This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perluniprops: Add info about unused Unicode files
[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
23e33b60
KW
7# Needs 'no overloading' to run faster on miniperl. Code commented out at the
8# subroutine objaddr can be used instead to work as far back (untested) as
f998e60c
KW
9# 5.8: needs pack "U". But almost all occurrences of objaddr have been
10# removed in favor of using 'no overloading'. You also would have to go
11# through and replace occurrences like:
ffe43484 12# my $addr = do { no overloading; pack 'J', $self; }
f998e60c
KW
13# with
14# my $addr = main::objaddr $self;
6c68572b 15# (or reverse commit 9b01bafde4b022706c3d6f947a0963f821b2e50b
051df77b
NC
16# that instituted the change to main::objaddr, and subsequent commits that
17# changed 0+$self to pack 'J', $self.)
6c68572b 18
cdcef19a 19my $start_time;
98dc9551 20BEGIN { # Get the time the script started running; do it at compilation to
cdcef19a
KW
21 # get it as close as possible
22 $start_time= time;
23}
24
23e33b60 25require 5.010_001;
d73e5302 26use strict;
99870f4d 27use warnings;
cf25bb62 28use Carp;
bd9ebcfd 29use Config;
99870f4d
KW
30use File::Find;
31use File::Path;
d07a55ed 32use File::Spec;
99870f4d
KW
33use Text::Tabs;
34
35sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 36my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d
KW
37
38##########################################################################
39#
40# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
41# from the Unicode database files (lib/unicore/.../*.txt), It also generates
42# a pod file and a .t file
43#
44# The structure of this file is:
45# First these introductory comments; then
46# code needed for everywhere, such as debugging stuff; then
47# code to handle input parameters; then
48# data structures likely to be of external interest (some of which depend on
49# the input parameters, so follows them; then
50# more data structures and subroutine and package (class) definitions; then
51# the small actual loop to process the input files and finish up; then
52# a __DATA__ section, for the .t tests
53#
5f7264c7 54# This program works on all releases of Unicode through at least 6.0. The
99870f4d
KW
55# outputs have been scrutinized most intently for release 5.1. The others
56# have been checked for somewhat more than just sanity. It can handle all
57# existing Unicode character properties in those releases.
58#
99870f4d
KW
59# This program is mostly about Unicode character (or code point) properties.
60# A property describes some attribute or quality of a code point, like if it
61# is lowercase or not, its name, what version of Unicode it was first defined
62# in, or what its uppercase equivalent is. Unicode deals with these disparate
63# possibilities by making all properties into mappings from each code point
64# into some corresponding value. In the case of it being lowercase or not,
65# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
66# property maps each Unicode code point to a single value, called a "property
67# value". (Hence each Unicode property is a true mathematical function with
68# exactly one value per code point.)
69#
70# When using a property in a regular expression, what is desired isn't the
71# mapping of the code point to its property's value, but the reverse (or the
72# mathematical "inverse relation"): starting with the property value, "Does a
73# code point map to it?" These are written in a "compound" form:
74# \p{property=value}, e.g., \p{category=punctuation}. This program generates
75# files containing the lists of code points that map to each such regular
76# expression property value, one file per list
77#
78# There is also a single form shortcut that Perl adds for many of the commonly
79# used properties. This happens for all binary properties, plus script,
80# general_category, and block properties.
81#
82# Thus the outputs of this program are files. There are map files, mostly in
83# the 'To' directory; and there are list files for use in regular expression
84# matching, all in subdirectories of the 'lib' directory, with each
85# subdirectory being named for the property that the lists in it are for.
86# Bookkeeping, test, and documentation files are also generated.
87
88my $matches_directory = 'lib'; # Where match (\p{}) files go.
89my $map_directory = 'To'; # Where map files go.
90
91# DATA STRUCTURES
92#
93# The major data structures of this program are Property, of course, but also
94# Table. There are two kinds of tables, very similar to each other.
95# "Match_Table" is the data structure giving the list of code points that have
96# a particular property value, mentioned above. There is also a "Map_Table"
97# data structure which gives the property's mapping from code point to value.
98# There are two structures because the match tables need to be combined in
99# various ways, such as constructing unions, intersections, complements, etc.,
100# and the map ones don't. And there would be problems, perhaps subtle, if
101# a map table were inadvertently operated on in some of those ways.
102# The use of separate classes with operations defined on one but not the other
103# prevents accidentally confusing the two.
104#
105# At the heart of each table's data structure is a "Range_List", which is just
106# an ordered list of "Ranges", plus ancillary information, and methods to
107# operate on them. A Range is a compact way to store property information.
108# Each range has a starting code point, an ending code point, and a value that
109# is meant to apply to all the code points between the two end points,
110# inclusive. For a map table, this value is the property value for those
111# code points. Two such ranges could be written like this:
112# 0x41 .. 0x5A, 'Upper',
113# 0x61 .. 0x7A, 'Lower'
114#
115# Each range also has a type used as a convenience to classify the values.
116# Most ranges in this program will be Type 0, or normal, but there are some
117# ranges that have a non-zero type. These are used only in map tables, and
118# are for mappings that don't fit into the normal scheme of things. Mappings
119# that require a hash entry to communicate with utf8.c are one example;
120# another example is mappings for charnames.pm to use which indicate a name
121# that is algorithmically determinable from its code point (and vice-versa).
122# These are used to significantly compact these tables, instead of listing
123# each one of the tens of thousands individually.
124#
125# In a match table, the value of a range is irrelevant (and hence the type as
126# well, which will always be 0), and arbitrarily set to the null string.
127# Using the example above, there would be two match tables for those two
128# entries, one named Upper would contain the 0x41..0x5A range, and the other
129# named Lower would contain 0x61..0x7A.
130#
131# Actually, there are two types of range lists, "Range_Map" is the one
132# associated with map tables, and "Range_List" with match tables.
133# Again, this is so that methods can be defined on one and not the other so as
134# to prevent operating on them in incorrect ways.
135#
136# Eventually, most tables are written out to files to be read by utf8_heavy.pl
137# in the perl core. All tables could in theory be written, but some are
138# suppressed because there is no current practical use for them. It is easy
139# to change which get written by changing various lists that are near the top
140# of the actual code in this file. The table data structures contain enough
141# ancillary information to allow them to be treated as separate entities for
142# writing, such as the path to each one's file. There is a heading in each
143# map table that gives the format of its entries, and what the map is for all
144# the code points missing from it. (This allows tables to be more compact.)
678f13d5 145#
99870f4d
KW
146# The Property data structure contains one or more tables. All properties
147# contain a map table (except the $perl property which is a
148# pseudo-property containing only match tables), and any properties that
149# are usable in regular expression matches also contain various matching
150# tables, one for each value the property can have. A binary property can
151# have two values, True and False (or Y and N, which are preferred by Unicode
152# terminology). Thus each of these properties will have a map table that
153# takes every code point and maps it to Y or N (but having ranges cuts the
154# number of entries in that table way down), and two match tables, one
155# which has a list of all the code points that map to Y, and one for all the
156# code points that map to N. (For each of these, a third table is also
157# generated for the pseudo Perl property. It contains the identical code
158# points as the Y table, but can be written, not in the compound form, but in
159# a "single" form like \p{IsUppercase}.) Many properties are binary, but some
160# properties have several possible values, some have many, and properties like
161# Name have a different value for every named code point. Those will not,
162# unless the controlling lists are changed, have their match tables written
163# out. But all the ones which can be used in regular expression \p{} and \P{}
c12f2655
KW
164# constructs will. Prior to 5.14, generally a property would have either its
165# map table or its match tables written but not both. Again, what gets
166# written is controlled by lists which can easily be changed. Starting in
167# 5.14, advantage was taken of this, and all the map tables needed to
168# reconstruct the Unicode db are now written out, while suppressing the
169# Unicode .txt files that contain the data. Our tables are much more compact
170# than the .txt files, so a significant space savings was achieved.
171
172# Properties have a 'Type', like binary, or string, or enum depending on how
173# many match tables there are and the content of the maps. This 'Type' is
174# different than a range 'Type', so don't get confused by the two concepts
175# having the same name.
678f13d5 176#
99870f4d
KW
177# For information about the Unicode properties, see Unicode's UAX44 document:
178
179my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
180
181# As stated earlier, this program will work on any release of Unicode so far.
182# Most obvious problems in earlier data have NOT been corrected except when
183# necessary to make Perl or this program work reasonably. For example, no
c12f2655
KW
184# folding information was given in early releases, so this program substitutes
185# lower case instead, just so that a regular expression with the /i option
186# will do something that actually gives the right results in many cases.
187# There are also a couple other corrections for version 1.1.5, commented at
188# the point they are made. As an example of corrections that weren't made
189# (but could be) is this statement from DerivedAge.txt: "The supplementary
190# private use code points and the non-character code points were assigned in
191# version 2.0, but not specifically listed in the UCD until versions 3.0 and
192# 3.1 respectively." (To be precise it was 3.0.1 not 3.0.0) More information
193# on Unicode version glitches is further down in these introductory comments.
99870f4d 194#
5f7264c7
KW
195# This program works on all non-provisional properties as of 6.0, though the
196# files for some are suppressed from apparent lack of demand for them. You
197# can change which are output by changing lists in this program.
678f13d5 198#
dc85bd38 199# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
200# loose matchings rules (from Unicode TR18):
201#
202# The recommended names for UCD properties and property values are in
203# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
204# [PropValue]. There are both abbreviated names and longer, more
205# descriptive names. It is strongly recommended that both names be
206# recognized, and that loose matching of property names be used,
207# whereby the case distinctions, whitespace, hyphens, and underbar
208# are ignored.
209# The program still allows Fuzzy to override its determination of if loose
210# matching should be used, but it isn't currently used, as it is no longer
211# needed; the calculations it makes are good enough.
678f13d5 212#
99870f4d
KW
213# SUMMARY OF HOW IT WORKS:
214#
215# Process arguments
216#
217# A list is constructed containing each input file that is to be processed
218#
219# Each file on the list is processed in a loop, using the associated handler
220# code for each:
221# The PropertyAliases.txt and PropValueAliases.txt files are processed
222# first. These files name the properties and property values.
223# Objects are created of all the property and property value names
224# that the rest of the input should expect, including all synonyms.
225# The other input files give mappings from properties to property
226# values. That is, they list code points and say what the mapping
227# is under the given property. Some files give the mappings for
228# just one property; and some for many. This program goes through
229# each file and populates the properties from them. Some properties
230# are listed in more than one file, and Unicode has set up a
231# precedence as to which has priority if there is a conflict. Thus
232# the order of processing matters, and this program handles the
233# conflict possibility by processing the overriding input files
234# last, so that if necessary they replace earlier values.
235# After this is all done, the program creates the property mappings not
236# furnished by Unicode, but derivable from what it does give.
237# The tables of code points that match each property value in each
238# property that is accessible by regular expressions are created.
239# The Perl-defined properties are created and populated. Many of these
240# require data determined from the earlier steps
241# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 242# and Unicode are reconciled and warned about.
99870f4d
KW
243# All the properties are written to files
244# Any other files are written, and final warnings issued.
678f13d5 245#
99870f4d
KW
246# For clarity, a number of operators have been overloaded to work on tables:
247# ~ means invert (take all characters not in the set). The more
248# conventional '!' is not used because of the possibility of confusing
249# it with the actual boolean operation.
250# + means union
251# - means subtraction
252# & means intersection
253# The precedence of these is the order listed. Parentheses should be
254# copiously used. These are not a general scheme. The operations aren't
255# defined for a number of things, deliberately, to avoid getting into trouble.
256# Operations are done on references and affect the underlying structures, so
257# that the copy constructors for them have been overloaded to not return a new
258# clone, but the input object itself.
678f13d5 259#
99870f4d
KW
260# The bool operator is deliberately not overloaded to avoid confusion with
261# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
262#
263# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
264#
265# This program needs to be able to run under miniperl. Therefore, it uses a
266# minimum of other modules, and hence implements some things itself that could
267# be gotten from CPAN
268#
269# This program uses inputs published by the Unicode Consortium. These can
270# change incompatibly between releases without the Perl maintainers realizing
271# it. Therefore this program is now designed to try to flag these. It looks
272# at the directories where the inputs are, and flags any unrecognized files.
273# It keeps track of all the properties in the files it handles, and flags any
274# that it doesn't know how to handle. It also flags any input lines that
275# don't match the expected syntax, among other checks.
276#
277# It is also designed so if a new input file matches one of the known
278# templates, one hopefully just needs to add it to a list to have it
279# processed.
280#
281# As mentioned earlier, some properties are given in more than one file. In
282# particular, the files in the extracted directory are supposedly just
283# reformattings of the others. But they contain information not easily
284# derivable from the other files, including results for Unihan, which this
285# program doesn't ordinarily look at, and for unassigned code points. They
286# also have historically had errors or been incomplete. In an attempt to
287# create the best possible data, this program thus processes them first to
288# glean information missing from the other files; then processes those other
289# files to override any errors in the extracted ones. Much of the design was
290# driven by this need to store things and then possibly override them.
291#
292# It tries to keep fatal errors to a minimum, to generate something usable for
293# testing purposes. It always looks for files that could be inputs, and will
294# warn about any that it doesn't know how to handle (the -q option suppresses
295# the warning).
99870f4d 296#
678f13d5
KW
297# Why is there more than one type of range?
298# This simplified things. There are some very specialized code points that
299# have to be handled specially for output, such as Hangul syllable names.
300# By creating a range type (done late in the development process), it
301# allowed this to be stored with the range, and overridden by other input.
302# Originally these were stored in another data structure, and it became a
303# mess trying to decide if a second file that was for the same property was
304# overriding the earlier one or not.
305#
306# Why are there two kinds of tables, match and map?
307# (And there is a base class shared by the two as well.) As stated above,
308# they actually are for different things. Development proceeded much more
309# smoothly when I (khw) realized the distinction. Map tables are used to
310# give the property value for every code point (actually every code point
311# that doesn't map to a default value). Match tables are used for regular
312# expression matches, and are essentially the inverse mapping. Separating
313# the two allows more specialized methods, and error checks so that one
314# can't just take the intersection of two map tables, for example, as that
315# is nonsensical.
99870f4d 316#
23e33b60
KW
317# DEBUGGING
318#
678f13d5
KW
319# This program is written so it will run under miniperl. Occasionally changes
320# will cause an error where the backtrace doesn't work well under miniperl.
321# To diagnose the problem, you can instead run it under regular perl, if you
322# have one compiled.
323#
324# There is a good trace facility. To enable it, first sub DEBUG must be set
325# to return true. Then a line like
326#
327# local $to_trace = 1 if main::DEBUG;
328#
329# can be added to enable tracing in its lexical scope or until you insert
330# another line:
331#
332# local $to_trace = 0 if main::DEBUG;
333#
334# then use a line like "trace $a, @b, %c, ...;
335#
336# Some of the more complex subroutines already have trace statements in them.
337# Permanent trace statements should be like:
338#
339# trace ... if main::DEBUG && $to_trace;
340#
341# If there is just one or a few files that you're debugging, you can easily
342# cause most everything else to be skipped. Change the line
343#
344# my $debug_skip = 0;
345#
346# to 1, and every file whose object is in @input_file_objects and doesn't have
347# a, 'non_skip => 1,' in its constructor will be skipped.
348#
b4a0206c 349# To compare the output tables, it may be useful to specify the -annotate
c4019d52
KW
350# flag. This causes the tables to expand so there is one entry for each
351# non-algorithmically named code point giving, currently its name, and its
352# graphic representation if printable (and you have a font that knows about
353# it). This makes it easier to see what the particular code points are in
354# each output table. The tables are usable, but because they don't have
355# ranges (for the most part), a Perl using them will run slower. Non-named
356# code points are annotated with a description of their status, and contiguous
357# ones with the same description will be output as a range rather than
358# individually. Algorithmically named characters are also output as ranges,
359# except when there are just a few contiguous ones.
360#
99870f4d
KW
361# FUTURE ISSUES
362#
363# The program would break if Unicode were to change its names so that
364# interior white space, underscores, or dashes differences were significant
365# within property and property value names.
366#
367# It might be easier to use the xml versions of the UCD if this program ever
368# would need heavy revision, and the ability to handle old versions was not
369# required.
370#
371# There is the potential for name collisions, in that Perl has chosen names
372# that Unicode could decide it also likes. There have been such collisions in
373# the past, with mostly Perl deciding to adopt the Unicode definition of the
374# name. However in the 5.2 Unicode beta testing, there were a number of such
375# collisions, which were withdrawn before the final release, because of Perl's
376# and other's protests. These all involved new properties which began with
377# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
378# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
379# Unicode document, so they are unlikely to be used by Unicode for another
380# purpose. However, they might try something beginning with 'In', or use any
381# of the other Perl-defined properties. This program will warn you of name
382# collisions, and refuse to generate tables with them, but manual intervention
383# will be required in this event. One scheme that could be implemented, if
384# necessary, would be to have this program generate another file, or add a
385# field to mktables.lst that gives the date of first definition of a property.
386# Each new release of Unicode would use that file as a basis for the next
387# iteration. And the Perl synonym addition code could sort based on the age
388# of the property, so older properties get priority, and newer ones that clash
389# would be refused; hence existing code would not be impacted, and some other
390# synonym would have to be used for the new property. This is ugly, and
391# manual intervention would certainly be easier to do in the short run; lets
392# hope it never comes to this.
678f13d5 393#
99870f4d
KW
394# A NOTE ON UNIHAN
395#
396# This program can generate tables from the Unihan database. But it doesn't
397# by default, letting the CPAN module Unicode::Unihan handle them. Prior to
398# version 5.2, this database was in a single file, Unihan.txt. In 5.2 the
399# database was split into 8 different files, all beginning with the letters
400# 'Unihan'. This program will read those file(s) if present, but it needs to
401# know which of the many properties in the file(s) should have tables created
402# for them. It will create tables for any properties listed in
403# PropertyAliases.txt and PropValueAliases.txt, plus any listed in the
404# @cjk_properties array and the @cjk_property_values array. Thus, if a
405# property you want is not in those files of the release you are building
406# against, you must add it to those two arrays. Starting in 4.0, the
407# Unicode_Radical_Stroke was listed in those files, so if the Unihan database
408# is present in the directory, a table will be generated for that property.
409# In 5.2, several more properties were added. For your convenience, the two
5f7264c7 410# arrays are initialized with all the 6.0 listed properties that are also in
99870f4d
KW
411# earlier releases. But these are commented out. You can just uncomment the
412# ones you want, or use them as a template for adding entries for other
413# properties.
414#
415# You may need to adjust the entries to suit your purposes. setup_unihan(),
416# and filter_unihan_line() are the functions where this is done. This program
417# already does some adjusting to make the lines look more like the rest of the
418# Unicode DB; You can see what that is in filter_unihan_line()
419#
420# There is a bug in the 3.2 data file in which some values for the
421# kPrimaryNumeric property have commas and an unexpected comment. A filter
422# could be added for these; or for a particular installation, the Unihan.txt
423# file could be edited to fix them.
99870f4d 424#
678f13d5
KW
425# HOW TO ADD A FILE TO BE PROCESSED
426#
427# A new file from Unicode needs to have an object constructed for it in
428# @input_file_objects, probably at the end or at the end of the extracted
429# ones. The program should warn you if its name will clash with others on
430# restrictive file systems, like DOS. If so, figure out a better name, and
431# add lines to the README.perl file giving that. If the file is a character
432# property, it should be in the format that Unicode has by default
433# standardized for such files for the more recently introduced ones.
434# If so, the Input_file constructor for @input_file_objects can just be the
435# file name and release it first appeared in. If not, then it should be
436# possible to construct an each_line_handler() to massage the line into the
437# standardized form.
438#
439# For non-character properties, more code will be needed. You can look at
440# the existing entries for clues.
441#
442# UNICODE VERSIONS NOTES
443#
444# The Unicode UCD has had a number of errors in it over the versions. And
445# these remain, by policy, in the standard for that version. Therefore it is
446# risky to correct them, because code may be expecting the error. So this
447# program doesn't generally make changes, unless the error breaks the Perl
448# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
449# for U+1105, which causes real problems for the algorithms for Jamo
450# calculations, so it is changed here.
451#
452# But it isn't so clear cut as to what to do about concepts that are
453# introduced in a later release; should they extend back to earlier releases
454# where the concept just didn't exist? It was easier to do this than to not,
455# so that's what was done. For example, the default value for code points not
456# in the files for various properties was probably undefined until changed by
457# some version. No_Block for blocks is such an example. This program will
458# assign No_Block even in Unicode versions that didn't have it. This has the
459# benefit that code being written doesn't have to special case earlier
460# versions; and the detriment that it doesn't match the Standard precisely for
461# the affected versions.
462#
463# Here are some observations about some of the issues in early versions:
464#
6426c51b 465# The number of code points in \p{alpha} halved in 2.1.9. It turns out that
678f13d5
KW
466# the reason is that the CJK block starting at 4E00 was removed from PropList,
467# and was not put back in until 3.1.0
468#
469# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
470# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
471# reason is that 3.2 introduced U+205F=medium math space, which was not
472# classed as white space, but Perl figured out that it should have been. 4.0
473# reclassified it correctly.
474#
475# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
476# this was erroneously a synonym for 202. In 4.0, ATB became 202, and ATBL
477# was left with no code points, as all the ones that mapped to 202 stayed
478# mapped to 202. Thus if your program used the numeric name for the class,
479# it would not have been affected, but if it used the mnemonic, it would have
480# been.
481#
482# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that code
483# points which eventually came to have this script property value, instead
484# mapped to "Unknown". But in the next release all these code points were
485# moved to \p{sc=common} instead.
99870f4d
KW
486#
487# The default for missing code points for BidiClass is complicated. Starting
488# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
489# tries to do the best it can for earlier releases. It is done in
490# process_PropertyAliases()
491#
492##############################################################################
493
494my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
495 # and errors
496my $MAX_LINE_WIDTH = 78;
497
498# Debugging aid to skip most files so as to not be distracted by them when
499# concentrating on the ones being debugged. Add
500# non_skip => 1,
501# to the constructor for those files you want processed when you set this.
502# Files with a first version number of 0 are special: they are always
c12f2655
KW
503# processed regardless of the state of this flag. Generally, Jamo.txt and
504# UnicodeData.txt must not be skipped if you want this program to not die
505# before normal completion.
99870f4d
KW
506my $debug_skip = 0;
507
508# Set to 1 to enable tracing.
509our $to_trace = 0;
510
511{ # Closure for trace: debugging aid
512 my $print_caller = 1; # ? Include calling subroutine name
513 my $main_with_colon = 'main::';
514 my $main_colon_length = length($main_with_colon);
515
516 sub trace {
517 return unless $to_trace; # Do nothing if global flag not set
518
519 my @input = @_;
520
521 local $DB::trace = 0;
522 $DB::trace = 0; # Quiet 'used only once' message
523
524 my $line_number;
525
526 # Loop looking up the stack to get the first non-trace caller
527 my $caller_line;
528 my $caller_name;
529 my $i = 0;
530 do {
531 $line_number = $caller_line;
532 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
533 $caller = $main_with_colon unless defined $caller;
534
535 $caller_name = $caller;
536
537 # get rid of pkg
538 $caller_name =~ s/.*:://;
539 if (substr($caller_name, 0, $main_colon_length)
540 eq $main_with_colon)
541 {
542 $caller_name = substr($caller_name, $main_colon_length);
543 }
544
545 } until ($caller_name ne 'trace');
546
547 # If the stack was empty, we were called from the top level
548 $caller_name = 'main' if ($caller_name eq ""
549 || $caller_name eq 'trace');
550
551 my $output = "";
552 foreach my $string (@input) {
553 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
554 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
555 $output .= simple_dumper($string);
556 }
557 else {
558 $string = "$string" if ref $string;
559 $string = $UNDEF unless defined $string;
560 chomp $string;
561 $string = '""' if $string eq "";
562 $output .= " " if $output ne ""
563 && $string ne ""
564 && substr($output, -1, 1) ne " "
565 && substr($string, 0, 1) ne " ";
566 $output .= $string;
567 }
568 }
569
99f78760
KW
570 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
571 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
572 print STDERR $output, "\n";
573 return;
574 }
575}
576
577# This is for a rarely used development feature that allows you to compare two
578# versions of the Unicode standard without having to deal with changes caused
c12f2655
KW
579# by the code points introduced in the later version. Change the 0 to a
580# string containing a SINGLE dotted Unicode release number (e.g. "2.1"). Only
581# code points introduced in that release and earlier will be used; later ones
582# are thrown away. You use the version number of the earliest one you want to
583# compare; then run this program on directory structures containing each
584# release, and compare the outputs. These outputs will therefore include only
585# the code points common to both releases, and you can see the changes caused
586# just by the underlying release semantic changes. For versions earlier than
587# 3.2, you must copy a version of DAge.txt into the directory.
588my $string_compare_versions = DEBUG && 0; # e.g., "2.1";
99870f4d
KW
589my $compare_versions = DEBUG
590 && $string_compare_versions
591 && pack "C*", split /\./, $string_compare_versions;
592
593sub uniques {
594 # Returns non-duplicated input values. From "Perl Best Practices:
595 # Encapsulated Cleverness". p. 455 in first edition.
596
597 my %seen;
0e407844
NC
598 # Arguably this breaks encapsulation, if the goal is to permit multiple
599 # distinct objects to stringify to the same value, and be interchangeable.
600 # However, for this program, no two objects stringify identically, and all
601 # lists passed to this function are either objects or strings. So this
602 # doesn't affect correctness, but it does give a couple of percent speedup.
603 no overloading;
99870f4d
KW
604 return grep { ! $seen{$_}++ } @_;
605}
606
607$0 = File::Spec->canonpath($0);
608
609my $make_test_script = 0; # ? Should we output a test script
610my $write_unchanged_files = 0; # ? Should we update the output files even if
611 # we don't think they have changed
612my $use_directory = ""; # ? Should we chdir somewhere.
613my $pod_directory; # input directory to store the pod file.
614my $pod_file = 'perluniprops';
615my $t_path; # Path to the .t test file
616my $file_list = 'mktables.lst'; # File to store input and output file names.
617 # This is used to speed up the build, by not
618 # executing the main body of the program if
619 # nothing on the list has changed since the
620 # previous build
621my $make_list = 1; # ? Should we write $file_list. Set to always
622 # make a list so that when the pumpking is
623 # preparing a release, s/he won't have to do
624 # special things
625my $glob_list = 0; # ? Should we try to include unknown .txt files
626 # in the input.
bd9ebcfd
KW
627my $output_range_counts = $debugging_build; # ? Should we include the number
628 # of code points in ranges in
629 # the output
558712cf 630my $annotate = 0; # ? Should character names be in the output
9ef2b94f 631
99870f4d
KW
632# Verbosity levels; 0 is quiet
633my $NORMAL_VERBOSITY = 1;
634my $PROGRESS = 2;
635my $VERBOSE = 3;
636
637my $verbosity = $NORMAL_VERBOSITY;
638
639# Process arguments
640while (@ARGV) {
cf25bb62
JH
641 my $arg = shift @ARGV;
642 if ($arg eq '-v') {
99870f4d
KW
643 $verbosity = $VERBOSE;
644 }
645 elsif ($arg eq '-p') {
646 $verbosity = $PROGRESS;
647 $| = 1; # Flush buffers as we go.
648 }
649 elsif ($arg eq '-q') {
650 $verbosity = 0;
651 }
652 elsif ($arg eq '-w') {
653 $write_unchanged_files = 1; # update the files even if havent changed
654 }
655 elsif ($arg eq '-check') {
6ae7e459
YO
656 my $this = shift @ARGV;
657 my $ok = shift @ARGV;
658 if ($this ne $ok) {
659 print "Skipping as check params are not the same.\n";
660 exit(0);
661 }
00a8df5c 662 }
99870f4d
KW
663 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
664 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
665 }
3df51b85
KW
666 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
667 {
99870f4d 668 $make_test_script = 1;
99870f4d
KW
669 }
670 elsif ($arg eq '-makelist') {
671 $make_list = 1;
672 }
673 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
674 -d $use_directory or croak "Unknown directory '$use_directory'";
675 }
676 elsif ($arg eq '-L') {
677
678 # Existence not tested until have chdir'd
679 $file_list = shift;
680 }
681 elsif ($arg eq '-globlist') {
682 $glob_list = 1;
683 }
684 elsif ($arg eq '-c') {
685 $output_range_counts = ! $output_range_counts
686 }
b4a0206c 687 elsif ($arg eq '-annotate') {
558712cf 688 $annotate = 1;
bd9ebcfd
KW
689 $debugging_build = 1;
690 $output_range_counts = 1;
9ef2b94f 691 }
99870f4d
KW
692 else {
693 my $with_c = 'with';
694 $with_c .= 'out' if $output_range_counts; # Complements the state
695 croak <<END;
696usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
697 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
698 [-check A B ]
699 -c : Output comments $with_c number of code points in ranges
700 -q : Quiet Mode: Only output serious warnings.
701 -p : Set verbosity level to normal plus show progress.
702 -v : Set Verbosity level high: Show progress and non-serious
703 warnings
704 -w : Write files regardless
705 -C dir : Change to this directory before proceeding. All relative paths
706 except those specified by the -P and -T options will be done
707 with respect to this directory.
708 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 709 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
710 -L filelist : Use alternate 'filelist' instead of standard one
711 -globlist : Take as input all non-Test *.txt files in current and sub
712 directories
3df51b85
KW
713 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
714 overrides -T
99870f4d 715 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 716 -annotate : Output an annotation for each character in the table files;
c4019d52
KW
717 useful for debugging mktables, looking at diffs; but is slow,
718 memory intensive; resulting tables are usable but slow and
719 very large.
99870f4d
KW
720 -check A B : Executes $0 only if A and B are the same
721END
722 }
723}
724
725# Stores the most-recently changed file. If none have changed, can skip the
726# build
aeab6150 727my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
728
729# Change directories now, because need to read 'version' early.
730if ($use_directory) {
3df51b85 731 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
732 $pod_directory = File::Spec->rel2abs($pod_directory);
733 }
3df51b85 734 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 735 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 736 }
99870f4d 737 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 738 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 739 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 740 }
3df51b85 741 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 742 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 743 }
00a8df5c
YO
744}
745
99870f4d
KW
746# Get Unicode version into regular and v-string. This is done now because
747# various tables below get populated based on it. These tables are populated
748# here to be near the top of the file, and so easily seeable by those needing
749# to modify things.
750open my $VERSION, "<", "version"
751 or croak "$0: can't open required file 'version': $!\n";
752my $string_version = <$VERSION>;
753close $VERSION;
754chomp $string_version;
755my $v_version = pack "C*", split /\./, $string_version; # v string
756
757# The following are the complete names of properties with property values that
758# are known to not match any code points in some versions of Unicode, but that
759# may change in the future so they should be matchable, hence an empty file is
760# generated for them.
761my @tables_that_may_be_empty = (
762 'Joining_Type=Left_Joining',
763 );
764push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
765push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
766push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
767 if $v_version ge v4.1.0;
82aed44a
KW
768push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
769 if $v_version ge v6.0.0;
99870f4d
KW
770
771# The lists below are hashes, so the key is the item in the list, and the
772# value is the reason why it is in the list. This makes generation of
773# documentation easier.
774
775my %why_suppressed; # No file generated for these.
776
777# Files aren't generated for empty extraneous properties. This is arguable.
778# Extraneous properties generally come about because a property is no longer
779# used in a newer version of Unicode. If we generated a file without code
780# points, programs that used to work on that property will still execute
781# without errors. It just won't ever match (or will always match, with \P{}).
782# This means that the logic is now likely wrong. I (khw) think its better to
783# find this out by getting an error message. Just move them to the table
784# above to change this behavior
785my %why_suppress_if_empty_warn_if_not = (
786
787 # It is the only property that has ever officially been removed from the
788 # Standard. The database never contained any code points for it.
789 'Special_Case_Condition' => 'Obsolete',
790
791 # Apparently never official, but there were code points in some versions of
792 # old-style PropList.txt
793 'Non_Break' => 'Obsolete',
794);
795
796# These would normally go in the warn table just above, but they were changed
797# a long time before this program was written, so warnings about them are
798# moot.
799if ($v_version gt v3.2.0) {
800 push @tables_that_may_be_empty,
801 'Canonical_Combining_Class=Attached_Below_Left'
802}
803
5f7264c7 804# These are listed in the Property aliases file in 6.0, but Unihan is ignored
99870f4d
KW
805# unless explicitly added.
806if ($v_version ge v5.2.0) {
807 my $unihan = 'Unihan; remove from list if using Unihan';
ea25a9b2 808 foreach my $table (qw (
99870f4d
KW
809 kAccountingNumeric
810 kOtherNumeric
811 kPrimaryNumeric
812 kCompatibilityVariant
813 kIICore
814 kIRG_GSource
815 kIRG_HSource
816 kIRG_JSource
817 kIRG_KPSource
818 kIRG_MSource
819 kIRG_KSource
820 kIRG_TSource
821 kIRG_USource
822 kIRG_VSource
823 kRSUnicode
ea25a9b2 824 ))
99870f4d
KW
825 {
826 $why_suppress_if_empty_warn_if_not{$table} = $unihan;
827 }
ca12659b
NC
828}
829
272501f6
KW
830# Enum values for to_output_map() method in the Map_Table package.
831my $EXTERNAL_MAP = 1;
832my $INTERNAL_MAP = 2;
833
fcf1973c
KW
834# To override computed values for writing the map tables for these properties.
835# The default for enum map tables is to write them out, so that the Unicode
836# .txt files can be removed, but all the data to compute any property value
837# for any code point is available in a more compact form.
838my %global_to_output_map = (
839 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
c12f2655
KW
840 # get stuck supporting it if things change. Since it is a STRING
841 # property, it normally would be listed in the pod, but INTERNAL_MAP
842 # suppresses that.
fcf1973c
KW
843 Unicode_1_Name => $INTERNAL_MAP,
844
845 Present_In => 0, # Suppress, as easily computed from Age
fcf1973c 846 Block => 0, # Suppress, as Blocks.txt is retained.
53d34b6c
KW
847
848 # Suppress, as mapping can be found instead from the
849 # Perl_Decomposition_Mapping file
850 Decomposition_Type => 0,
fcf1973c
KW
851);
852
99870f4d
KW
853# Properties that this program ignores.
854my @unimplemented_properties = (
855'Unicode_Radical_Stroke' # Remove if changing to handle this one.
856);
d73e5302 857
99870f4d
KW
858# There are several types of obsolete properties defined by Unicode. These
859# must be hand-edited for every new Unicode release.
860my %why_deprecated; # Generates a deprecated warning message if used.
861my %why_stabilized; # Documentation only
862my %why_obsolete; # Documentation only
863
864{ # Closure
865 my $simple = 'Perl uses the more complete version of this property';
866 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
867
868 my $other_properties = 'other properties';
869 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 870 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
871
872 %why_deprecated = (
5f7264c7 873 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
874 'Jamo_Short_Name' => $contributory,
875 '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',
876 'Other_Alphabetic' => $contributory,
877 'Other_Default_Ignorable_Code_Point' => $contributory,
878 'Other_Grapheme_Extend' => $contributory,
879 'Other_ID_Continue' => $contributory,
880 'Other_ID_Start' => $contributory,
881 'Other_Lowercase' => $contributory,
882 'Other_Math' => $contributory,
883 'Other_Uppercase' => $contributory,
e22aaf5c
KW
884 'Expands_On_NFC' => $why_no_expand,
885 'Expands_On_NFD' => $why_no_expand,
886 'Expands_On_NFKC' => $why_no_expand,
887 'Expands_On_NFKD' => $why_no_expand,
99870f4d
KW
888 );
889
890 %why_suppressed = (
5f7264c7 891 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
892 # contains the same information, but without the algorithmically
893 # determinable Hangul syllables'. This file is not published, so it's
894 # existence is not noted in the comment.
895 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize',
896
897 'ISO_Comment' => 'Apparently no demand for it, but can access it through Unicode::UCD::charinfo. Obsoleted, and code points for it removed in Unicode 5.2',
99870f4d
KW
898
899 'Simple_Case_Folding' => "$simple. Can access this through Unicode::UCD::casefold",
900 'Simple_Lowercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
901 'Simple_Titlecase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
902 'Simple_Uppercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
903
904 'Name' => "Accessible via 'use charnames;'",
905 'Name_Alias' => "Accessible via 'use charnames;'",
906
5f7264c7 907 FC_NFKC_Closure => 'Supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
908 );
909
910 # The following are suppressed because they were made contributory or
911 # deprecated by Unicode before Perl ever thought about supporting them.
e22aaf5c
KW
912 foreach my $property ('Jamo_Short_Name',
913 'Grapheme_Link',
914 'Expands_On_NFC',
915 'Expands_On_NFD',
916 'Expands_On_NFKC',
917 'Expands_On_NFKD'
918 ) {
99870f4d
KW
919 $why_suppressed{$property} = $why_deprecated{$property};
920 }
cf25bb62 921
99870f4d
KW
922 # Customize the message for all the 'Other_' properties
923 foreach my $property (keys %why_deprecated) {
924 next if (my $main_property = $property) !~ s/^Other_//;
925 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
926 }
927}
928
929if ($v_version ge 4.0.0) {
930 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
931 if ($v_version ge 6.0.0) {
932 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
933 }
99870f4d 934}
5f7264c7 935if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 936 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7 937 if ($v_version ge 6.0.0) {
63f74647 938 $why_deprecated{'ISO_Comment'} = 'No longer needed for Unicode\'s internal chart generation; otherwise not useful, and code points for it have been removed';
5f7264c7 939 }
99870f4d
KW
940}
941
942# Probably obsolete forever
943if ($v_version ge v4.1.0) {
82aed44a
KW
944 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
945}
946if ($v_version ge v6.0.0) {
2b352efd
KW
947 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana (or both)"';
948 $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
949}
950
951# This program can create files for enumerated-like properties, such as
952# 'Numeric_Type'. This file would be the same format as for a string
953# property, with a mapping from code point to its value, so you could look up,
954# for example, the script a code point is in. But no one so far wants this
955# mapping, or they have found another way to get it since this is a new
956# feature. So no file is generated except if it is in this list.
957my @output_mapped_properties = split "\n", <<END;
958END
959
c12f2655
KW
960# If you are using the Unihan database in a Unicode version before 5.2, you
961# need to add the properties that you want to extract from it to this table.
962# For your convenience, the properties in the 6.0 PropertyAliases.txt file are
963# listed, commented out
99870f4d
KW
964my @cjk_properties = split "\n", <<'END';
965#cjkAccountingNumeric; kAccountingNumeric
966#cjkOtherNumeric; kOtherNumeric
967#cjkPrimaryNumeric; kPrimaryNumeric
968#cjkCompatibilityVariant; kCompatibilityVariant
969#cjkIICore ; kIICore
970#cjkIRG_GSource; kIRG_GSource
971#cjkIRG_HSource; kIRG_HSource
972#cjkIRG_JSource; kIRG_JSource
973#cjkIRG_KPSource; kIRG_KPSource
974#cjkIRG_KSource; kIRG_KSource
975#cjkIRG_TSource; kIRG_TSource
976#cjkIRG_USource; kIRG_USource
977#cjkIRG_VSource; kIRG_VSource
978#cjkRSUnicode; kRSUnicode ; Unicode_Radical_Stroke; URS
979END
980
981# Similarly for the property values. For your convenience, the lines in the
5f7264c7 982# 6.0 PropertyAliases.txt file are listed. Just remove the first BUT NOT both
c12f2655 983# '#' marks (for Unicode versions before 5.2)
99870f4d
KW
984my @cjk_property_values = split "\n", <<'END';
985## @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
986## @missing: 0000..10FFFF; cjkCompatibilityVariant; <code point>
987## @missing: 0000..10FFFF; cjkIICore; <none>
988## @missing: 0000..10FFFF; cjkIRG_GSource; <none>
989## @missing: 0000..10FFFF; cjkIRG_HSource; <none>
990## @missing: 0000..10FFFF; cjkIRG_JSource; <none>
991## @missing: 0000..10FFFF; cjkIRG_KPSource; <none>
992## @missing: 0000..10FFFF; cjkIRG_KSource; <none>
993## @missing: 0000..10FFFF; cjkIRG_TSource; <none>
994## @missing: 0000..10FFFF; cjkIRG_USource; <none>
995## @missing: 0000..10FFFF; cjkIRG_VSource; <none>
996## @missing: 0000..10FFFF; cjkOtherNumeric; NaN
997## @missing: 0000..10FFFF; cjkPrimaryNumeric; NaN
998## @missing: 0000..10FFFF; cjkRSUnicode; <none>
999END
1000
1001# The input files don't list every code point. Those not listed are to be
1002# defaulted to some value. Below are hard-coded what those values are for
1003# non-binary properties as of 5.1. Starting in 5.0, there are
1004# machine-parsable comment lines in the files the give the defaults; so this
1005# list shouldn't have to be extended. The claim is that all missing entries
1006# for binary properties will default to 'N'. Unicode tried to change that in
1007# 5.2, but the beta period produced enough protest that they backed off.
1008#
1009# The defaults for the fields that appear in UnicodeData.txt in this hash must
1010# be in the form that it expects. The others may be synonyms.
1011my $CODE_POINT = '<code point>';
1012my %default_mapping = (
1013 Age => "Unassigned",
1014 # Bidi_Class => Complicated; set in code
1015 Bidi_Mirroring_Glyph => "",
1016 Block => 'No_Block',
1017 Canonical_Combining_Class => 0,
1018 Case_Folding => $CODE_POINT,
1019 Decomposition_Mapping => $CODE_POINT,
1020 Decomposition_Type => 'None',
1021 East_Asian_Width => "Neutral",
1022 FC_NFKC_Closure => $CODE_POINT,
1023 General_Category => 'Cn',
1024 Grapheme_Cluster_Break => 'Other',
1025 Hangul_Syllable_Type => 'NA',
1026 ISO_Comment => "",
1027 Jamo_Short_Name => "",
1028 Joining_Group => "No_Joining_Group",
1029 # Joining_Type => Complicated; set in code
1030 kIICore => 'N', # Is converted to binary
1031 #Line_Break => Complicated; set in code
1032 Lowercase_Mapping => $CODE_POINT,
1033 Name => "",
1034 Name_Alias => "",
1035 NFC_QC => 'Yes',
1036 NFD_QC => 'Yes',
1037 NFKC_QC => 'Yes',
1038 NFKD_QC => 'Yes',
1039 Numeric_Type => 'None',
1040 Numeric_Value => 'NaN',
1041 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1042 Sentence_Break => 'Other',
1043 Simple_Case_Folding => $CODE_POINT,
1044 Simple_Lowercase_Mapping => $CODE_POINT,
1045 Simple_Titlecase_Mapping => $CODE_POINT,
1046 Simple_Uppercase_Mapping => $CODE_POINT,
1047 Titlecase_Mapping => $CODE_POINT,
1048 Unicode_1_Name => "",
1049 Unicode_Radical_Stroke => "",
1050 Uppercase_Mapping => $CODE_POINT,
1051 Word_Break => 'Other',
1052);
1053
1054# Below are files that Unicode furnishes, but this program ignores, and why
1055my %ignored_files = (
73ba1144
KW
1056 'CJKRadicals.txt' => 'Maps the kRSUnicode property values to corresponding code points',
1057 'Index.txt' => 'Alphabetical index of Unicode characters',
1058 'NamedSqProv.txt' => 'Named sequences proposed for inclusion in a later version of the Unicode Standard; if you need them now, you can append this file to F<NamedSequences.txt> and recompile perl',
1059 'NamesList.txt' => 'Annotated list of characters',
1060 'NormalizationCorrections.txt' => 'Documentation of corrections already incorporated into the Unicode data base',
1061 'Props.txt' => 'Only in very early releases; is a subset of F<PropList.txt> (which is used instead)',
1062 'ReadMe.txt' => 'Documentation',
1063 'StandardizedVariants.txt' => 'Certain glyph variations for character display are standardized. This lists the non-Unihan ones; the Unihan ones are also not used by Perl, and are in a separate Unicode data base L<http://www.unicode.org/ivd>',
1064 'EmojiSources.txt' => 'Maps certain Unicode code points to their legacy Japanese cell-phone values',
1065 'IndicMatraCategory.txt' => 'Provisional; for the analysis and processing of Indic scripts',
1066 'IndicSyllabicCategory.txt' => 'Provisional; for the analysis and processing of Indic scripts',
1067 'auxiliary/WordBreakTest.html' => 'Documentation of validation tests',
1068 'auxiliary/SentenceBreakTest.html' => 'Documentation of validation tests',
1069 'auxiliary/GraphemeBreakTest.html' => 'Documentation of validation tests',
1070 'auxiliary/LineBreakTest.html' => 'Documentation of validation tests',
99870f4d
KW
1071);
1072
678f13d5 1073### End of externally interesting definitions, except for @input_file_objects
99870f4d
KW
1074
1075my $HEADER=<<"EOF";
1076# !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3df51b85
KW
1077# This file is machine-generated by $0 from the Unicode
1078# database, Version $string_version. Any changes made here will be lost!
cf25bb62
JH
1079EOF
1080
b6922eda 1081my $INTERNAL_ONLY=<<"EOF";
99870f4d
KW
1082
1083# !!!!!!! INTERNAL PERL USE ONLY !!!!!!!
fac53429
KW
1084# This file is for internal use by core Perl only. The format and even the
1085# name or existence of this file are subject to change without notice. Don't
1086# use it directly.
99870f4d
KW
1087EOF
1088
1089my $DEVELOPMENT_ONLY=<<"EOF";
1090# !!!!!!! DEVELOPMENT USE ONLY !!!!!!!
1091# This file contains information artificially constrained to code points
1092# present in Unicode release $string_compare_versions.
1093# IT CANNOT BE RELIED ON. It is for use during development only and should
23e33b60 1094# not be used for production.
b6922eda
KW
1095
1096EOF
1097
6189eadc
KW
1098my $MAX_UNICODE_CODEPOINT_STRING = "10FFFF";
1099my $MAX_UNICODE_CODEPOINT = hex $MAX_UNICODE_CODEPOINT_STRING;
1100my $MAX_UNICODE_CODEPOINTS = $MAX_UNICODE_CODEPOINT + 1;
99870f4d
KW
1101
1102# Matches legal code point. 4-6 hex numbers, If there are 6, the first
1103# two must be 10; if there are 5, the first must not be a 0. Written this way
92199589
KW
1104# to decrease backtracking. The first regex allows the code point to be at
1105# the end of a word, but to work properly, the word shouldn't end with a valid
1106# hex character. The second one won't match a code point at the end of a
1107# word, and doesn't have the run-on issue
8c32d378
KW
1108my $run_on_code_point_re =
1109 qr/ (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
1110my $code_point_re = qr/\b$run_on_code_point_re/;
99870f4d
KW
1111
1112# This matches the beginning of the line in the Unicode db files that give the
1113# defaults for code points not listed (i.e., missing) in the file. The code
1114# depends on this ending with a semi-colon, so it can assume it is a valid
1115# field when the line is split() by semi-colons
1116my $missing_defaults_prefix =
6189eadc 1117 qr/^#\s+\@missing:\s+0000\.\.$MAX_UNICODE_CODEPOINT_STRING\s*;/;
99870f4d
KW
1118
1119# Property types. Unicode has more types, but these are sufficient for our
1120# purposes.
1121my $UNKNOWN = -1; # initialized to illegal value
1122my $NON_STRING = 1; # Either binary or enum
1123my $BINARY = 2;
06f26c45
KW
1124my $FORCED_BINARY = 3; # Not a binary property, but, besides its normal
1125 # tables, additional true and false tables are
1126 # generated so that false is anything matching the
1127 # default value, and true is everything else.
1128my $ENUM = 4; # Include catalog
1129my $STRING = 5; # Anything else: string or misc
99870f4d
KW
1130
1131# Some input files have lines that give default values for code points not
1132# contained in the file. Sometimes these should be ignored.
1133my $NO_DEFAULTS = 0; # Must evaluate to false
1134my $NOT_IGNORED = 1;
1135my $IGNORED = 2;
1136
1137# Range types. Each range has a type. Most ranges are type 0, for normal,
1138# and will appear in the main body of the tables in the output files, but
1139# there are other types of ranges as well, listed below, that are specially
1140# handled. There are pseudo-types as well that will never be stored as a
1141# type, but will affect the calculation of the type.
1142
1143# 0 is for normal, non-specials
1144my $MULTI_CP = 1; # Sequence of more than code point
1145my $HANGUL_SYLLABLE = 2;
1146my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1147my $NULL = 4; # The map is to the null string; utf8.c can't
1148 # handle these, nor is there an accepted syntax
1149 # for them in \p{} constructs
f86864ac 1150my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1151 # otherwise be $MULTI_CP type are instead type 0
1152
1153# process_generic_property_file() can accept certain overrides in its input.
1154# Each of these must begin AND end with $CMD_DELIM.
1155my $CMD_DELIM = "\a";
1156my $REPLACE_CMD = 'replace'; # Override the Replace
1157my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1158
1159my $NO = 0;
1160my $YES = 1;
1161
1162# Values for the Replace argument to add_range.
1163# $NO # Don't replace; add only the code points not
1164 # already present.
1165my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1166 # the comments at the subroutine definition.
1167my $UNCONDITIONALLY = 2; # Replace without conditions.
1168my $MULTIPLE = 4; # Don't replace, but add a duplicate record if
1169 # already there
56343c78 1170my $CROAK = 5; # Die with an error if is already there
99870f4d
KW
1171
1172# Flags to give property statuses. The phrases are to remind maintainers that
1173# if the flag is changed, the indefinite article referring to it in the
1174# documentation may need to be as well.
1175my $NORMAL = "";
1176my $SUPPRESSED = 'z'; # The character should never actually be seen, since
1177 # it is suppressed
c12f2655
KW
1178my $PLACEHOLDER = 'P'; # A property that is defined as a placeholder in a
1179 # Unicode version that doesn't have it, but we need it
1180 # to be defined, if empty, to have things work.
1181 # Implies no pod entry generated
99870f4d
KW
1182my $DEPRECATED = 'D';
1183my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1184my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1185my $DISCOURAGED = 'X';
1186my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1187my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1188my $STRICTER = 'T';
1189my $a_bold_stricter = "a 'B<$STRICTER>'";
1190my $A_bold_stricter = "A 'B<$STRICTER>'";
1191my $STABILIZED = 'S';
1192my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1193my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1194my $OBSOLETE = 'O';
1195my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1196my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1197
1198my %status_past_participles = (
1199 $DISCOURAGED => 'discouraged',
1200 $SUPPRESSED => 'should never be generated',
1201 $STABILIZED => 'stabilized',
1202 $OBSOLETE => 'obsolete',
37e2e78e 1203 $DEPRECATED => 'deprecated',
99870f4d
KW
1204);
1205
f5817e0a
KW
1206# The format of the values of the tables:
1207my $EMPTY_FORMAT = "";
99870f4d
KW
1208my $BINARY_FORMAT = 'b';
1209my $DECIMAL_FORMAT = 'd';
1210my $FLOAT_FORMAT = 'f';
1211my $INTEGER_FORMAT = 'i';
1212my $HEX_FORMAT = 'x';
1213my $RATIONAL_FORMAT = 'r';
1214my $STRING_FORMAT = 's';
a14f3cb1 1215my $DECOMP_STRING_FORMAT = 'c';
c3ff2976 1216my $STRING_WHITE_SPACE_LIST = 'sw';
99870f4d
KW
1217
1218my %map_table_formats = (
1219 $BINARY_FORMAT => 'binary',
1220 $DECIMAL_FORMAT => 'single decimal digit',
1221 $FLOAT_FORMAT => 'floating point number',
1222 $INTEGER_FORMAT => 'integer',
add63c13 1223 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1224 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1225 $STRING_FORMAT => 'string',
92f9d56c 1226 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
c3ff2976 1227 $STRING_WHITE_SPACE_LIST => 'string, but some elements are interpreted as a list; white space occurs only as list item separators'
99870f4d
KW
1228);
1229
1230# Unicode didn't put such derived files in a separate directory at first.
1231my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1232my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1233my $AUXILIARY = 'auxiliary';
1234
1235# Hashes that will eventually go into Heavy.pl for the use of utf8_heavy.pl
1236my %loose_to_file_of; # loosely maps table names to their respective
1237 # files
1238my %stricter_to_file_of; # same; but for stricter mapping.
1239my %nv_floating_to_rational; # maps numeric values floating point numbers to
1240 # their rational equivalent
c12f2655
KW
1241my %loose_property_name_of; # Loosely maps (non_string) property names to
1242 # standard form
99870f4d 1243
d867ccfb
KW
1244# Most properties are immune to caseless matching, otherwise you would get
1245# nonsensical results, as properties are a function of a code point, not
1246# everything that is caselessly equivalent to that code point. For example,
1247# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1248# be true because 's' and 'S' are equivalent caselessly. However,
1249# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1250# extend that concept to those very few properties that are like this. Each
1251# such property will match the full range caselessly. They are hard-coded in
1252# the program; it's not worth trying to make it general as it's extremely
1253# unlikely that they will ever change.
1254my %caseless_equivalent_to;
1255
99870f4d
KW
1256# These constants names and values were taken from the Unicode standard,
1257# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1258# syllables. The '_string' versions are so generated tables can retain the
1259# hex format, which is the more familiar value
1260my $SBase_string = "0xAC00";
1261my $SBase = CORE::hex $SBase_string;
1262my $LBase_string = "0x1100";
1263my $LBase = CORE::hex $LBase_string;
1264my $VBase_string = "0x1161";
1265my $VBase = CORE::hex $VBase_string;
1266my $TBase_string = "0x11A7";
1267my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1268my $SCount = 11172;
1269my $LCount = 19;
1270my $VCount = 21;
1271my $TCount = 28;
1272my $NCount = $VCount * $TCount;
1273
1274# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1275# with the above published constants.
1276my %Jamo;
1277my %Jamo_L; # Leading consonants
1278my %Jamo_V; # Vowels
1279my %Jamo_T; # Trailing consonants
1280
bb1dd3da
KW
1281# For code points whose name contains its ordinal as a '-ABCD' suffix.
1282# The key is the base name of the code point, and the value is an
1283# array giving all the ranges that use this base name. Each range
1284# is actually a hash giving the 'low' and 'high' values of it.
1285my %names_ending_in_code_point;
1286my %loose_names_ending_in_code_point; # Same as above, but has blanks, dashes
1287 # removed from the names
1288# Inverse mapping. The list of ranges that have these kinds of
1289# names. Each element contains the low, high, and base names in an
1290# anonymous hash.
1291my @code_points_ending_in_code_point;
1292
1293# Boolean: does this Unicode version have the hangul syllables, and are we
1294# writing out a table for them?
1295my $has_hangul_syllables = 0;
1296
1297# Does this Unicode version have code points whose names end in their
1298# respective code points, and are we writing out a table for them? 0 for no;
1299# otherwise points to first property that a table is needed for them, so that
1300# if multiple tables are needed, we don't create duplicates
1301my $needing_code_points_ending_in_code_point = 0;
1302
37e2e78e 1303my @backslash_X_tests; # List of tests read in for testing \X
99870f4d
KW
1304my @unhandled_properties; # Will contain a list of properties found in
1305 # the input that we didn't process.
f86864ac 1306my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1307 # listed in the pod
1308my @map_properties; # Properties that get map files written
1309my @named_sequences; # NamedSequences.txt contents.
1310my %potential_files; # Generated list of all .txt files in the directory
1311 # structure so we can warn if something is being
1312 # ignored.
1313my @files_actually_output; # List of files we generated.
1314my @more_Names; # Some code point names are compound; this is used
1315 # to store the extra components of them.
1316my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1317 # the minimum before we consider it equivalent to a
1318 # candidate rational
1319my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1320
1321# These store references to certain commonly used property objects
1322my $gc;
1323my $perl;
1324my $block;
3e20195b
KW
1325my $perl_charname;
1326my $print;
7fc6cb55 1327my $Any;
359523e2 1328my $script;
99870f4d
KW
1329
1330# Are there conflicting names because of beginning with 'In_', or 'Is_'
1331my $has_In_conflicts = 0;
1332my $has_Is_conflicts = 0;
1333
1334sub internal_file_to_platform ($) {
1335 # Convert our file paths which have '/' separators to those of the
1336 # platform.
1337
1338 my $file = shift;
1339 return undef unless defined $file;
1340
1341 return File::Spec->join(split '/', $file);
d07a55ed 1342}
5beb625e 1343
99870f4d
KW
1344sub file_exists ($) { # platform independent '-e'. This program internally
1345 # uses slash as a path separator.
1346 my $file = shift;
1347 return 0 if ! defined $file;
1348 return -e internal_file_to_platform($file);
1349}
5beb625e 1350
99870f4d 1351sub objaddr($) {
23e33b60
KW
1352 # Returns the address of the blessed input object.
1353 # It doesn't check for blessedness because that would do a string eval
1354 # every call, and the program is structured so that this is never called
1355 # for a non-blessed object.
99870f4d 1356
23e33b60 1357 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1358
1359 # Numifying a ref gives its address.
051df77b 1360 return pack 'J', $_[0];
99870f4d
KW
1361}
1362
558712cf 1363# These are used only if $annotate is true.
c4019d52
KW
1364# The entire range of Unicode characters is examined to populate these
1365# after all the input has been processed. But most can be skipped, as they
1366# have the same descriptive phrases, such as being unassigned
1367my @viacode; # Contains the 1 million character names
1368my @printable; # boolean: And are those characters printable?
1369my @annotate_char_type; # Contains a type of those characters, specifically
1370 # for the purposes of annotation.
1371my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1372 # name for the purposes of annotation. They map to the
c4019d52
KW
1373 # upper edge of the range, so that the end point can
1374 # be immediately found. This is used to skip ahead to
1375 # the end of a range, and avoid processing each
1376 # individual code point in it.
1377my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1378 # characters, but excluding those which are
1379 # also noncharacter code points
1380
1381# The annotation types are an extension of the regular range types, though
1382# some of the latter are folded into one. Make the new types negative to
1383# avoid conflicting with the regular types
1384my $SURROGATE_TYPE = -1;
1385my $UNASSIGNED_TYPE = -2;
1386my $PRIVATE_USE_TYPE = -3;
1387my $NONCHARACTER_TYPE = -4;
1388my $CONTROL_TYPE = -5;
1389my $UNKNOWN_TYPE = -6; # Used only if there is a bug in this program
1390
1391sub populate_char_info ($) {
558712cf 1392 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1393 # input code point's info that are needed for outputting more detailed
1394 # comments. If calling context wants a return, it is the end point of
1395 # any contiguous range of characters that share essentially the same info
1396
1397 my $i = shift;
1398 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1399
1400 $viacode[$i] = $perl_charname->value_of($i) || "";
1401
1402 # A character is generally printable if Unicode says it is,
1403 # but below we make sure that most Unicode general category 'C' types
1404 # aren't.
1405 $printable[$i] = $print->contains($i);
1406
1407 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1408
1409 # Only these two regular types are treated specially for annotations
1410 # purposes
1411 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1412 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1413
1414 # Give a generic name to all code points that don't have a real name.
1415 # We output ranges, if applicable, for these. Also calculate the end
1416 # point of the range.
1417 my $end;
1418 if (! $viacode[$i]) {
1419 if ($gc-> table('Surrogate')->contains($i)) {
1420 $viacode[$i] = 'Surrogate';
1421 $annotate_char_type[$i] = $SURROGATE_TYPE;
1422 $printable[$i] = 0;
1423 $end = $gc->table('Surrogate')->containing_range($i)->end;
1424 }
1425 elsif ($gc-> table('Private_use')->contains($i)) {
1426 $viacode[$i] = 'Private Use';
1427 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1428 $printable[$i] = 0;
1429 $end = $gc->table('Private_Use')->containing_range($i)->end;
1430 }
1431 elsif (Property::property_ref('Noncharacter_Code_Point')-> table('Y')->
1432 contains($i))
1433 {
1434 $viacode[$i] = 'Noncharacter';
1435 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1436 $printable[$i] = 0;
1437 $end = property_ref('Noncharacter_Code_Point')->table('Y')->
1438 containing_range($i)->end;
1439 }
1440 elsif ($gc-> table('Control')->contains($i)) {
1441 $viacode[$i] = 'Control';
1442 $annotate_char_type[$i] = $CONTROL_TYPE;
1443 $printable[$i] = 0;
1444 $end = 0x81 if $i == 0x80; # Hard-code this one known case
1445 }
1446 elsif ($gc-> table('Unassigned')->contains($i)) {
1447 $viacode[$i] = 'Unassigned, block=' . $block-> value_of($i);
1448 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1449 $printable[$i] = 0;
1450
1451 # Because we name the unassigned by the blocks they are in, it
1452 # can't go past the end of that block, and it also can't go past
1453 # the unassigned range it is in. The special table makes sure
1454 # that the non-characters, which are unassigned, are separated
1455 # out.
1456 $end = min($block->containing_range($i)->end,
1457 $unassigned_sans_noncharacters-> containing_range($i)->
1458 end);
13ca76ff
KW
1459 }
1460 else {
1461 Carp::my_carp_bug("Can't figure out how to annotate "
1462 . sprintf("U+%04X", $i)
1463 . ". Proceeding anyway.");
c4019d52
KW
1464 $viacode[$i] = 'UNKNOWN';
1465 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1466 $printable[$i] = 0;
1467 }
1468 }
1469
1470 # Here, has a name, but if it's one in which the code point number is
1471 # appended to the name, do that.
1472 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1473 $viacode[$i] .= sprintf("-%04X", $i);
1474 $end = $perl_charname->containing_range($i)->end;
1475 }
1476
1477 # And here, has a name, but if it's a hangul syllable one, replace it with
1478 # the correct name from the Unicode algorithm
1479 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1480 use integer;
1481 my $SIndex = $i - $SBase;
1482 my $L = $LBase + $SIndex / $NCount;
1483 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1484 my $T = $TBase + $SIndex % $TCount;
1485 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1486 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1487 $end = $perl_charname->containing_range($i)->end;
1488 }
1489
1490 return if ! defined wantarray;
1491 return $i if ! defined $end; # If not a range, return the input
1492
1493 # Save this whole range so can find the end point quickly
1494 $annotate_ranges->add_map($i, $end, $end);
1495
1496 return $end;
1497}
1498
23e33b60
KW
1499# Commented code below should work on Perl 5.8.
1500## This 'require' doesn't necessarily work in miniperl, and even if it does,
1501## the native perl version of it (which is what would operate under miniperl)
1502## is extremely slow, as it does a string eval every call.
1503#my $has_fast_scalar_util = $\18 !~ /miniperl/
1504# && defined eval "require Scalar::Util";
1505#
1506#sub objaddr($) {
1507# # Returns the address of the blessed input object. Uses the XS version if
1508# # available. It doesn't check for blessedness because that would do a
1509# # string eval every call, and the program is structured so that this is
1510# # never called for a non-blessed object.
1511#
1512# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1513#
1514# # Check at least that is a ref.
1515# my $pkg = ref($_[0]) or return undef;
1516#
1517# # Change to a fake package to defeat any overloaded stringify
1518# bless $_[0], 'main::Fake';
1519#
1520# # Numifying a ref gives its address.
051df77b 1521# my $addr = pack 'J', $_[0];
23e33b60
KW
1522#
1523# # Return to original class
1524# bless $_[0], $pkg;
1525# return $addr;
1526#}
1527
99870f4d
KW
1528sub max ($$) {
1529 my $a = shift;
1530 my $b = shift;
1531 return $a if $a >= $b;
1532 return $b;
1533}
1534
1535sub min ($$) {
1536 my $a = shift;
1537 my $b = shift;
1538 return $a if $a <= $b;
1539 return $b;
1540}
1541
1542sub clarify_number ($) {
1543 # This returns the input number with underscores inserted every 3 digits
1544 # in large (5 digits or more) numbers. Input must be entirely digits, not
1545 # checked.
1546
1547 my $number = shift;
1548 my $pos = length($number) - 3;
1549 return $number if $pos <= 1;
1550 while ($pos > 0) {
1551 substr($number, $pos, 0) = '_';
1552 $pos -= 3;
5beb625e 1553 }
99870f4d 1554 return $number;
99598c8c
JH
1555}
1556
12ac2576 1557
99870f4d 1558package Carp;
7ebf06b3 1559
99870f4d
KW
1560# These routines give a uniform treatment of messages in this program. They
1561# are placed in the Carp package to cause the stack trace to not include them,
1562# although an alternative would be to use another package and set @CARP_NOT
1563# for it.
12ac2576 1564
99870f4d 1565our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1566
99f78760
KW
1567# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1568# and overload trying to load Scalar:Util under miniperl. See
1569# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1570undef $overload::VERSION;
1571
99870f4d
KW
1572sub my_carp {
1573 my $message = shift || "";
1574 my $nofold = shift || 0;
7ebf06b3 1575
99870f4d
KW
1576 if ($message) {
1577 $message = main::join_lines($message);
1578 $message =~ s/^$0: *//; # Remove initial program name
1579 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1580 $message = "\n$0: $message;";
12ac2576 1581
99870f4d
KW
1582 # Fold the message with program name, semi-colon end punctuation
1583 # (which looks good with the message that carp appends to it), and a
1584 # hanging indent for continuation lines.
1585 $message = main::simple_fold($message, "", 4) unless $nofold;
1586 $message =~ s/\n$//; # Remove the trailing nl so what carp
1587 # appends is to the same line
1588 }
12ac2576 1589
99870f4d 1590 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1591
99870f4d
KW
1592 carp $message;
1593 return;
1594}
7ebf06b3 1595
99870f4d
KW
1596sub my_carp_bug {
1597 # This is called when it is clear that the problem is caused by a bug in
1598 # this program.
7ebf06b3 1599
99870f4d
KW
1600 my $message = shift;
1601 $message =~ s/^$0: *//;
1602 $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");
1603 carp $message;
1604 return;
1605}
7ebf06b3 1606
99870f4d
KW
1607sub carp_too_few_args {
1608 if (@_ != 2) {
1609 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1610 return;
12ac2576 1611 }
7ebf06b3 1612
99870f4d
KW
1613 my $args_ref = shift;
1614 my $count = shift;
7ebf06b3 1615
99870f4d
KW
1616 my_carp_bug("Need at least $count arguments to "
1617 . (caller 1)[3]
1618 . ". Instead got: '"
1619 . join ', ', @$args_ref
1620 . "'. No action taken.");
1621 return;
12ac2576
JP
1622}
1623
99870f4d
KW
1624sub carp_extra_args {
1625 my $args_ref = shift;
1626 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1627
99870f4d
KW
1628 unless (ref $args_ref) {
1629 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1630 return;
1631 }
1632 my ($package, $file, $line) = caller;
1633 my $subroutine = (caller 1)[3];
cf25bb62 1634
99870f4d
KW
1635 my $list;
1636 if (ref $args_ref eq 'HASH') {
1637 foreach my $key (keys %$args_ref) {
1638 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1639 }
99870f4d 1640 $list = join ', ', each %{$args_ref};
cf25bb62 1641 }
99870f4d
KW
1642 elsif (ref $args_ref eq 'ARRAY') {
1643 foreach my $arg (@$args_ref) {
1644 $arg = $UNDEF unless defined $arg;
1645 }
1646 $list = join ', ', @$args_ref;
1647 }
1648 else {
1649 my_carp_bug("Can't cope with ref "
1650 . ref($args_ref)
1651 . " . argument to 'carp_extra_args'. Not checking arguments.");
1652 return;
1653 }
1654
1655 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1656 return;
d73e5302
JH
1657}
1658
99870f4d
KW
1659package main;
1660
1661{ # Closure
1662
1663 # This program uses the inside-out method for objects, as recommended in
1664 # "Perl Best Practices". This closure aids in generating those. There
1665 # are two routines. setup_package() is called once per package to set
1666 # things up, and then set_access() is called for each hash representing a
1667 # field in the object. These routines arrange for the object to be
1668 # properly destroyed when no longer used, and for standard accessor
1669 # functions to be generated. If you need more complex accessors, just
1670 # write your own and leave those accesses out of the call to set_access().
1671 # More details below.
1672
1673 my %constructor_fields; # fields that are to be used in constructors; see
1674 # below
1675
1676 # The values of this hash will be the package names as keys to other
1677 # hashes containing the name of each field in the package as keys, and
1678 # references to their respective hashes as values.
1679 my %package_fields;
1680
1681 sub setup_package {
1682 # Sets up the package, creating standard DESTROY and dump methods
1683 # (unless already defined). The dump method is used in debugging by
1684 # simple_dumper().
1685 # The optional parameters are:
1686 # a) a reference to a hash, that gets populated by later
1687 # set_access() calls with one of the accesses being
1688 # 'constructor'. The caller can then refer to this, but it is
1689 # not otherwise used by these two routines.
1690 # b) a reference to a callback routine to call during destruction
1691 # of the object, before any fields are actually destroyed
1692
1693 my %args = @_;
1694 my $constructor_ref = delete $args{'Constructor_Fields'};
1695 my $destroy_callback = delete $args{'Destroy_Callback'};
1696 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1697
1698 my %fields;
1699 my $package = (caller)[0];
1700
1701 $package_fields{$package} = \%fields;
1702 $constructor_fields{$package} = $constructor_ref;
1703
1704 unless ($package->can('DESTROY')) {
1705 my $destroy_name = "${package}::DESTROY";
1706 no strict "refs";
1707
1708 # Use typeglob to give the anonymous subroutine the name we want
1709 *$destroy_name = sub {
1710 my $self = shift;
ffe43484 1711 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1712
1713 $self->$destroy_callback if $destroy_callback;
1714 foreach my $field (keys %{$package_fields{$package}}) {
1715 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1716 delete $package_fields{$package}{$field}{$addr};
1717 }
1718 return;
1719 }
1720 }
1721
1722 unless ($package->can('dump')) {
1723 my $dump_name = "${package}::dump";
1724 no strict "refs";
1725 *$dump_name = sub {
1726 my $self = shift;
1727 return dump_inside_out($self, $package_fields{$package}, @_);
1728 }
1729 }
1730 return;
1731 }
1732
1733 sub set_access {
1734 # Arrange for the input field to be garbage collected when no longer
1735 # needed. Also, creates standard accessor functions for the field
1736 # based on the optional parameters-- none if none of these parameters:
1737 # 'addable' creates an 'add_NAME()' accessor function.
1738 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1739 # function.
1740 # 'settable' creates a 'set_NAME()' accessor function.
1741 # 'constructor' doesn't create an accessor function, but adds the
1742 # field to the hash that was previously passed to
1743 # setup_package();
1744 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1745 # 'add' etc. all mean 'addable'.
1746 # The read accessor function will work on both array and scalar
1747 # values. If another accessor in the parameter list is 'a', the read
1748 # access assumes an array. You can also force it to be array access
1749 # by specifying 'readable_array' instead of 'readable'
1750 #
1751 # A sort-of 'protected' access can be set-up by preceding the addable,
1752 # readable or settable with some initial portion of 'protected_' (but,
1753 # the underscore is required), like 'p_a', 'pro_set', etc. The
1754 # "protection" is only by convention. All that happens is that the
1755 # accessor functions' names begin with an underscore. So instead of
1756 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1757 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1758 # end of each package, and then storing the __LINE__ ranges and
1759 # checking them on every accessor. But that is way overkill.)
1760
1761 # We create anonymous subroutines as the accessors and then use
1762 # typeglobs to assign them to the proper package and name
1763
1764 my $name = shift; # Name of the field
1765 my $field = shift; # Reference to the inside-out hash containing the
1766 # field
1767
1768 my $package = (caller)[0];
1769
1770 if (! exists $package_fields{$package}) {
1771 croak "$0: Must call 'setup_package' before 'set_access'";
1772 }
d73e5302 1773
99870f4d
KW
1774 # Stash the field so DESTROY can get it.
1775 $package_fields{$package}{$name} = $field;
cf25bb62 1776
99870f4d
KW
1777 # Remaining arguments are the accessors. For each...
1778 foreach my $access (@_) {
1779 my $access = lc $access;
cf25bb62 1780
99870f4d 1781 my $protected = "";
cf25bb62 1782
99870f4d
KW
1783 # Match the input as far as it goes.
1784 if ($access =~ /^(p[^_]*)_/) {
1785 $protected = $1;
1786 if (substr('protected_', 0, length $protected)
1787 eq $protected)
1788 {
1789
1790 # Add 1 for the underscore not included in $protected
1791 $access = substr($access, length($protected) + 1);
1792 $protected = '_';
1793 }
1794 else {
1795 $protected = "";
1796 }
1797 }
1798
1799 if (substr('addable', 0, length $access) eq $access) {
1800 my $subname = "${package}::${protected}add_$name";
1801 no strict "refs";
1802
1803 # add_ accessor. Don't add if already there, which we
1804 # determine using 'eq' for scalars and '==' otherwise.
1805 *$subname = sub {
1806 use strict "refs";
1807 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1808 my $self = shift;
1809 my $value = shift;
ffe43484 1810 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1811 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1812 if (ref $value) {
f998e60c 1813 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1814 }
1815 else {
f998e60c 1816 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1817 }
f998e60c 1818 push @{$field->{$addr}}, $value;
99870f4d
KW
1819 return;
1820 }
1821 }
1822 elsif (substr('constructor', 0, length $access) eq $access) {
1823 if ($protected) {
1824 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1825 }
1826 else {
1827 $constructor_fields{$package}{$name} = $field;
1828 }
1829 }
1830 elsif (substr('readable_array', 0, length $access) eq $access) {
1831
1832 # Here has read access. If one of the other parameters for
1833 # access is array, or this one specifies array (by being more
1834 # than just 'readable_'), then create a subroutine that
1835 # assumes the data is an array. Otherwise just a scalar
1836 my $subname = "${package}::${protected}$name";
1837 if (grep { /^a/i } @_
1838 or length($access) > length('readable_'))
1839 {
1840 no strict "refs";
1841 *$subname = sub {
1842 use strict "refs";
23e33b60 1843 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 1844 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
1845 if (ref $field->{$addr} ne 'ARRAY') {
1846 my $type = ref $field->{$addr};
1847 $type = 'scalar' unless $type;
1848 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
1849 return;
1850 }
1851 return scalar @{$field->{$addr}} unless wantarray;
1852
1853 # Make a copy; had problems with caller modifying the
1854 # original otherwise
1855 my @return = @{$field->{$addr}};
1856 return @return;
1857 }
1858 }
1859 else {
1860
1861 # Here not an array value, a simpler function.
1862 no strict "refs";
1863 *$subname = sub {
1864 use strict "refs";
23e33b60 1865 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 1866 no overloading;
051df77b 1867 return $field->{pack 'J', $_[0]};
99870f4d
KW
1868 }
1869 }
1870 }
1871 elsif (substr('settable', 0, length $access) eq $access) {
1872 my $subname = "${package}::${protected}set_$name";
1873 no strict "refs";
1874 *$subname = sub {
1875 use strict "refs";
23e33b60
KW
1876 if (main::DEBUG) {
1877 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
1878 Carp::carp_extra_args(\@_) if @_ > 2;
1879 }
1880 # $self is $_[0]; $value is $_[1]
f998e60c 1881 no overloading;
051df77b 1882 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
1883 return;
1884 }
1885 }
1886 else {
1887 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
1888 }
cf25bb62 1889 }
99870f4d 1890 return;
cf25bb62 1891 }
99870f4d
KW
1892}
1893
1894package Input_file;
1895
1896# All input files use this object, which stores various attributes about them,
1897# and provides for convenient, uniform handling. The run method wraps the
1898# processing. It handles all the bookkeeping of opening, reading, and closing
1899# the file, returning only significant input lines.
1900#
1901# Each object gets a handler which processes the body of the file, and is
1902# called by run(). Most should use the generic, default handler, which has
1903# code scrubbed to handle things you might not expect. A handler should
1904# basically be a while(next_line()) {...} loop.
1905#
1906# You can also set up handlers to
1907# 1) call before the first line is read for pre processing
1908# 2) call to adjust each line of the input before the main handler gets them
1909# 3) call upon EOF before the main handler exits its loop
1910# 4) call at the end for post processing
1911#
1912# $_ is used to store the input line, and is to be filtered by the
1913# each_line_handler()s. So, if the format of the line is not in the desired
1914# format for the main handler, these are used to do that adjusting. They can
1915# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
1916# so the $_ output of one is used as the input to the next. None of the other
1917# handlers are stackable, but could easily be changed to be so.
1918#
1919# Most of the handlers can call insert_lines() or insert_adjusted_lines()
1920# which insert the parameters as lines to be processed before the next input
1921# file line is read. This allows the EOF handler to flush buffers, for
1922# example. The difference between the two routines is that the lines inserted
1923# by insert_lines() are subjected to the each_line_handler()s. (So if you
1924# called it from such a handler, you would get infinite recursion.) Lines
1925# inserted by insert_adjusted_lines() go directly to the main handler without
1926# any adjustments. If the post-processing handler calls any of these, there
1927# will be no effect. Some error checking for these conditions could be added,
1928# but it hasn't been done.
1929#
1930# carp_bad_line() should be called to warn of bad input lines, which clears $_
1931# to prevent further processing of the line. This routine will output the
1932# message as a warning once, and then keep a count of the lines that have the
1933# same message, and output that count at the end of the file's processing.
1934# This keeps the number of messages down to a manageable amount.
1935#
1936# get_missings() should be called to retrieve any @missing input lines.
1937# Messages will be raised if this isn't done if the options aren't to ignore
1938# missings.
1939
1940sub trace { return main::trace(@_); }
1941
99870f4d
KW
1942{ # Closure
1943 # Keep track of fields that are to be put into the constructor.
1944 my %constructor_fields;
1945
1946 main::setup_package(Constructor_Fields => \%constructor_fields);
1947
1948 my %file; # Input file name, required
1949 main::set_access('file', \%file, qw{ c r });
1950
1951 my %first_released; # Unicode version file was first released in, required
1952 main::set_access('first_released', \%first_released, qw{ c r });
1953
1954 my %handler; # Subroutine to process the input file, defaults to
1955 # 'process_generic_property_file'
1956 main::set_access('handler', \%handler, qw{ c });
1957
1958 my %property;
1959 # name of property this file is for. defaults to none, meaning not
1960 # applicable, or is otherwise determinable, for example, from each line.
1961 main::set_access('property', \%property, qw{ c });
1962
1963 my %optional;
1964 # If this is true, the file is optional. If not present, no warning is
1965 # output. If it is present, the string given by this parameter is
1966 # evaluated, and if false the file is not processed.
1967 main::set_access('optional', \%optional, 'c', 'r');
1968
1969 my %non_skip;
1970 # This is used for debugging, to skip processing of all but a few input
1971 # files. Add 'non_skip => 1' to the constructor for those files you want
1972 # processed when you set the $debug_skip global.
1973 main::set_access('non_skip', \%non_skip, 'c');
1974
37e2e78e 1975 my %skip;
09ca89ce
KW
1976 # This is used to skip processing of this input file semi-permanently,
1977 # when it evaluates to true. The value should be the reason the file is
1978 # being skipped. It is used for files that we aren't planning to process
1979 # anytime soon, but want to allow to be in the directory and not raise a
1980 # message that we are not handling. Mostly for test files. This is in
1981 # contrast to the non_skip element, which is supposed to be used very
1982 # temporarily for debugging. Sets 'optional' to 1. Also, files that we
1983 # pretty much will never look at can be placed in the global
1984 # %ignored_files instead. Ones used here will be added to that list.
37e2e78e
KW
1985 main::set_access('skip', \%skip, 'c');
1986
99870f4d
KW
1987 my %each_line_handler;
1988 # list of subroutines to look at and filter each non-comment line in the
1989 # file. defaults to none. The subroutines are called in order, each is
1990 # to adjust $_ for the next one, and the final one adjusts it for
1991 # 'handler'
1992 main::set_access('each_line_handler', \%each_line_handler, 'c');
1993
1994 my %has_missings_defaults;
1995 # ? Are there lines in the file giving default values for code points
1996 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
1997 # the norm, but IGNORED means it has such lines, but the handler doesn't
1998 # use them. Having these three states allows us to catch changes to the
1999 # UCD that this program should track
2000 main::set_access('has_missings_defaults',
2001 \%has_missings_defaults, qw{ c r });
2002
2003 my %pre_handler;
2004 # Subroutine to call before doing anything else in the file. If undef, no
2005 # such handler is called.
2006 main::set_access('pre_handler', \%pre_handler, qw{ c });
2007
2008 my %eof_handler;
2009 # Subroutine to call upon getting an EOF on the input file, but before
2010 # that is returned to the main handler. This is to allow buffers to be
2011 # flushed. The handler is expected to call insert_lines() or
2012 # insert_adjusted() with the buffered material
2013 main::set_access('eof_handler', \%eof_handler, qw{ c r });
2014
2015 my %post_handler;
2016 # Subroutine to call after all the lines of the file are read in and
2017 # processed. If undef, no such handler is called.
2018 main::set_access('post_handler', \%post_handler, qw{ c });
2019
2020 my %progress_message;
2021 # Message to print to display progress in lieu of the standard one
2022 main::set_access('progress_message', \%progress_message, qw{ c });
2023
2024 my %handle;
2025 # cache open file handle, internal. Is undef if file hasn't been
2026 # processed at all, empty if has;
2027 main::set_access('handle', \%handle);
2028
2029 my %added_lines;
2030 # cache of lines added virtually to the file, internal
2031 main::set_access('added_lines', \%added_lines);
2032
2033 my %errors;
2034 # cache of errors found, internal
2035 main::set_access('errors', \%errors);
2036
2037 my %missings;
2038 # storage of '@missing' defaults lines
2039 main::set_access('missings', \%missings);
2040
2041 sub new {
2042 my $class = shift;
2043
2044 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2045 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2046
2047 # Set defaults
2048 $handler{$addr} = \&main::process_generic_property_file;
2049 $non_skip{$addr} = 0;
37e2e78e 2050 $skip{$addr} = 0;
99870f4d
KW
2051 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2052 $handle{$addr} = undef;
2053 $added_lines{$addr} = [ ];
2054 $each_line_handler{$addr} = [ ];
2055 $errors{$addr} = { };
2056 $missings{$addr} = [ ];
2057
2058 # Two positional parameters.
99f78760 2059 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2060 $file{$addr} = main::internal_file_to_platform(shift);
2061 $first_released{$addr} = shift;
2062
2063 # The rest of the arguments are key => value pairs
2064 # %constructor_fields has been set up earlier to list all possible
2065 # ones. Either set or push, depending on how the default has been set
2066 # up just above.
2067 my %args = @_;
2068 foreach my $key (keys %args) {
2069 my $argument = $args{$key};
2070
2071 # Note that the fields are the lower case of the constructor keys
2072 my $hash = $constructor_fields{lc $key};
2073 if (! defined $hash) {
2074 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2075 next;
2076 }
2077 if (ref $hash->{$addr} eq 'ARRAY') {
2078 if (ref $argument eq 'ARRAY') {
2079 foreach my $argument (@{$argument}) {
2080 next if ! defined $argument;
2081 push @{$hash->{$addr}}, $argument;
2082 }
2083 }
2084 else {
2085 push @{$hash->{$addr}}, $argument if defined $argument;
2086 }
2087 }
2088 else {
2089 $hash->{$addr} = $argument;
2090 }
2091 delete $args{$key};
2092 };
2093
2094 # If the file has a property for it, it means that the property is not
2095 # listed in the file's entries. So add a handler to the list of line
2096 # handlers to insert the property name into the lines, to provide a
2097 # uniform interface to the final processing subroutine.
2098 # the final code doesn't have to worry about that.
2099 if ($property{$addr}) {
2100 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
2101 }
2102
2103 if ($non_skip{$addr} && ! $debug_skip && $verbosity) {
2104 print "Warning: " . __PACKAGE__ . " constructor for $file{$addr} has useless 'non_skip' in it\n";
a3a8c5f0 2105 }
99870f4d 2106
09ca89ce
KW
2107 # If skipping, set to optional, and add to list of ignored files,
2108 # including its reason
2109 if ($skip{$addr}) {
2110 $optional{$addr} = 1;
2111 $ignored_files{$file{$addr}} = $skip{$addr}
2112 }
37e2e78e 2113
99870f4d 2114 return $self;
d73e5302
JH
2115 }
2116
cf25bb62 2117
99870f4d
KW
2118 use overload
2119 fallback => 0,
2120 qw("") => "_operator_stringify",
2121 "." => \&main::_operator_dot,
2122 ;
cf25bb62 2123
99870f4d
KW
2124 sub _operator_stringify {
2125 my $self = shift;
cf25bb62 2126
99870f4d 2127 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2128 }
d73e5302 2129
99870f4d
KW
2130 # flag to make sure extracted files are processed early
2131 my $seen_non_extracted_non_age = 0;
d73e5302 2132
99870f4d
KW
2133 sub run {
2134 # Process the input object $self. This opens and closes the file and
2135 # calls all the handlers for it. Currently, this can only be called
2136 # once per file, as it destroy's the EOF handler
d73e5302 2137
99870f4d
KW
2138 my $self = shift;
2139 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2140
ffe43484 2141 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2142
99870f4d 2143 my $file = $file{$addr};
d73e5302 2144
99870f4d
KW
2145 # Don't process if not expecting this file (because released later
2146 # than this Unicode version), and isn't there. This means if someone
2147 # copies it into an earlier version's directory, we will go ahead and
2148 # process it.
2149 return if $first_released{$addr} gt $v_version && ! -e $file;
2150
2151 # If in debugging mode and this file doesn't have the non-skip
2152 # flag set, and isn't one of the critical files, skip it.
2153 if ($debug_skip
2154 && $first_released{$addr} ne v0
2155 && ! $non_skip{$addr})
2156 {
2157 print "Skipping $file in debugging\n" if $verbosity;
2158 return;
2159 }
2160
2161 # File could be optional
37e2e78e 2162 if ($optional{$addr}) {
99870f4d
KW
2163 return unless -e $file;
2164 my $result = eval $optional{$addr};
2165 if (! defined $result) {
2166 Carp::my_carp_bug("Got '$@' when tried to eval $optional{$addr}. $file Skipped.");
2167 return;
2168 }
2169 if (! $result) {
2170 if ($verbosity) {
2171 print STDERR "Skipping processing input file '$file' because '$optional{$addr}' is not true\n";
2172 }
2173 return;
2174 }
2175 }
2176
2177 if (! defined $file || ! -e $file) {
2178
2179 # If the file doesn't exist, see if have internal data for it
2180 # (based on first_released being 0).
2181 if ($first_released{$addr} eq v0) {
2182 $handle{$addr} = 'pretend_is_open';
2183 }
2184 else {
2185 if (! $optional{$addr} # File could be optional
2186 && $v_version ge $first_released{$addr})
2187 {
2188 print STDERR "Skipping processing input file '$file' because not found\n" if $v_version ge $first_released{$addr};
2189 }
2190 return;
2191 }
2192 }
2193 else {
2194
37e2e78e
KW
2195 # Here, the file exists. Some platforms may change the case of
2196 # its name
99870f4d 2197 if ($seen_non_extracted_non_age) {
517956bf 2198 if ($file =~ /$EXTRACTED/i) {
99870f4d 2199 Carp::my_carp_bug(join_lines(<<END
99f78760 2200$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2201anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2202have subtle problems
2203END
2204 ));
2205 }
2206 }
2207 elsif ($EXTRACTED_DIR
2208 && $first_released{$addr} ne v0
517956bf
CB
2209 && $file !~ /$EXTRACTED/i
2210 && lc($file) ne 'dage.txt')
99870f4d
KW
2211 {
2212 # We don't set this (by the 'if' above) if we have no
2213 # extracted directory, so if running on an early version,
2214 # this test won't work. Not worth worrying about.
2215 $seen_non_extracted_non_age = 1;
2216 }
2217
2218 # And mark the file as having being processed, and warn if it
2219 # isn't a file we are expecting. As we process the files,
2220 # they are deleted from the hash, so any that remain at the
2221 # end of the program are files that we didn't process.
517956bf
CB
2222 my $fkey = File::Spec->rel2abs($file);
2223 my $expecting = delete $potential_files{$fkey};
2224 $expecting = delete $potential_files{lc($fkey)} unless defined $expecting;
678f13d5
KW
2225 Carp::my_carp("Was not expecting '$file'.") if
2226 ! $expecting
99870f4d
KW
2227 && ! defined $handle{$addr};
2228
37e2e78e
KW
2229 # Having deleted from expected files, we can quit if not to do
2230 # anything. Don't print progress unless really want verbosity
2231 if ($skip{$addr}) {
2232 print "Skipping $file.\n" if $verbosity >= $VERBOSE;
2233 return;
2234 }
2235
99870f4d
KW
2236 # Open the file, converting the slashes used in this program
2237 # into the proper form for the OS
2238 my $file_handle;
2239 if (not open $file_handle, "<", $file) {
2240 Carp::my_carp("Can't open $file. Skipping: $!");
2241 return 0;
2242 }
2243 $handle{$addr} = $file_handle; # Cache the open file handle
2244 }
2245
2246 if ($verbosity >= $PROGRESS) {
2247 if ($progress_message{$addr}) {
2248 print "$progress_message{$addr}\n";
2249 }
2250 else {
2251 # If using a virtual file, say so.
2252 print "Processing ", (-e $file)
2253 ? $file
2254 : "substitute $file",
2255 "\n";
2256 }
2257 }
2258
2259
2260 # Call any special handler for before the file.
2261 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2262
2263 # Then the main handler
2264 &{$handler{$addr}}($self);
2265
2266 # Then any special post-file handler.
2267 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2268
2269 # If any errors have been accumulated, output the counts (as the first
2270 # error message in each class was output when it was encountered).
2271 if ($errors{$addr}) {
2272 my $total = 0;
2273 my $types = 0;
2274 foreach my $error (keys %{$errors{$addr}}) {
2275 $total += $errors{$addr}->{$error};
2276 delete $errors{$addr}->{$error};
2277 $types++;
2278 }
2279 if ($total > 1) {
2280 my $message
2281 = "A total of $total lines had errors in $file. ";
2282
2283 $message .= ($types == 1)
2284 ? '(Only the first one was displayed.)'
2285 : '(Only the first of each type was displayed.)';
2286 Carp::my_carp($message);
2287 }
2288 }
2289
2290 if (@{$missings{$addr}}) {
2291 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2292 }
2293
2294 # If a real file handle, close it.
2295 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2296 ref $handle{$addr};
2297 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2298 # the file, as opposed to undef
2299 return;
2300 }
2301
2302 sub next_line {
2303 # Sets $_ to be the next logical input line, if any. Returns non-zero
2304 # if such a line exists. 'logical' means that any lines that have
2305 # been added via insert_lines() will be returned in $_ before the file
2306 # is read again.
2307
2308 my $self = shift;
2309 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2310
ffe43484 2311 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2312
2313 # Here the file is open (or if the handle is not a ref, is an open
2314 # 'virtual' file). Get the next line; any inserted lines get priority
2315 # over the file itself.
2316 my $adjusted;
2317
2318 LINE:
2319 while (1) { # Loop until find non-comment, non-empty line
2320 #local $to_trace = 1 if main::DEBUG;
2321 my $inserted_ref = shift @{$added_lines{$addr}};
2322 if (defined $inserted_ref) {
2323 ($adjusted, $_) = @{$inserted_ref};
2324 trace $adjusted, $_ if main::DEBUG && $to_trace;
2325 return 1 if $adjusted;
2326 }
2327 else {
2328 last if ! ref $handle{$addr}; # Don't read unless is real file
2329 last if ! defined ($_ = readline $handle{$addr});
2330 }
2331 chomp;
2332 trace $_ if main::DEBUG && $to_trace;
2333
2334 # See if this line is the comment line that defines what property
2335 # value that code points that are not listed in the file should
2336 # have. The format or existence of these lines is not guaranteed
2337 # by Unicode since they are comments, but the documentation says
2338 # that this was added for machine-readability, so probably won't
2339 # change. This works starting in Unicode Version 5.0. They look
2340 # like:
2341 #
2342 # @missing: 0000..10FFFF; Not_Reordered
2343 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2344 # @missing: 0000..10FFFF; ; NaN
2345 #
2346 # Save the line for a later get_missings() call.
2347 if (/$missing_defaults_prefix/) {
2348 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2349 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2350 }
2351 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2352 my @defaults = split /\s* ; \s*/x, $_;
2353
2354 # The first field is the @missing, which ends in a
2355 # semi-colon, so can safely shift.
2356 shift @defaults;
2357
2358 # Some of these lines may have empty field placeholders
2359 # which get in the way. An example is:
2360 # @missing: 0000..10FFFF; ; NaN
2361 # Remove them. Process starting from the top so the
2362 # splice doesn't affect things still to be looked at.
2363 for (my $i = @defaults - 1; $i >= 0; $i--) {
2364 next if $defaults[$i] ne "";
2365 splice @defaults, $i, 1;
2366 }
2367
2368 # What's left should be just the property (maybe) and the
2369 # default. Having only one element means it doesn't have
2370 # the property.
2371 my $default;
2372 my $property;
2373 if (@defaults >= 1) {
2374 if (@defaults == 1) {
2375 $default = $defaults[0];
2376 }
2377 else {
2378 $property = $defaults[0];
2379 $default = $defaults[1];
2380 }
2381 }
2382
2383 if (@defaults < 1
2384 || @defaults > 2
2385 || ($default =~ /^</
2386 && $default !~ /^<code *point>$/i
2387 && $default !~ /^<none>$/i))
2388 {
2389 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2390 }
2391 else {
2392
2393 # If the property is missing from the line, it should
2394 # be the one for the whole file
2395 $property = $property{$addr} if ! defined $property;
2396
2397 # Change <none> to the null string, which is what it
2398 # really means. If the default is the code point
2399 # itself, set it to <code point>, which is what
2400 # Unicode uses (but sometimes they've forgotten the
2401 # space)
2402 if ($default =~ /^<none>$/i) {
2403 $default = "";
2404 }
2405 elsif ($default =~ /^<code *point>$/i) {
2406 $default = $CODE_POINT;
2407 }
2408
2409 # Store them as a sub-arrays with both components.
2410 push @{$missings{$addr}}, [ $default, $property ];
2411 }
2412 }
2413
2414 # There is nothing for the caller to process on this comment
2415 # line.
2416 next;
2417 }
2418
2419 # Remove comments and trailing space, and skip this line if the
2420 # result is empty
2421 s/#.*//;
2422 s/\s+$//;
2423 next if /^$/;
2424
2425 # Call any handlers for this line, and skip further processing of
2426 # the line if the handler sets the line to null.
2427 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2428 &{$sub_ref}($self);
2429 next LINE if /^$/;
2430 }
2431
2432 # Here the line is ok. return success.
2433 return 1;
2434 } # End of looping through lines.
2435
2436 # If there is an EOF handler, call it (only once) and if it generates
2437 # more lines to process go back in the loop to handle them.
2438 if ($eof_handler{$addr}) {
2439 &{$eof_handler{$addr}}($self);
2440 $eof_handler{$addr} = ""; # Currently only get one shot at it.
2441 goto LINE if $added_lines{$addr};
2442 }
2443
2444 # Return failure -- no more lines.
2445 return 0;
2446
2447 }
2448
2449# Not currently used, not fully tested.
2450# sub peek {
2451# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
2452# # record. Not callable from an each_line_handler(), nor does it call
2453# # an each_line_handler() on the line.
2454#
2455# my $self = shift;
ffe43484 2456# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2457#
2458# foreach my $inserted_ref (@{$added_lines{$addr}}) {
2459# my ($adjusted, $line) = @{$inserted_ref};
2460# next if $adjusted;
2461#
2462# # Remove comments and trailing space, and return a non-empty
2463# # resulting line
2464# $line =~ s/#.*//;
2465# $line =~ s/\s+$//;
2466# return $line if $line ne "";
2467# }
2468#
2469# return if ! ref $handle{$addr}; # Don't read unless is real file
2470# while (1) { # Loop until find non-comment, non-empty line
2471# local $to_trace = 1 if main::DEBUG;
2472# trace $_ if main::DEBUG && $to_trace;
2473# return if ! defined (my $line = readline $handle{$addr});
2474# chomp $line;
2475# push @{$added_lines{$addr}}, [ 0, $line ];
2476#
2477# $line =~ s/#.*//;
2478# $line =~ s/\s+$//;
2479# return $line if $line ne "";
2480# }
2481#
2482# return;
2483# }
2484
2485
2486 sub insert_lines {
2487 # Lines can be inserted so that it looks like they were in the input
2488 # file at the place it was when this routine is called. See also
2489 # insert_adjusted_lines(). Lines inserted via this routine go through
2490 # any each_line_handler()
2491
2492 my $self = shift;
2493
2494 # Each inserted line is an array, with the first element being 0 to
2495 # indicate that this line hasn't been adjusted, and needs to be
2496 # processed.
f998e60c 2497 no overloading;
051df77b 2498 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
2499 return;
2500 }
2501
2502 sub insert_adjusted_lines {
2503 # Lines can be inserted so that it looks like they were in the input
2504 # file at the place it was when this routine is called. See also
2505 # insert_lines(). Lines inserted via this routine are already fully
2506 # adjusted, ready to be processed; each_line_handler()s handlers will
2507 # not be called. This means this is not a completely general
2508 # facility, as only the last each_line_handler on the stack should
2509 # call this. It could be made more general, by passing to each of the
2510 # line_handlers their position on the stack, which they would pass on
2511 # to this routine, and that would replace the boolean first element in
2512 # the anonymous array pushed here, so that the next_line routine could
2513 # use that to call only those handlers whose index is after it on the
2514 # stack. But this is overkill for what is needed now.
2515
2516 my $self = shift;
2517 trace $_[0] if main::DEBUG && $to_trace;
2518
2519 # Each inserted line is an array, with the first element being 1 to
2520 # indicate that this line has been adjusted
f998e60c 2521 no overloading;
051df77b 2522 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
2523 return;
2524 }
2525
2526 sub get_missings {
2527 # Returns the stored up @missings lines' values, and clears the list.
2528 # The values are in an array, consisting of the default in the first
2529 # element, and the property in the 2nd. However, since these lines
2530 # can be stacked up, the return is an array of all these arrays.
2531
2532 my $self = shift;
2533 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2534
ffe43484 2535 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2536
2537 # If not accepting a list return, just return the first one.
2538 return shift @{$missings{$addr}} unless wantarray;
2539
2540 my @return = @{$missings{$addr}};
2541 undef @{$missings{$addr}};
2542 return @return;
2543 }
2544
2545 sub _insert_property_into_line {
2546 # Add a property field to $_, if this file requires it.
2547
f998e60c 2548 my $self = shift;
ffe43484 2549 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 2550 my $property = $property{$addr};
99870f4d
KW
2551 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2552
2553 $_ =~ s/(;|$)/; $property$1/;
2554 return;
2555 }
2556
2557 sub carp_bad_line {
2558 # Output consistent error messages, using either a generic one, or the
2559 # one given by the optional parameter. To avoid gazillions of the
2560 # same message in case the syntax of a file is way off, this routine
2561 # only outputs the first instance of each message, incrementing a
2562 # count so the totals can be output at the end of the file.
2563
2564 my $self = shift;
2565 my $message = shift;
2566 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2567
ffe43484 2568 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2569
2570 $message = 'Unexpected line' unless $message;
2571
2572 # No trailing punctuation so as to fit with our addenda.
2573 $message =~ s/[.:;,]$//;
2574
2575 # If haven't seen this exact message before, output it now. Otherwise
2576 # increment the count of how many times it has occurred
2577 unless ($errors{$addr}->{$message}) {
2578 Carp::my_carp("$message in '$_' in "
f998e60c 2579 . $file{$addr}
99870f4d
KW
2580 . " at line $.. Skipping this line;");
2581 $errors{$addr}->{$message} = 1;
2582 }
2583 else {
2584 $errors{$addr}->{$message}++;
2585 }
2586
2587 # Clear the line to prevent any further (meaningful) processing of it.
2588 $_ = "";
2589
2590 return;
2591 }
2592} # End closure
2593
2594package Multi_Default;
2595
2596# Certain properties in early versions of Unicode had more than one possible
2597# default for code points missing from the files. In these cases, one
2598# default applies to everything left over after all the others are applied,
2599# and for each of the others, there is a description of which class of code
2600# points applies to it. This object helps implement this by storing the
2601# defaults, and for all but that final default, an eval string that generates
2602# the class that it applies to.
2603
2604
2605{ # Closure
2606
2607 main::setup_package();
2608
2609 my %class_defaults;
2610 # The defaults structure for the classes
2611 main::set_access('class_defaults', \%class_defaults);
2612
2613 my %other_default;
2614 # The default that applies to everything left over.
2615 main::set_access('other_default', \%other_default, 'r');
2616
2617
2618 sub new {
2619 # The constructor is called with default => eval pairs, terminated by
2620 # the left-over default. e.g.
2621 # Multi_Default->new(
2622 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
2623 # - 0x200D',
2624 # 'R' => 'some other expression that evaluates to code points',
2625 # .
2626 # .
2627 # .
2628 # 'U'));
2629
2630 my $class = shift;
2631
2632 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 2633 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2634
2635 while (@_ > 1) {
2636 my $default = shift;
2637 my $eval = shift;
2638 $class_defaults{$addr}->{$default} = $eval;
2639 }
2640
2641 $other_default{$addr} = shift;
2642
2643 return $self;
2644 }
2645
2646 sub get_next_defaults {
2647 # Iterates and returns the next class of defaults.
2648 my $self = shift;
2649 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2650
ffe43484 2651 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2652
2653 return each %{$class_defaults{$addr}};
2654 }
2655}
2656
2657package Alias;
2658
2659# An alias is one of the names that a table goes by. This class defines them
2660# including some attributes. Everything is currently setup in the
2661# constructor.
2662
2663
2664{ # Closure
2665
2666 main::setup_package();
2667
2668 my %name;
2669 main::set_access('name', \%name, 'r');
2670
2671 my %loose_match;
c12f2655 2672 # Should this name match loosely or not.
99870f4d
KW
2673 main::set_access('loose_match', \%loose_match, 'r');
2674
33e96e72
KW
2675 my %make_re_pod_entry;
2676 # Some aliases should not get their own entries in the re section of the
2677 # pod, because they are covered by a wild-card, and some we want to
2678 # discourage use of. Binary
2679 main::set_access('make_re_pod_entry', \%make_re_pod_entry, 'r');
99870f4d
KW
2680
2681 my %status;
2682 # Aliases have a status, like deprecated, or even suppressed (which means
2683 # they don't appear in documentation). Enum
2684 main::set_access('status', \%status, 'r');
2685
2686 my %externally_ok;
2687 # Similarly, some aliases should not be considered as usable ones for
2688 # external use, such as file names, or we don't want documentation to
2689 # recommend them. Boolean
2690 main::set_access('externally_ok', \%externally_ok, 'r');
2691
2692 sub new {
2693 my $class = shift;
2694
2695 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2696 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2697
2698 $name{$addr} = shift;
2699 $loose_match{$addr} = shift;
33e96e72 2700 $make_re_pod_entry{$addr} = shift;
99870f4d
KW
2701 $externally_ok{$addr} = shift;
2702 $status{$addr} = shift;
2703
2704 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2705
2706 # Null names are never ok externally
2707 $externally_ok{$addr} = 0 if $name{$addr} eq "";
2708
2709 return $self;
2710 }
2711}
2712
2713package Range;
2714
2715# A range is the basic unit for storing code points, and is described in the
2716# comments at the beginning of the program. Each range has a starting code
2717# point; an ending code point (not less than the starting one); a value
2718# that applies to every code point in between the two end-points, inclusive;
2719# and an enum type that applies to the value. The type is for the user's
2720# convenience, and has no meaning here, except that a non-zero type is
2721# considered to not obey the normal Unicode rules for having standard forms.
2722#
2723# The same structure is used for both map and match tables, even though in the
2724# latter, the value (and hence type) is irrelevant and could be used as a
2725# comment. In map tables, the value is what all the code points in the range
2726# map to. Type 0 values have the standardized version of the value stored as
2727# well, so as to not have to recalculate it a lot.
2728
2729sub trace { return main::trace(@_); }
2730
2731{ # Closure
2732
2733 main::setup_package();
2734
2735 my %start;
2736 main::set_access('start', \%start, 'r', 's');
2737
2738 my %end;
2739 main::set_access('end', \%end, 'r', 's');
2740
2741 my %value;
2742 main::set_access('value', \%value, 'r');
2743
2744 my %type;
2745 main::set_access('type', \%type, 'r');
2746
2747 my %standard_form;
2748 # The value in internal standard form. Defined only if the type is 0.
2749 main::set_access('standard_form', \%standard_form);
2750
2751 # Note that if these fields change, the dump() method should as well
2752
2753 sub new {
2754 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
2755 my $class = shift;
2756
2757 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2758 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2759
2760 $start{$addr} = shift;
2761 $end{$addr} = shift;
2762
2763 my %args = @_;
2764
2765 my $value = delete $args{'Value'}; # Can be 0
2766 $value = "" unless defined $value;
2767 $value{$addr} = $value;
2768
2769 $type{$addr} = delete $args{'Type'} || 0;
2770
2771 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2772
2773 if (! $type{$addr}) {
2774 $standard_form{$addr} = main::standardize($value);
2775 }
2776
2777 return $self;
2778 }
2779
2780 use overload
2781 fallback => 0,
2782 qw("") => "_operator_stringify",
2783 "." => \&main::_operator_dot,
2784 ;
2785
2786 sub _operator_stringify {
2787 my $self = shift;
ffe43484 2788 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2789
2790 # Output it like '0041..0065 (value)'
2791 my $return = sprintf("%04X", $start{$addr})
2792 . '..'
2793 . sprintf("%04X", $end{$addr});
2794 my $value = $value{$addr};
2795 my $type = $type{$addr};
2796 $return .= ' (';
2797 $return .= "$value";
2798 $return .= ", Type=$type" if $type != 0;
2799 $return .= ')';
2800
2801 return $return;
2802 }
2803
2804 sub standard_form {
2805 # The standard form is the value itself if the standard form is
2806 # undefined (that is if the value is special)
2807
2808 my $self = shift;
2809 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2810
ffe43484 2811 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2812
2813 return $standard_form{$addr} if defined $standard_form{$addr};
2814 return $value{$addr};
2815 }
2816
2817 sub dump {
2818 # Human, not machine readable. For machine readable, comment out this
2819 # entire routine and let the standard one take effect.
2820 my $self = shift;
2821 my $indent = shift;
2822 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2823
ffe43484 2824 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2825
2826 my $return = $indent
2827 . sprintf("%04X", $start{$addr})
2828 . '..'
2829 . sprintf("%04X", $end{$addr})
2830 . " '$value{$addr}';";
2831 if (! defined $standard_form{$addr}) {
2832 $return .= "(type=$type{$addr})";
2833 }
2834 elsif ($standard_form{$addr} ne $value{$addr}) {
2835 $return .= "(standard '$standard_form{$addr}')";
2836 }
2837 return $return;
2838 }
2839} # End closure
2840
2841package _Range_List_Base;
2842
2843# Base class for range lists. A range list is simply an ordered list of
2844# ranges, so that the ranges with the lowest starting numbers are first in it.
2845#
2846# When a new range is added that is adjacent to an existing range that has the
2847# same value and type, it merges with it to form a larger range.
2848#
2849# Ranges generally do not overlap, except that there can be multiple entries
2850# of single code point ranges. This is because of NameAliases.txt.
2851#
2852# In this program, there is a standard value such that if two different
2853# values, have the same standard value, they are considered equivalent. This
2854# value was chosen so that it gives correct results on Unicode data
2855
2856# There are a number of methods to manipulate range lists, and some operators
2857# are overloaded to handle them.
2858
99870f4d
KW
2859sub trace { return main::trace(@_); }
2860
2861{ # Closure
2862
2863 our $addr;
2864
2865 main::setup_package();
2866
2867 my %ranges;
2868 # The list of ranges
2869 main::set_access('ranges', \%ranges, 'readable_array');
2870
2871 my %max;
2872 # The highest code point in the list. This was originally a method, but
2873 # actual measurements said it was used a lot.
2874 main::set_access('max', \%max, 'r');
2875
2876 my %each_range_iterator;
2877 # Iterator position for each_range()
2878 main::set_access('each_range_iterator', \%each_range_iterator);
2879
2880 my %owner_name_of;
2881 # Name of parent this is attached to, if any. Solely for better error
2882 # messages.
2883 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
2884
2885 my %_search_ranges_cache;
2886 # A cache of the previous result from _search_ranges(), for better
2887 # performance
2888 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
2889
2890 sub new {
2891 my $class = shift;
2892 my %args = @_;
2893
2894 # Optional initialization data for the range list.
2895 my $initialize = delete $args{'Initialize'};
2896
2897 my $self;
2898
2899 # Use _union() to initialize. _union() returns an object of this
2900 # class, which means that it will call this constructor recursively.
2901 # But it won't have this $initialize parameter so that it won't
2902 # infinitely loop on this.
2903 return _union($class, $initialize, %args) if defined $initialize;
2904
2905 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2906 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2907
2908 # Optional parent object, only for debug info.
2909 $owner_name_of{$addr} = delete $args{'Owner'};
2910 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
2911
2912 # Stringify, in case it is an object.
2913 $owner_name_of{$addr} = "$owner_name_of{$addr}";
2914
2915 # This is used only for error messages, and so a colon is added
2916 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
2917
2918 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2919
2920 # Max is initialized to a negative value that isn't adjacent to 0,
2921 # for simpler tests
2922 $max{$addr} = -2;
2923
2924 $_search_ranges_cache{$addr} = 0;
2925 $ranges{$addr} = [];
2926
2927 return $self;
2928 }
2929
2930 use overload
2931 fallback => 0,
2932 qw("") => "_operator_stringify",
2933 "." => \&main::_operator_dot,
2934 ;
2935
2936 sub _operator_stringify {
2937 my $self = shift;
ffe43484 2938 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2939
2940 return "Range_List attached to '$owner_name_of{$addr}'"
2941 if $owner_name_of{$addr};
2942 return "anonymous Range_List " . \$self;
2943 }
2944
2945 sub _union {
2946 # Returns the union of the input code points. It can be called as
2947 # either a constructor or a method. If called as a method, the result
2948 # will be a new() instance of the calling object, containing the union
2949 # of that object with the other parameter's code points; if called as
2950 # a constructor, the first parameter gives the class the new object
2951 # should be, and the second parameter gives the code points to go into
2952 # it.
2953 # In either case, there are two parameters looked at by this routine;
2954 # any additional parameters are passed to the new() constructor.
2955 #
2956 # The code points can come in the form of some object that contains
2957 # ranges, and has a conventionally named method to access them; or
2958 # they can be an array of individual code points (as integers); or
2959 # just a single code point.
2960 #
2961 # If they are ranges, this routine doesn't make any effort to preserve
2962 # the range values of one input over the other. Therefore this base
2963 # class should not allow _union to be called from other than
2964 # initialization code, so as to prevent two tables from being added
2965 # together where the range values matter. The general form of this
2966 # routine therefore belongs in a derived class, but it was moved here
2967 # to avoid duplication of code. The failure to overload this in this
2968 # class keeps it safe.
2969 #
2970
2971 my $self;
2972 my @args; # Arguments to pass to the constructor
2973
2974 my $class = shift;
2975
2976 # If a method call, will start the union with the object itself, and
2977 # the class of the new object will be the same as self.
2978 if (ref $class) {
2979 $self = $class;
2980 $class = ref $self;
2981 push @args, $self;
2982 }
2983
2984 # Add the other required parameter.
2985 push @args, shift;
2986 # Rest of parameters are passed on to the constructor
2987
2988 # Accumulate all records from both lists.
2989 my @records;
2990 for my $arg (@args) {
2991 #local $to_trace = 0 if main::DEBUG;
2992 trace "argument = $arg" if main::DEBUG && $to_trace;
2993 if (! defined $arg) {
2994 my $message = "";
2995 if (defined $self) {
f998e60c 2996 no overloading;
051df77b 2997 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2998 }
2999 Carp::my_carp_bug($message .= "Undefined argument to _union. No union done.");
3000 return;
3001 }
3002 $arg = [ $arg ] if ! ref $arg;
3003 my $type = ref $arg;
3004 if ($type eq 'ARRAY') {
3005 foreach my $element (@$arg) {
3006 push @records, Range->new($element, $element);
3007 }
3008 }
3009 elsif ($arg->isa('Range')) {
3010 push @records, $arg;
3011 }
3012 elsif ($arg->can('ranges')) {
3013 push @records, $arg->ranges;
3014 }
3015 else {
3016 my $message = "";
3017 if (defined $self) {
f998e60c 3018 no overloading;
051df77b 3019 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
3020 }
3021 Carp::my_carp_bug($message . "Cannot take the union of a $type. No union done.");
3022 return;
3023 }
3024 }
3025
3026 # Sort with the range containing the lowest ordinal first, but if
3027 # two ranges start at the same code point, sort with the bigger range
3028 # of the two first, because it takes fewer cycles.
3029 @records = sort { ($a->start <=> $b->start)
3030 or
3031 # if b is shorter than a, b->end will be
3032 # less than a->end, and we want to select
3033 # a, so want to return -1
3034 ($b->end <=> $a->end)
3035 } @records;
3036
3037 my $new = $class->new(@_);
3038
3039 # Fold in records so long as they add new information.
3040 for my $set (@records) {
3041 my $start = $set->start;
3042 my $end = $set->end;
3043 my $value = $set->value;
3044 if ($start > $new->max) {
3045 $new->_add_delete('+', $start, $end, $value);
3046 }
3047 elsif ($end > $new->max) {
3048 $new->_add_delete('+', $new->max +1, $end, $value);
3049 }
3050 }
3051
3052 return $new;
3053 }
3054
3055 sub range_count { # Return the number of ranges in the range list
3056 my $self = shift;
3057 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3058
f998e60c 3059 no overloading;
051df77b 3060 return scalar @{$ranges{pack 'J', $self}};
99870f4d
KW
3061 }
3062
3063 sub min {
3064 # Returns the minimum code point currently in the range list, or if
3065 # the range list is empty, 2 beyond the max possible. This is a
3066 # method because used so rarely, that not worth saving between calls,
3067 # and having to worry about changing it as ranges are added and
3068 # deleted.
3069
3070 my $self = shift;
3071 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3072
ffe43484 3073 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3074
3075 # If the range list is empty, return a large value that isn't adjacent
3076 # to any that could be in the range list, for simpler tests
6189eadc 3077 return $MAX_UNICODE_CODEPOINT + 2 unless scalar @{$ranges{$addr}};
99870f4d
KW
3078 return $ranges{$addr}->[0]->start;
3079 }
3080
3081 sub contains {
3082 # Boolean: Is argument in the range list? If so returns $i such that:
3083 # range[$i]->end < $codepoint <= range[$i+1]->end
3084 # which is one beyond what you want; this is so that the 0th range
3085 # doesn't return false
3086 my $self = shift;
3087 my $codepoint = shift;
3088 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3089
99870f4d
KW
3090 my $i = $self->_search_ranges($codepoint);
3091 return 0 unless defined $i;
3092
3093 # The search returns $i, such that
3094 # range[$i-1]->end < $codepoint <= range[$i]->end
3095 # So is in the table if and only iff it is at least the start position
3096 # of range $i.
f998e60c 3097 no overloading;
051df77b 3098 return 0 if $ranges{pack 'J', $self}->[$i]->start > $codepoint;
99870f4d
KW
3099 return $i + 1;
3100 }
3101
2f7a8815
KW
3102 sub containing_range {
3103 # Returns the range object that contains the code point, undef if none
3104
3105 my $self = shift;
3106 my $codepoint = shift;
3107 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3108
3109 my $i = $self->contains($codepoint);
3110 return unless $i;
3111
3112 # contains() returns 1 beyond where we should look
3113 no overloading;
3114 return $ranges{pack 'J', $self}->[$i-1];
3115 }
3116
99870f4d
KW
3117 sub value_of {
3118 # Returns the value associated with the code point, undef if none
3119
3120 my $self = shift;
3121 my $codepoint = shift;
3122 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3123
d69c231b
KW
3124 my $range = $self->containing_range($codepoint);
3125 return unless defined $range;
99870f4d 3126
d69c231b 3127 return $range->value;
99870f4d
KW
3128 }
3129
0a9dbafc
KW
3130 sub type_of {
3131 # Returns the type of the range containing the code point, undef if
3132 # the code point is not in the table
3133
3134 my $self = shift;
3135 my $codepoint = shift;
3136 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3137
3138 my $range = $self->containing_range($codepoint);
3139 return unless defined $range;
3140
3141 return $range->type;
3142 }
3143
99870f4d
KW
3144 sub _search_ranges {
3145 # Find the range in the list which contains a code point, or where it
3146 # should go if were to add it. That is, it returns $i, such that:
3147 # range[$i-1]->end < $codepoint <= range[$i]->end
3148 # Returns undef if no such $i is possible (e.g. at end of table), or
3149 # if there is an error.
3150
3151 my $self = shift;
3152 my $code_point = shift;
3153 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3154
ffe43484 3155 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3156
3157 return if $code_point > $max{$addr};
3158 my $r = $ranges{$addr}; # The current list of ranges
3159 my $range_list_size = scalar @$r;
3160 my $i;
3161
3162 use integer; # want integer division
3163
3164 # Use the cached result as the starting guess for this one, because,
3165 # an experiment on 5.1 showed that 90% of the time the cache was the
3166 # same as the result on the next call (and 7% it was one less).
3167 $i = $_search_ranges_cache{$addr};
3168 $i = 0 if $i >= $range_list_size; # Reset if no longer valid (prob.
3169 # from an intervening deletion
3170 #local $to_trace = 1 if main::DEBUG;
3171 trace "previous \$i is still valid: $i" if main::DEBUG && $to_trace && $code_point <= $r->[$i]->end && ($i == 0 || $r->[$i-1]->end < $code_point);
3172 return $i if $code_point <= $r->[$i]->end
3173 && ($i == 0 || $r->[$i-1]->end < $code_point);
3174
3175 # Here the cache doesn't yield the correct $i. Try adding 1.
3176 if ($i < $range_list_size - 1
3177 && $r->[$i]->end < $code_point &&
3178 $code_point <= $r->[$i+1]->end)
3179 {
3180 $i++;
3181 trace "next \$i is correct: $i" if main::DEBUG && $to_trace;
3182 $_search_ranges_cache{$addr} = $i;
3183 return $i;
3184 }
3185
3186 # Here, adding 1 also didn't work. We do a binary search to
3187 # find the correct position, starting with current $i
3188 my $lower = 0;
3189 my $upper = $range_list_size - 1;
3190 while (1) {
3191 trace "top of loop i=$i:", sprintf("%04X", $r->[$lower]->start), "[$lower] .. ", sprintf("%04X", $r->[$i]->start), "[$i] .. ", sprintf("%04X", $r->[$upper]->start), "[$upper]" if main::DEBUG && $to_trace;
3192
3193 if ($code_point <= $r->[$i]->end) {
3194
3195 # Here we have met the upper constraint. We can quit if we
3196 # also meet the lower one.
3197 last if $i == 0 || $r->[$i-1]->end < $code_point;
3198
3199 $upper = $i; # Still too high.
3200
3201 }
3202 else {
3203
3204 # Here, $r[$i]->end < $code_point, so look higher up.
3205 $lower = $i;
3206 }
3207
3208 # Split search domain in half to try again.
3209 my $temp = ($upper + $lower) / 2;
3210
3211 # No point in continuing unless $i changes for next time
3212 # in the loop.
3213 if ($temp == $i) {
3214
3215 # We can't reach the highest element because of the averaging.
3216 # So if one below the upper edge, force it there and try one
3217 # more time.
3218 if ($i == $range_list_size - 2) {
3219
3220 trace "Forcing to upper edge" if main::DEBUG && $to_trace;
3221 $i = $range_list_size - 1;
3222
3223 # Change $lower as well so if fails next time through,
3224 # taking the average will yield the same $i, and we will
3225 # quit with the error message just below.
3226 $lower = $i;
3227 next;
3228 }
3229 Carp::my_carp_bug("$owner_name_of{$addr}Can't find where the range ought to go. No action taken.");
3230 return;
3231 }
3232 $i = $temp;
3233 } # End of while loop
3234
3235 if (main::DEBUG && $to_trace) {
3236 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i;
3237 trace "i= [ $i ]", $r->[$i];
3238 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < $range_list_size - 1;
3239 }
3240
3241 # Here we have found the offset. Cache it as a starting point for the
3242 # next call.
3243 $_search_ranges_cache{$addr} = $i;
3244 return $i;
3245 }
3246
3247 sub _add_delete {
3248 # Add, replace or delete ranges to or from a list. The $type
3249 # parameter gives which:
3250 # '+' => insert or replace a range, returning a list of any changed
3251 # ranges.
3252 # '-' => delete a range, returning a list of any deleted ranges.
3253 #
3254 # The next three parameters give respectively the start, end, and
3255 # value associated with the range. 'value' should be null unless the
3256 # operation is '+';
3257 #
3258 # The range list is kept sorted so that the range with the lowest
3259 # starting position is first in the list, and generally, adjacent
c1739a4a 3260 # ranges with the same values are merged into a single larger one (see
99870f4d
KW
3261 # exceptions below).
3262 #
c1739a4a 3263 # There are more parameters; all are key => value pairs:
99870f4d
KW
3264 # Type gives the type of the value. It is only valid for '+'.
3265 # All ranges have types; if this parameter is omitted, 0 is
3266 # assumed. Ranges with type 0 are assumed to obey the
3267 # Unicode rules for casing, etc; ranges with other types are
3268 # not. Otherwise, the type is arbitrary, for the caller's
3269 # convenience, and looked at only by this routine to keep
3270 # adjacent ranges of different types from being merged into
3271 # a single larger range, and when Replace =>
3272 # $IF_NOT_EQUIVALENT is specified (see just below).
3273 # Replace determines what to do if the range list already contains
3274 # ranges which coincide with all or portions of the input
3275 # range. It is only valid for '+':
3276 # => $NO means that the new value is not to replace
3277 # any existing ones, but any empty gaps of the
3278 # range list coinciding with the input range
3279 # will be filled in with the new value.
3280 # => $UNCONDITIONALLY means to replace the existing values with
3281 # this one unconditionally. However, if the
3282 # new and old values are identical, the
3283 # replacement is skipped to save cycles
3284 # => $IF_NOT_EQUIVALENT means to replace the existing values
3285 # with this one if they are not equivalent.
3286 # Ranges are equivalent if their types are the
c1739a4a 3287 # same, and they are the same string; or if
99870f4d
KW
3288 # both are type 0 ranges, if their Unicode
3289 # standard forms are identical. In this last
3290 # case, the routine chooses the more "modern"
3291 # one to use. This is because some of the
3292 # older files are formatted with values that
3293 # are, for example, ALL CAPs, whereas the
3294 # derived files have a more modern style,
3295 # which looks better. By looking for this
3296 # style when the pre-existing and replacement
3297 # standard forms are the same, we can move to
3298 # the modern style
3299 # => $MULTIPLE means that if this range duplicates an
3300 # existing one, but has a different value,
3301 # don't replace the existing one, but insert
3302 # this, one so that the same range can occur
53d84487
KW
3303 # multiple times. They are stored LIFO, so
3304 # that the final one inserted is the first one
3305 # returned in an ordered search of the table.
99870f4d
KW
3306 # => anything else is the same as => $IF_NOT_EQUIVALENT
3307 #
c1739a4a
KW
3308 # "same value" means identical for non-type-0 ranges, and it means
3309 # having the same standard forms for type-0 ranges.
99870f4d
KW
3310
3311 return Carp::carp_too_few_args(\@_, 5) if main::DEBUG && @_ < 5;
3312
3313 my $self = shift;
3314 my $operation = shift; # '+' for add/replace; '-' for delete;
3315 my $start = shift;
3316 my $end = shift;
3317 my $value = shift;
3318
3319 my %args = @_;
3320
3321 $value = "" if not defined $value; # warning: $value can be "0"
3322
3323 my $replace = delete $args{'Replace'};
3324 $replace = $IF_NOT_EQUIVALENT unless defined $replace;
3325
3326 my $type = delete $args{'Type'};
3327 $type = 0 unless defined $type;
3328
3329 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3330
ffe43484 3331 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3332
3333 if ($operation ne '+' && $operation ne '-') {
3334 Carp::my_carp_bug("$owner_name_of{$addr}First parameter to _add_delete must be '+' or '-'. No action taken.");
3335 return;
3336 }
3337 unless (defined $start && defined $end) {
3338 Carp::my_carp_bug("$owner_name_of{$addr}Undefined start and/or end to _add_delete. No action taken.");
3339 return;
3340 }
3341 unless ($end >= $start) {
3342 Carp::my_carp_bug("$owner_name_of{$addr}End of range (" . sprintf("%04X", $end) . ") must not be before start (" . sprintf("%04X", $start) . "). No action taken.");
3343 return;
3344 }
3345 #local $to_trace = 1 if main::DEBUG;
3346
3347 if ($operation eq '-') {
3348 if ($replace != $IF_NOT_EQUIVALENT) {
3349 Carp::my_carp_bug("$owner_name_of{$addr}Replace => \$IF_NOT_EQUIVALENT is required when deleting a range from a range list. Assuming Replace => \$IF_NOT_EQUIVALENT.");
3350 $replace = $IF_NOT_EQUIVALENT;
3351 }
3352 if ($type) {
3353 Carp::my_carp_bug("$owner_name_of{$addr}Type => 0 is required when deleting a range from a range list. Assuming Type => 0.");
3354 $type = 0;
3355 }
3356 if ($value ne "") {
3357 Carp::my_carp_bug("$owner_name_of{$addr}Value => \"\" is required when deleting a range from a range list. Assuming Value => \"\".");
3358 $value = "";
3359 }
3360 }
3361
3362 my $r = $ranges{$addr}; # The current list of ranges
3363 my $range_list_size = scalar @$r; # And its size
3364 my $max = $max{$addr}; # The current high code point in
3365 # the list of ranges
3366
3367 # Do a special case requiring fewer machine cycles when the new range
3368 # starts after the current highest point. The Unicode input data is
3369 # structured so this is common.
3370 if ($start > $max) {
3371
3372 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) type=$type" if main::DEBUG && $to_trace;
3373 return if $operation eq '-'; # Deleting a non-existing range is a
3374 # no-op
3375
3376 # If the new range doesn't logically extend the current final one
3377 # in the range list, create a new range at the end of the range
3378 # list. (max cleverly is initialized to a negative number not
3379 # adjacent to 0 if the range list is empty, so even adding a range
3380 # to an empty range list starting at 0 will have this 'if'
3381 # succeed.)
3382 if ($start > $max + 1 # non-adjacent means can't extend.
3383 || @{$r}[-1]->value ne $value # values differ, can't extend.
3384 || @{$r}[-1]->type != $type # types differ, can't extend.
3385 ) {
3386 push @$r, Range->new($start, $end,
3387 Value => $value,
3388 Type => $type);
3389 }
3390 else {
3391
3392 # Here, the new range starts just after the current highest in
3393 # the range list, and they have the same type and value.
3394 # Extend the current range to incorporate the new one.
3395 @{$r}[-1]->set_end($end);
3396 }
3397
3398 # This becomes the new maximum.
3399 $max{$addr} = $end;
3400
3401 return;
3402 }
3403 #local $to_trace = 0 if main::DEBUG;
3404
3405 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) replace=$replace" if main::DEBUG && $to_trace;
3406
3407 # Here, the input range isn't after the whole rest of the range list.
3408 # Most likely 'splice' will be needed. The rest of the routine finds
3409 # the needed splice parameters, and if necessary, does the splice.
3410 # First, find the offset parameter needed by the splice function for
3411 # the input range. Note that the input range may span multiple
3412 # existing ones, but we'll worry about that later. For now, just find
3413 # the beginning. If the input range is to be inserted starting in a
3414 # position not currently in the range list, it must (obviously) come
3415 # just after the range below it, and just before the range above it.
3416 # Slightly less obviously, it will occupy the position currently
3417 # occupied by the range that is to come after it. More formally, we
3418 # are looking for the position, $i, in the array of ranges, such that:
3419 #
3420 # r[$i-1]->start <= r[$i-1]->end < $start < r[$i]->start <= r[$i]->end
3421 #
3422 # (The ordered relationships within existing ranges are also shown in
3423 # the equation above). However, if the start of the input range is
3424 # within an existing range, the splice offset should point to that
3425 # existing range's position in the list; that is $i satisfies a
3426 # somewhat different equation, namely:
3427 #
3428 #r[$i-1]->start <= r[$i-1]->end < r[$i]->start <= $start <= r[$i]->end
3429 #
3430 # More briefly, $start can come before or after r[$i]->start, and at
3431 # this point, we don't know which it will be. However, these
3432 # two equations share these constraints:
3433 #
3434 # r[$i-1]->end < $start <= r[$i]->end
3435 #
3436 # And that is good enough to find $i.
3437
3438 my $i = $self->_search_ranges($start);
3439 if (! defined $i) {
3440 Carp::my_carp_bug("Searching $self for range beginning with $start unexpectedly returned undefined. Operation '$operation' not performed");
3441 return;
3442 }
3443
3444 # The search function returns $i such that:
3445 #
3446 # r[$i-1]->end < $start <= r[$i]->end
3447 #
3448 # That means that $i points to the first range in the range list
3449 # that could possibly be affected by this operation. We still don't
3450 # know if the start of the input range is within r[$i], or if it
3451 # points to empty space between r[$i-1] and r[$i].
3452 trace "[$i] is the beginning splice point. Existing range there is ", $r->[$i] if main::DEBUG && $to_trace;
3453
3454 # Special case the insertion of data that is not to replace any
3455 # existing data.
3456 if ($replace == $NO) { # If $NO, has to be operation '+'
3457 #local $to_trace = 1 if main::DEBUG;
3458 trace "Doesn't replace" if main::DEBUG && $to_trace;
3459
3460 # Here, the new range is to take effect only on those code points
3461 # that aren't already in an existing range. This can be done by
3462 # looking through the existing range list and finding the gaps in
3463 # the ranges that this new range affects, and then calling this
3464 # function recursively on each of those gaps, leaving untouched
3465 # anything already in the list. Gather up a list of the changed
3466 # gaps first so that changes to the internal state as new ranges
3467 # are added won't be a problem.
3468 my @gap_list;
3469
3470 # First, if the starting point of the input range is outside an
3471 # existing one, there is a gap from there to the beginning of the
3472 # existing range -- add a span to fill the part that this new
3473 # range occupies
3474 if ($start < $r->[$i]->start) {
3475 push @gap_list, Range->new($start,
3476 main::min($end,
3477 $r->[$i]->start - 1),
3478 Type => $type);
3479 trace "gap before $r->[$i] [$i], will add", $gap_list[-1] if main::DEBUG && $to_trace;
3480 }
3481
3482 # Then look through the range list for other gaps until we reach
3483 # the highest range affected by the input one.
3484 my $j;
3485 for ($j = $i+1; $j < $range_list_size; $j++) {
3486 trace "j=[$j]", $r->[$j] if main::DEBUG && $to_trace;
3487 last if $end < $r->[$j]->start;
3488
3489 # If there is a gap between when this range starts and the
3490 # previous one ends, add a span to fill it. Note that just
3491 # because there are two ranges doesn't mean there is a
3492 # non-zero gap between them. It could be that they have
3493 # different values or types
3494 if ($r->[$j-1]->end + 1 != $r->[$j]->start) {
3495 push @gap_list,
3496 Range->new($r->[$j-1]->end + 1,
3497 $r->[$j]->start - 1,
3498 Type => $type);
3499 trace "gap between $r->[$j-1] and $r->[$j] [$j], will add: $gap_list[-1]" if main::DEBUG && $to_trace;
3500 }
3501 }
3502
3503 # Here, we have either found an existing range in the range list,
3504 # beyond the area affected by the input one, or we fell off the
3505 # end of the loop because the input range affects the whole rest
3506 # of the range list. In either case, $j is 1 higher than the
3507 # highest affected range. If $j == $i, it means that there are no
3508 # affected ranges, that the entire insertion is in the gap between
3509 # r[$i-1], and r[$i], which we already have taken care of before
3510 # the loop.
3511 # On the other hand, if there are affected ranges, it might be
3512 # that there is a gap that needs filling after the final such
3513 # range to the end of the input range
3514 if ($r->[$j-1]->end < $end) {
3515 push @gap_list, Range->new(main::max($start,
3516 $r->[$j-1]->end + 1),
3517 $end,
3518 Type => $type);
3519 trace "gap after $r->[$j-1], will add $gap_list[-1]" if main::DEBUG && $to_trace;
3520 }
3521
3522 # Call recursively to fill in all the gaps.
3523 foreach my $gap (@gap_list) {
3524 $self->_add_delete($operation,
3525 $gap->start,
3526 $gap->end,
3527 $value,
3528 Type => $type);
3529 }
3530
3531 return;
3532 }
3533
53d84487
KW
3534 # Here, we have taken care of the case where $replace is $NO.
3535 # Remember that here, r[$i-1]->end < $start <= r[$i]->end
3536 # If inserting a multiple record, this is where it goes, before the
3537 # first (if any) existing one. This implies an insertion, and no
3538 # change to any existing ranges. Note that $i can be -1 if this new
3539 # range doesn't actually duplicate any existing, and comes at the
3540 # beginning of the list.
3541 if ($replace == $MULTIPLE) {
3542
3543 if ($start != $end) {
3544 Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple record when the range ($start..$end) contains more than one code point. No action taken.");
3545 return;
3546 }
3547
3548 # Don't add an exact duplicate, as it isn't really a multiple
3549 if ($end >= $r->[$i]->start) {
1f6798c4
KW
3550 my $existing_value = $r->[$i]->value;
3551 my $existing_type = $r->[$i]->type;
3552 return if $value eq $existing_value && $type eq $existing_type;
3553
3554 # If the multiple value is part of an existing range, we want
3555 # to split up that range, so that only the single code point
3556 # is affected. To do this, we first call ourselves
3557 # recursively to delete that code point from the table, having
3558 # preserved its current data above. Then we call ourselves
3559 # recursively again to add the new multiple, which we know by
3560 # the test just above is different than the current code
3561 # point's value, so it will become a range containing a single
3562 # code point: just itself. Finally, we add back in the
3563 # pre-existing code point, which will again be a single code
3564 # point range. Because 'i' likely will have changed as a
3565 # result of these operations, we can't just continue on, but
3566 # do this operation recursively as well.
53d84487 3567 if ($r->[$i]->start != $r->[$i]->end) {
1f6798c4
KW
3568 $self->_add_delete('-', $start, $end, "");
3569 $self->_add_delete('+', $start, $end, $value, Type => $type);
3570 return $self->_add_delete('+', $start, $end, $existing_value, Type => $existing_type, Replace => $MULTIPLE);
53d84487 3571 }
53d84487
KW
3572 }
3573
3574 trace "Adding multiple record at $i with $start..$end, $value" if main::DEBUG && $to_trace;
3575 my @return = splice @$r,
3576 $i,
3577 0,
3578 Range->new($start,
3579 $end,
3580 Value => $value,
3581 Type => $type);
3582 if (main::DEBUG && $to_trace) {
3583 trace "After splice:";
3584 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3585 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3586 trace "i =[", $i, "]", $r->[$i] if $i >= 0;
3587 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3588 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3589 trace 'i+3=[', $i+3, ']', $r->[$i+3] if $i < @$r - 3;
3590 }
3591 return @return;
3592 }
3593
3594 # Here, we have taken care of $NO and $MULTIPLE replaces. This leaves
3595 # delete, insert, and replace either unconditionally or if not
3596 # equivalent. $i still points to the first potential affected range.
3597 # Now find the highest range affected, which will determine the length
3598 # parameter to splice. (The input range can span multiple existing
3599 # ones.) If this isn't a deletion, while we are looking through the
3600 # range list, see also if this is a replacement rather than a clean
3601 # insertion; that is if it will change the values of at least one
3602 # existing range. Start off assuming it is an insert, until find it
3603 # isn't.
3604 my $clean_insert = $operation eq '+';
99870f4d
KW
3605 my $j; # This will point to the highest affected range
3606
3607 # For non-zero types, the standard form is the value itself;
3608 my $standard_form = ($type) ? $value : main::standardize($value);
3609
3610 for ($j = $i; $j < $range_list_size; $j++) {
3611 trace "Looking for highest affected range; the one at $j is ", $r->[$j] if main::DEBUG && $to_trace;
3612
3613 # If find a range that it doesn't overlap into, we can stop
3614 # searching
3615 last if $end < $r->[$j]->start;
3616
969a34cc
KW
3617 # Here, overlaps the range at $j. If the values don't match,
3618 # and so far we think this is a clean insertion, it becomes a
3619 # non-clean insertion, i.e., a 'change' or 'replace' instead.
3620 if ($clean_insert) {
99870f4d 3621 if ($r->[$j]->standard_form ne $standard_form) {
969a34cc 3622 $clean_insert = 0;
56343c78
KW
3623 if ($replace == $CROAK) {
3624 main::croak("The range to add "
3625 . sprintf("%04X", $start)
3626 . '-'
3627 . sprintf("%04X", $end)
3628 . " with value '$value' overlaps an existing range $r->[$j]");
3629 }
99870f4d
KW
3630 }
3631 else {
3632
3633 # Here, the two values are essentially the same. If the
3634 # two are actually identical, replacing wouldn't change
3635 # anything so skip it.
3636 my $pre_existing = $r->[$j]->value;
3637 if ($pre_existing ne $value) {
3638
3639 # Here the new and old standardized values are the
3640 # same, but the non-standardized values aren't. If
3641 # replacing unconditionally, then replace
3642 if( $replace == $UNCONDITIONALLY) {
969a34cc 3643 $clean_insert = 0;
99870f4d
KW
3644 }
3645 else {
3646
3647 # Here, are replacing conditionally. Decide to
3648 # replace or not based on which appears to look
3649 # the "nicest". If one is mixed case and the
3650 # other isn't, choose the mixed case one.
3651 my $new_mixed = $value =~ /[A-Z]/
3652 && $value =~ /[a-z]/;
3653 my $old_mixed = $pre_existing =~ /[A-Z]/
3654 && $pre_existing =~ /[a-z]/;
3655
3656 if ($old_mixed != $new_mixed) {
969a34cc 3657 $clean_insert = 0 if $new_mixed;
99870f4d 3658 if (main::DEBUG && $to_trace) {
969a34cc
KW
3659 if ($clean_insert) {
3660 trace "Retaining $pre_existing over $value";
99870f4d
KW
3661 }
3662 else {
969a34cc 3663 trace "Replacing $pre_existing with $value";
99870f4d
KW
3664 }
3665 }
3666 }
3667 else {
3668
3669 # Here casing wasn't different between the two.
3670 # If one has hyphens or underscores and the
3671 # other doesn't, choose the one with the
3672 # punctuation.
3673 my $new_punct = $value =~ /[-_]/;
3674 my $old_punct = $pre_existing =~ /[-_]/;
3675
3676 if ($old_punct != $new_punct) {
969a34cc 3677 $clean_insert = 0 if $new_punct;
99870f4d 3678 if (main::DEBUG && $to_trace) {
969a34cc
KW
3679 if ($clean_insert) {
3680 trace "Retaining $pre_existing over $value";
99870f4d
KW
3681 }
3682 else {
969a34cc 3683 trace "Replacing $pre_existing with $value";
99870f4d
KW
3684 }
3685 }
3686 } # else existing one is just as "good";
3687 # retain it to save cycles.
3688 }
3689 }
3690 }
3691 }
3692 }
3693 } # End of loop looking for highest affected range.
3694
3695 # Here, $j points to one beyond the highest range that this insertion
3696 # affects (hence to beyond the range list if that range is the final
3697 # one in the range list).
3698
3699 # The splice length is all the affected ranges. Get it before
3700 # subtracting, for efficiency, so we don't have to later add 1.
3701 my $length = $j - $i;
3702
3703 $j--; # $j now points to the highest affected range.
3704 trace "Final affected range is $j: $r->[$j]" if main::DEBUG && $to_trace;
3705
99870f4d
KW
3706 # Here, have taken care of $NO and $MULTIPLE replaces.
3707 # $j points to the highest affected range. But it can be < $i or even
3708 # -1. These happen only if the insertion is entirely in the gap
3709 # between r[$i-1] and r[$i]. Here's why: j < i means that the j loop
3710 # above exited first time through with $end < $r->[$i]->start. (And
3711 # then we subtracted one from j) This implies also that $start <
3712 # $r->[$i]->start, but we know from above that $r->[$i-1]->end <
3713 # $start, so the entire input range is in the gap.
3714 if ($j < $i) {
3715
3716 # Here the entire input range is in the gap before $i.
3717
3718 if (main::DEBUG && $to_trace) {
3719 if ($i) {
3720 trace "Entire range is between $r->[$i-1] and $r->[$i]";
3721 }
3722 else {
3723 trace "Entire range is before $r->[$i]";
3724 }
3725 }
3726 return if $operation ne '+'; # Deletion of a non-existent range is
3727 # a no-op
3728 }
3729 else {
3730
969a34cc
KW
3731 # Here part of the input range is not in the gap before $i. Thus,
3732 # there is at least one affected one, and $j points to the highest
3733 # such one.
99870f4d
KW
3734
3735 # At this point, here is the situation:
3736 # This is not an insertion of a multiple, nor of tentative ($NO)
3737 # data.
3738 # $i points to the first element in the current range list that
3739 # may be affected by this operation. In fact, we know
3740 # that the range at $i is affected because we are in
3741 # the else branch of this 'if'
3742 # $j points to the highest affected range.
3743 # In other words,
3744 # r[$i-1]->end < $start <= r[$i]->end
3745 # And:
3746 # r[$i-1]->end < $start <= $end <= r[$j]->end
3747 #
3748 # Also:
969a34cc
KW
3749 # $clean_insert is a boolean which is set true if and only if
3750 # this is a "clean insertion", i.e., not a change nor a
3751 # deletion (multiple was handled above).
99870f4d
KW
3752
3753 # We now have enough information to decide if this call is a no-op
969a34cc
KW
3754 # or not. It is a no-op if this is an insertion of already
3755 # existing data.
99870f4d 3756
969a34cc 3757 if (main::DEBUG && $to_trace && $clean_insert
99870f4d
KW
3758 && $i == $j
3759 && $start >= $r->[$i]->start)
3760 {
3761 trace "no-op";
3762 }
969a34cc 3763 return if $clean_insert
99870f4d
KW
3764 && $i == $j # more than one affected range => not no-op
3765
3766 # Here, r[$i-1]->end < $start <= $end <= r[$i]->end
3767 # Further, $start and/or $end is >= r[$i]->start
3768 # The test below hence guarantees that
3769 # r[$i]->start < $start <= $end <= r[$i]->end
3770 # This means the input range is contained entirely in
3771 # the one at $i, so is a no-op
3772 && $start >= $r->[$i]->start;
3773 }
3774
3775 # Here, we know that some action will have to be taken. We have
3776 # calculated the offset and length (though adjustments may be needed)
3777 # for the splice. Now start constructing the replacement list.
3778 my @replacement;
3779 my $splice_start = $i;
3780
3781 my $extends_below;
3782 my $extends_above;
3783
3784 # See if should extend any adjacent ranges.
3785 if ($operation eq '-') { # Don't extend deletions
3786 $extends_below = $extends_above = 0;
3787 }
3788 else { # Here, should extend any adjacent ranges. See if there are
3789 # any.
3790 $extends_below = ($i > 0
3791 # can't extend unless adjacent
3792 && $r->[$i-1]->end == $start -1
3793 # can't extend unless are same standard value
3794 && $r->[$i-1]->standard_form eq $standard_form
3795 # can't extend unless share type
3796 && $r->[$i-1]->type == $type);
3797 $extends_above = ($j+1 < $range_list_size
3798 && $r->[$j+1]->start == $end +1
3799 && $r->[$j+1]->standard_form eq $standard_form
23822bda 3800 && $r->[$j+1]->type == $type);
99870f4d
KW
3801 }
3802 if ($extends_below && $extends_above) { # Adds to both
3803 $splice_start--; # start replace at element below
3804 $length += 2; # will replace on both sides
3805 trace "Extends both below and above ranges" if main::DEBUG && $to_trace;
3806
3807 # The result will fill in any gap, replacing both sides, and
3808 # create one large range.
3809 @replacement = Range->new($r->[$i-1]->start,
3810 $r->[$j+1]->end,
3811 Value => $value,
3812 Type => $type);
3813 }
3814 else {
3815
3816 # Here we know that the result won't just be the conglomeration of
3817 # a new range with both its adjacent neighbors. But it could
3818 # extend one of them.
3819
3820 if ($extends_below) {
3821
3822 # Here the new element adds to the one below, but not to the
3823 # one above. If inserting, and only to that one range, can
3824 # just change its ending to include the new one.
969a34cc 3825 if ($length == 0 && $clean_insert) {
99870f4d
KW
3826 $r->[$i-1]->set_end($end);
3827 trace "inserted range extends range to below so it is now $r->[$i-1]" if main::DEBUG && $to_trace;
3828 return;
3829 }
3830 else {
3831 trace "Changing inserted range to start at ", sprintf("%04X", $r->[$i-1]->start), " instead of ", sprintf("%04X", $start) if main::DEBUG && $to_trace;
3832 $splice_start--; # start replace at element below
3833 $length++; # will replace the element below
3834 $start = $r->[$i-1]->start;
3835 }
3836 }
3837 elsif ($extends_above) {
3838
3839 # Here the new element adds to the one above, but not below.
3840 # Mirror the code above
969a34cc 3841 if ($length == 0 && $clean_insert) {
99870f4d
KW
3842 $r->[$j+1]->set_start($start);
3843 trace "inserted range extends range to above so it is now $r->[$j+1]" if main::DEBUG && $to_trace;
3844 return;
3845 }
3846 else {
3847 trace "Changing inserted range to end at ", sprintf("%04X", $r->[$j+1]->end), " instead of ", sprintf("%04X", $end) if main::DEBUG && $to_trace;
3848 $length++; # will replace the element above
3849 $end = $r->[$j+1]->end;
3850 }
3851 }
3852
3853 trace "Range at $i is $r->[$i]" if main::DEBUG && $to_trace;
3854
3855 # Finally, here we know there will have to be a splice.
3856 # If the change or delete affects only the highest portion of the
3857 # first affected range, the range will have to be split. The
3858 # splice will remove the whole range, but will replace it by a new
3859 # range containing just the unaffected part. So, in this case,
3860 # add to the replacement list just this unaffected portion.
3861 if (! $extends_below
3862 && $start > $r->[$i]->start && $start <= $r->[$i]->end)
3863 {
3864 push @replacement,
3865 Range->new($r->[$i]->start,
3866 $start - 1,
3867 Value => $r->[$i]->value,
3868 Type => $r->[$i]->type);
3869 }
3870
3871 # In the case of an insert or change, but not a delete, we have to
3872 # put in the new stuff; this comes next.
3873 if ($operation eq '+') {
3874 push @replacement, Range->new($start,
3875 $end,
3876 Value => $value,
3877 Type => $type);
3878 }
3879
3880 trace "Range at $j is $r->[$j]" if main::DEBUG && $to_trace && $j != $i;
3881 #trace "$end >=", $r->[$j]->start, " && $end <", $r->[$j]->end if main::DEBUG && $to_trace;
3882
3883 # And finally, if we're changing or deleting only a portion of the
3884 # highest affected range, it must be split, as the lowest one was.
3885 if (! $extends_above
3886 && $j >= 0 # Remember that j can be -1 if before first
3887 # current element
3888 && $end >= $r->[$j]->start
3889 && $end < $r->[$j]->e