is described in the
3421# comments at the beginning of the program. Each range has a starting code
3422# point; an ending code point (not less than the starting one); a value
3423# that applies to every code point in between the two end-points, inclusive;
3424# and an enum type that applies to the value. The type is for the user's
3425# convenience, and has no meaning here, except that a non-zero type is
3426# considered to not obey the normal Unicode rules for having standard forms.
3427#
3428# The same structure is used for both map and match tables, even though in the
3429# latter, the value (and hence type) is irrelevant and could be used as a
3430# comment. In map tables, the value is what all the code points in the range
3431# map to. Type 0 values have the standardized version of the value stored as
3432# well, so as to not have to recalculate it a lot.
3433
3434sub trace { return main::trace(@_); }
3435
3436{ # Closure
3437
3438 main::setup_package();
3439
3440 my %start;
3441 main::set_access('start', \%start, 'r', 's');
3442
3443 my %end;
3444 main::set_access('end', \%end, 'r', 's');
3445
3446 my %value;
3447 main::set_access('value', \%value, 'r');
3448
3449 my %type;
3450 main::set_access('type', \%type, 'r');
3451
3452 my %standard_form;
3453 # The value in internal standard form. Defined only if the type is 0.
3454 main::set_access('standard_form', \%standard_form);
3455
3456 # Note that if these fields change, the dump() method should as well
3457
3458 sub new {
3459 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
3460 my $class = shift;
3461
3462 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3463 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3464
3465 $start{$addr} = shift;
3466 $end{$addr} = shift;
3467
3468 my %args = @_;
3469
3470 my $value = delete $args{'Value'}; # Can be 0
3471 $value = "" unless defined $value;
3472 $value{$addr} = $value;
3473
3474 $type{$addr} = delete $args{'Type'} || 0;
3475
3476 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3477
99870f4d
KW
3478 return $self;
3479 }
3480
3481 use overload
3482 fallback => 0,
3483 qw("") => "_operator_stringify",
3484 "." => \&main::_operator_dot,
1285127e 3485 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3486 ;
3487
3488 sub _operator_stringify {
3489 my $self = shift;
ffe43484 3490 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3491
3492 # Output it like '0041..0065 (value)'
3493 my $return = sprintf("%04X", $start{$addr})
3494 . '..'
3495 . sprintf("%04X", $end{$addr});
3496 my $value = $value{$addr};
3497 my $type = $type{$addr};
3498 $return .= ' (';
3499 $return .= "$value";
3500 $return .= ", Type=$type" if $type != 0;
3501 $return .= ')';
3502
3503 return $return;
3504 }
3505
3506 sub standard_form {
c292d35a
NC
3507 # Calculate the standard form only if needed, and cache the result.
3508 # The standard form is the value itself if the type is special.
3509 # This represents a considerable CPU and memory saving - at the time
3510 # of writing there are 368676 non-special objects, but the standard
3511 # form is only requested for 22047 of them - ie about 6%.
99870f4d
KW
3512
3513 my $self = shift;
3514 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3515
ffe43484 3516 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3517
3518 return $standard_form{$addr} if defined $standard_form{$addr};
c292d35a
NC
3519
3520 my $value = $value{$addr};
3521 return $value if $type{$addr};
3522 return $standard_form{$addr} = main::standardize($value);
99870f4d
KW
3523 }
3524
3525 sub dump {
3526 # Human, not machine readable. For machine readable, comment out this
3527 # entire routine and let the standard one take effect.
3528 my $self = shift;
3529 my $indent = shift;
3530 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3531
ffe43484 3532 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3533
3534 my $return = $indent
3535 . sprintf("%04X", $start{$addr})
3536 . '..'
3537 . sprintf("%04X", $end{$addr})
3538 . " '$value{$addr}';";
3539 if (! defined $standard_form{$addr}) {
3540 $return .= "(type=$type{$addr})";
3541 }
3542 elsif ($standard_form{$addr} ne $value{$addr}) {
3543 $return .= "(standard '$standard_form{$addr}')";
3544 }
3545 return $return;
3546 }
3547} # End closure
3548
3549package _Range_List_Base;
3550
3551# Base class for range lists. A range list is simply an ordered list of
3552# ranges, so that the ranges with the lowest starting numbers are first in it.
3553#
3554# When a new range is added that is adjacent to an existing range that has the
3555# same value and type, it merges with it to form a larger range.
3556#
3557# Ranges generally do not overlap, except that there can be multiple entries
3558# of single code point ranges. This is because of NameAliases.txt.
3559#
3560# In this program, there is a standard value such that if two different
3561# values, have the same standard value, they are considered equivalent. This
3562# value was chosen so that it gives correct results on Unicode data
3563
3564# There are a number of methods to manipulate range lists, and some operators
3565# are overloaded to handle them.
3566
99870f4d
KW
3567sub trace { return main::trace(@_); }
3568
3569{ # Closure
3570
3571 our $addr;
3572
5b348b71
KW
3573 # Max is initialized to a negative value that isn't adjacent to 0, for
3574 # simpler tests
3575 my $max_init = -2;
3576
99870f4d
KW
3577 main::setup_package();
3578
3579 my %ranges;
3580 # The list of ranges
3581 main::set_access('ranges', \%ranges, 'readable_array');
3582
3583 my %max;
3584 # The highest code point in the list. This was originally a method, but
3585 # actual measurements said it was used a lot.
3586 main::set_access('max', \%max, 'r');
3587
3588 my %each_range_iterator;
3589 # Iterator position for each_range()
3590 main::set_access('each_range_iterator', \%each_range_iterator);
3591
3592 my %owner_name_of;
3593 # Name of parent this is attached to, if any. Solely for better error
3594 # messages.
3595 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
3596
3597 my %_search_ranges_cache;
3598 # A cache of the previous result from _search_ranges(), for better
3599 # performance
3600 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
3601
3602 sub new {
3603 my $class = shift;
3604 my %args = @_;
3605
3606 # Optional initialization data for the range list.
3607 my $initialize = delete $args{'Initialize'};
3608
3609 my $self;
3610
3611 # Use _union() to initialize. _union() returns an object of this
3612 # class, which means that it will call this constructor recursively.
3613 # But it won't have this $initialize parameter so that it won't
3614 # infinitely loop on this.
3615 return _union($class, $initialize, %args) if defined $initialize;
3616
3617 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 3618 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3619
3620 # Optional parent object, only for debug info.
3621 $owner_name_of{$addr} = delete $args{'Owner'};
3622 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
3623
3624 # Stringify, in case it is an object.
3625 $owner_name_of{$addr} = "$owner_name_of{$addr}";
3626
3627 # This is used only for error messages, and so a colon is added
3628 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
3629
3630 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3631
5b348b71 3632 $max{$addr} = $max_init;
99870f4d
KW
3633
3634 $_search_ranges_cache{$addr} = 0;
3635 $ranges{$addr} = [];
3636
3637 return $self;
3638 }
3639
3640 use overload
3641 fallback => 0,
3642 qw("") => "_operator_stringify",
3643 "." => \&main::_operator_dot,
1285127e 3644 ".=" => \&main::_operator_dot_equal,
99870f4d
KW
3645 ;
3646
3647 sub _operator_stringify {
3648 my $self = shift;
ffe43484 3649 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3650
3651 return "Range_List attached to '$owner_name_of{$addr}'"
3652 if $owner_name_of{$addr};
3653 return "anonymous Range_List " . \$self;
3654 }
3655
3656 sub _union {
3657 # Returns the union of the input code points. It can be called as
3658 # either a constructor or a method. If called as a method, the result
3659 # will be a new() instance of the calling object, containing the union
3660 # of that object with the other parameter's code points; if called as
d59563d0 3661 # a constructor, the first parameter gives the class that the new object
99870f4d
KW
3662 # should be, and the second parameter gives the code points to go into
3663 # it.
3664 # In either case, there are two parameters looked at by this routine;
3665 # any additional parameters are passed to the new() constructor.
3666 #
3667 # The code points can come in the form of some object that contains
3668 # ranges, and has a conventionally named method to access them; or
3669 # they can be an array of individual code points (as integers); or
3670 # just a single code point.
3671 #
3672 # If they are ranges, this routine doesn't make any effort to preserve
3198cc57
KW
3673 # the range values and types of one input over the other. Therefore
3674 # this base class should not allow _union to be called from other than
99870f4d
KW
3675 # initialization code, so as to prevent two tables from being added
3676 # together where the range values matter. The general form of this
3677 # routine therefore belongs in a derived class, but it was moved here
3678 # to avoid duplication of code. The failure to overload this in this
3679 # class keeps it safe.
3198cc57
KW
3680 #
3681 # It does make the effort during initialization to accept tables with
3682 # multiple values for the same code point, and to preserve the order
3683 # of these. If there is only one input range or range set, it doesn't
3684 # sort (as it should already be sorted to the desired order), and will
3685 # accept multiple values per code point. Otherwise it will merge
3686 # multiple values into a single one.
99870f4d
KW
3687
3688 my $self;
3689 my @args; # Arguments to pass to the constructor
3690
3691 my $class = shift;
3692
3693 # If a method call, will start the union with the object itself, and
3694 # the class of the new object will be the same as self.
3695 if (ref $class) {
3696 $self = $class;
3697 $class = ref $self;
3698 push @args, $self;
3699 }
3700
3701 # Add the other required parameter.
3702 push @args, shift;
3703 # Rest of parameters are passed on to the constructor
3704
3705 # Accumulate all records from both lists.
3706 my @records;
3198cc57 3707 my $input_count = 0;
99870f4d
KW
3708 for my $arg (@args) {
3709 #local $to_trace = 0 if main::DEBUG;
3710 trace "argument = $arg" if main::DEBUG && $to_trace;
3