This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
mktables: Work around deficiency in Unicode data 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
25
23e33b60 26require 5.010_001;
d73e5302 27use strict;
99870f4d 28use warnings;
cf25bb62 29use Carp;
bd9ebcfd 30use Config;
99870f4d
KW
31use File::Find;
32use File::Path;
d07a55ed 33use File::Spec;
99870f4d
KW
34use Text::Tabs;
35
36sub DEBUG () { 0 } # Set to 0 for production; 1 for development
bd9ebcfd 37my $debugging_build = $Config{"ccflags"} =~ /-DDEBUGGING/;
99870f4d
KW
38
39##########################################################################
40#
41# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
42# from the Unicode database files (lib/unicore/.../*.txt), It also generates
43# a pod file and a .t file
44#
45# The structure of this file is:
46# First these introductory comments; then
47# code needed for everywhere, such as debugging stuff; then
48# code to handle input parameters; then
49# data structures likely to be of external interest (some of which depend on
50# the input parameters, so follows them; then
51# more data structures and subroutine and package (class) definitions; then
52# the small actual loop to process the input files and finish up; then
53# a __DATA__ section, for the .t tests
54#
5f7264c7 55# This program works on all releases of Unicode through at least 6.0. The
99870f4d
KW
56# outputs have been scrutinized most intently for release 5.1. The others
57# have been checked for somewhat more than just sanity. It can handle all
58# existing Unicode character properties in those releases.
59#
99870f4d
KW
60# This program is mostly about Unicode character (or code point) properties.
61# A property describes some attribute or quality of a code point, like if it
62# is lowercase or not, its name, what version of Unicode it was first defined
63# in, or what its uppercase equivalent is. Unicode deals with these disparate
64# possibilities by making all properties into mappings from each code point
65# into some corresponding value. In the case of it being lowercase or not,
66# the mapping is either to 'Y' or 'N' (or various synonyms thereof). Each
67# property maps each Unicode code point to a single value, called a "property
68# value". (Hence each Unicode property is a true mathematical function with
69# exactly one value per code point.)
70#
71# When using a property in a regular expression, what is desired isn't the
72# mapping of the code point to its property's value, but the reverse (or the
73# mathematical "inverse relation"): starting with the property value, "Does a
74# code point map to it?" These are written in a "compound" form:
75# \p{property=value}, e.g., \p{category=punctuation}. This program generates
76# files containing the lists of code points that map to each such regular
77# expression property value, one file per list
78#
79# There is also a single form shortcut that Perl adds for many of the commonly
80# used properties. This happens for all binary properties, plus script,
81# general_category, and block properties.
82#
83# Thus the outputs of this program are files. There are map files, mostly in
84# the 'To' directory; and there are list files for use in regular expression
85# matching, all in subdirectories of the 'lib' directory, with each
86# subdirectory being named for the property that the lists in it are for.
87# Bookkeeping, test, and documentation files are also generated.
88
89my $matches_directory = 'lib'; # Where match (\p{}) files go.
90my $map_directory = 'To'; # Where map files go.
91
92# DATA STRUCTURES
93#
94# The major data structures of this program are Property, of course, but also
95# Table. There are two kinds of tables, very similar to each other.
96# "Match_Table" is the data structure giving the list of code points that have
97# a particular property value, mentioned above. There is also a "Map_Table"
98# data structure which gives the property's mapping from code point to value.
99# There are two structures because the match tables need to be combined in
100# various ways, such as constructing unions, intersections, complements, etc.,
101# and the map ones don't. And there would be problems, perhaps subtle, if
102# a map table were inadvertently operated on in some of those ways.
103# The use of separate classes with operations defined on one but not the other
104# prevents accidentally confusing the two.
105#
106# At the heart of each table's data structure is a "Range_List", which is just
107# an ordered list of "Ranges", plus ancillary information, and methods to
108# operate on them. A Range is a compact way to store property information.
109# Each range has a starting code point, an ending code point, and a value that
110# is meant to apply to all the code points between the two end points,
111# inclusive. For a map table, this value is the property value for those
112# code points. Two such ranges could be written like this:
113# 0x41 .. 0x5A, 'Upper',
114# 0x61 .. 0x7A, 'Lower'
115#
116# Each range also has a type used as a convenience to classify the values.
117# Most ranges in this program will be Type 0, or normal, but there are some
118# ranges that have a non-zero type. These are used only in map tables, and
119# are for mappings that don't fit into the normal scheme of things. Mappings
120# that require a hash entry to communicate with utf8.c are one example;
121# another example is mappings for charnames.pm to use which indicate a name
122# that is algorithmically determinable from its code point (and vice-versa).
123# These are used to significantly compact these tables, instead of listing
124# each one of the tens of thousands individually.
125#
126# In a match table, the value of a range is irrelevant (and hence the type as
127# well, which will always be 0), and arbitrarily set to the null string.
128# Using the example above, there would be two match tables for those two
129# entries, one named Upper would contain the 0x41..0x5A range, and the other
130# named Lower would contain 0x61..0x7A.
131#
132# Actually, there are two types of range lists, "Range_Map" is the one
133# associated with map tables, and "Range_List" with match tables.
134# Again, this is so that methods can be defined on one and not the other so as
135# to prevent operating on them in incorrect ways.
136#
137# Eventually, most tables are written out to files to be read by utf8_heavy.pl
138# in the perl core. All tables could in theory be written, but some are
139# suppressed because there is no current practical use for them. It is easy
140# to change which get written by changing various lists that are near the top
141# of the actual code in this file. The table data structures contain enough
142# ancillary information to allow them to be treated as separate entities for
143# writing, such as the path to each one's file. There is a heading in each
144# map table that gives the format of its entries, and what the map is for all
145# the code points missing from it. (This allows tables to be more compact.)
678f13d5 146#
99870f4d
KW
147# The Property data structure contains one or more tables. All properties
148# contain a map table (except the $perl property which is a
149# pseudo-property containing only match tables), and any properties that
150# are usable in regular expression matches also contain various matching
151# tables, one for each value the property can have. A binary property can
152# have two values, True and False (or Y and N, which are preferred by Unicode
153# terminology). Thus each of these properties will have a map table that
154# takes every code point and maps it to Y or N (but having ranges cuts the
155# number of entries in that table way down), and two match tables, one
156# which has a list of all the code points that map to Y, and one for all the
157# code points that map to N. (For each of these, a third table is also
158# generated for the pseudo Perl property. It contains the identical code
159# points as the Y table, but can be written, not in the compound form, but in
160# a "single" form like \p{IsUppercase}.) Many properties are binary, but some
161# properties have several possible values, some have many, and properties like
162# Name have a different value for every named code point. Those will not,
163# unless the controlling lists are changed, have their match tables written
164# out. But all the ones which can be used in regular expression \p{} and \P{}
165# constructs will. Generally a property will have either its map table or its
166# match tables written but not both. Again, what gets written is controlled
dc85bd38
KW
167# by lists which can easily be changed. Properties have a 'Type', like
168# binary, or string, or enum depending on how many match tables there are and
169# the content of the maps. This 'Type' is different than a range 'Type', so
170# don't get confused by the two concepts having the same name.
678f13d5 171#
99870f4d
KW
172# For information about the Unicode properties, see Unicode's UAX44 document:
173
174my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
175
176# As stated earlier, this program will work on any release of Unicode so far.
177# Most obvious problems in earlier data have NOT been corrected except when
178# necessary to make Perl or this program work reasonably. For example, no
179# folding information was given in early releases, so this program uses the
180# substitute of lower case, just so that a regular expression with the /i
181# option will do something that actually gives the right results in many
182# cases. There are also a couple other corrections for version 1.1.5,
183# commented at the point they are made. As an example of corrections that
184# weren't made (but could be) is this statement from DerivedAge.txt: "The
185# supplementary private use code points and the non-character code points were
186# assigned in version 2.0, but not specifically listed in the UCD until
187# versions 3.0 and 3.1 respectively." (To be precise it was 3.0.1 not 3.0.0)
188# More information on Unicode version glitches is further down in these
189# introductory comments.
190#
5f7264c7
KW
191# This program works on all non-provisional properties as of 6.0, though the
192# files for some are suppressed from apparent lack of demand for them. You
193# can change which are output by changing lists in this program.
678f13d5 194#
dc85bd38 195# The old version of mktables emphasized the term "Fuzzy" to mean Unicode's
99870f4d
KW
196# loose matchings rules (from Unicode TR18):
197#
198# The recommended names for UCD properties and property values are in
199# PropertyAliases.txt [Prop] and PropertyValueAliases.txt
200# [PropValue]. There are both abbreviated names and longer, more
201# descriptive names. It is strongly recommended that both names be
202# recognized, and that loose matching of property names be used,
203# whereby the case distinctions, whitespace, hyphens, and underbar
204# are ignored.
205# The program still allows Fuzzy to override its determination of if loose
206# matching should be used, but it isn't currently used, as it is no longer
207# needed; the calculations it makes are good enough.
678f13d5 208#
99870f4d
KW
209# SUMMARY OF HOW IT WORKS:
210#
211# Process arguments
212#
213# A list is constructed containing each input file that is to be processed
214#
215# Each file on the list is processed in a loop, using the associated handler
216# code for each:
217# The PropertyAliases.txt and PropValueAliases.txt files are processed
218# first. These files name the properties and property values.
219# Objects are created of all the property and property value names
220# that the rest of the input should expect, including all synonyms.
221# The other input files give mappings from properties to property
222# values. That is, they list code points and say what the mapping
223# is under the given property. Some files give the mappings for
224# just one property; and some for many. This program goes through
225# each file and populates the properties from them. Some properties
226# are listed in more than one file, and Unicode has set up a
227# precedence as to which has priority if there is a conflict. Thus
228# the order of processing matters, and this program handles the
229# conflict possibility by processing the overriding input files
230# last, so that if necessary they replace earlier values.
231# After this is all done, the program creates the property mappings not
232# furnished by Unicode, but derivable from what it does give.
233# The tables of code points that match each property value in each
234# property that is accessible by regular expressions are created.
235# The Perl-defined properties are created and populated. Many of these
236# require data determined from the earlier steps
237# Any Perl-defined synonyms are created, and name clashes between Perl
678f13d5 238# and Unicode are reconciled and warned about.
99870f4d
KW
239# All the properties are written to files
240# Any other files are written, and final warnings issued.
678f13d5 241#
99870f4d
KW
242# For clarity, a number of operators have been overloaded to work on tables:
243# ~ means invert (take all characters not in the set). The more
244# conventional '!' is not used because of the possibility of confusing
245# it with the actual boolean operation.
246# + means union
247# - means subtraction
248# & means intersection
249# The precedence of these is the order listed. Parentheses should be
250# copiously used. These are not a general scheme. The operations aren't
251# defined for a number of things, deliberately, to avoid getting into trouble.
252# Operations are done on references and affect the underlying structures, so
253# that the copy constructors for them have been overloaded to not return a new
254# clone, but the input object itself.
678f13d5 255#
99870f4d
KW
256# The bool operator is deliberately not overloaded to avoid confusion with
257# "should it mean if the object merely exists, or also is non-empty?".
99870f4d
KW
258#
259# WHY CERTAIN DESIGN DECISIONS WERE MADE
678f13d5
KW
260#
261# This program needs to be able to run under miniperl. Therefore, it uses a
262# minimum of other modules, and hence implements some things itself that could
263# be gotten from CPAN
264#
265# This program uses inputs published by the Unicode Consortium. These can
266# change incompatibly between releases without the Perl maintainers realizing
267# it. Therefore this program is now designed to try to flag these. It looks
268# at the directories where the inputs are, and flags any unrecognized files.
269# It keeps track of all the properties in the files it handles, and flags any
270# that it doesn't know how to handle. It also flags any input lines that
271# don't match the expected syntax, among other checks.
272#
273# It is also designed so if a new input file matches one of the known
274# templates, one hopefully just needs to add it to a list to have it
275# processed.
276#
277# As mentioned earlier, some properties are given in more than one file. In
278# particular, the files in the extracted directory are supposedly just
279# reformattings of the others. But they contain information not easily
280# derivable from the other files, including results for Unihan, which this
281# program doesn't ordinarily look at, and for unassigned code points. They
282# also have historically had errors or been incomplete. In an attempt to
283# create the best possible data, this program thus processes them first to
284# glean information missing from the other files; then processes those other
285# files to override any errors in the extracted ones. Much of the design was
286# driven by this need to store things and then possibly override them.
287#
288# It tries to keep fatal errors to a minimum, to generate something usable for
289# testing purposes. It always looks for files that could be inputs, and will
290# warn about any that it doesn't know how to handle (the -q option suppresses
291# the warning).
99870f4d
KW
292#
293# Why have files written out for binary 'N' matches?
294# For binary properties, if you know the mapping for either Y or N; the
678f13d5
KW
295# other is trivial to construct, so could be done at Perl run-time by just
296# complementing the result, instead of having a file for it. That is, if
297# someone types in \p{foo: N}, Perl could translate that to \P{foo: Y} and
298# not need a file. The problem is communicating to Perl that a given
299# property is binary. Perl can't figure it out from looking at the N (or
300# No), as some non-binary properties have these as property values. So
301# rather than inventing a way to communicate this info back to the core,
302# which would have required changes there as well, it was simpler just to
303# add the extra tables.
304#
305# Why is there more than one type of range?
306# This simplified things. There are some very specialized code points that
307# have to be handled specially for output, such as Hangul syllable names.
308# By creating a range type (done late in the development process), it
309# allowed this to be stored with the range, and overridden by other input.
310# Originally these were stored in another data structure, and it became a
311# mess trying to decide if a second file that was for the same property was
312# overriding the earlier one or not.
313#
314# Why are there two kinds of tables, match and map?
315# (And there is a base class shared by the two as well.) As stated above,
316# they actually are for different things. Development proceeded much more
317# smoothly when I (khw) realized the distinction. Map tables are used to
318# give the property value for every code point (actually every code point
319# that doesn't map to a default value). Match tables are used for regular
320# expression matches, and are essentially the inverse mapping. Separating
321# the two allows more specialized methods, and error checks so that one
322# can't just take the intersection of two map tables, for example, as that
323# is nonsensical.
99870f4d
KW
324#
325# There are no match tables generated for matches of the null string. These
c1739a4a 326# would look like qr/\p{JSN=}/ currently without modifying the regex code.
678f13d5
KW
327# Perhaps something like them could be added if necessary. The JSN does have
328# a real code point U+110B that maps to the null string, but it is a
329# contributory property, and therefore not output by default. And it's easily
330# handled so far by making the null string the default where it is a
331# possibility.
99870f4d 332#
23e33b60
KW
333# DEBUGGING
334#
678f13d5
KW
335# This program is written so it will run under miniperl. Occasionally changes
336# will cause an error where the backtrace doesn't work well under miniperl.
337# To diagnose the problem, you can instead run it under regular perl, if you
338# have one compiled.
339#
340# There is a good trace facility. To enable it, first sub DEBUG must be set
341# to return true. Then a line like
342#
343# local $to_trace = 1 if main::DEBUG;
344#
345# can be added to enable tracing in its lexical scope or until you insert
346# another line:
347#
348# local $to_trace = 0 if main::DEBUG;
349#
350# then use a line like "trace $a, @b, %c, ...;
351#
352# Some of the more complex subroutines already have trace statements in them.
353# Permanent trace statements should be like:
354#
355# trace ... if main::DEBUG && $to_trace;
356#
357# If there is just one or a few files that you're debugging, you can easily
358# cause most everything else to be skipped. Change the line
359#
360# my $debug_skip = 0;
361#
362# to 1, and every file whose object is in @input_file_objects and doesn't have
363# a, 'non_skip => 1,' in its constructor will be skipped.
364#
b4a0206c 365# To compare the output tables, it may be useful to specify the -annotate
c4019d52
KW
366# flag. This causes the tables to expand so there is one entry for each
367# non-algorithmically named code point giving, currently its name, and its
368# graphic representation if printable (and you have a font that knows about
369# it). This makes it easier to see what the particular code points are in
370# each output table. The tables are usable, but because they don't have
371# ranges (for the most part), a Perl using them will run slower. Non-named
372# code points are annotated with a description of their status, and contiguous
373# ones with the same description will be output as a range rather than
374# individually. Algorithmically named characters are also output as ranges,
375# except when there are just a few contiguous ones.
376#
99870f4d
KW
377# FUTURE ISSUES
378#
379# The program would break if Unicode were to change its names so that
380# interior white space, underscores, or dashes differences were significant
381# within property and property value names.
382#
383# It might be easier to use the xml versions of the UCD if this program ever
384# would need heavy revision, and the ability to handle old versions was not
385# required.
386#
387# There is the potential for name collisions, in that Perl has chosen names
388# that Unicode could decide it also likes. There have been such collisions in
389# the past, with mostly Perl deciding to adopt the Unicode definition of the
390# name. However in the 5.2 Unicode beta testing, there were a number of such
391# collisions, which were withdrawn before the final release, because of Perl's
392# and other's protests. These all involved new properties which began with
393# 'Is'. Based on the protests, Unicode is unlikely to try that again. Also,
394# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
395# Unicode document, so they are unlikely to be used by Unicode for another
396# purpose. However, they might try something beginning with 'In', or use any
397# of the other Perl-defined properties. This program will warn you of name
398# collisions, and refuse to generate tables with them, but manual intervention
399# will be required in this event. One scheme that could be implemented, if
400# necessary, would be to have this program generate another file, or add a
401# field to mktables.lst that gives the date of first definition of a property.
402# Each new release of Unicode would use that file as a basis for the next
403# iteration. And the Perl synonym addition code could sort based on the age
404# of the property, so older properties get priority, and newer ones that clash
405# would be refused; hence existing code would not be impacted, and some other
406# synonym would have to be used for the new property. This is ugly, and
407# manual intervention would certainly be easier to do in the short run; lets
408# hope it never comes to this.
678f13d5 409#
99870f4d
KW
410# A NOTE ON UNIHAN
411#
412# This program can generate tables from the Unihan database. But it doesn't
413# by default, letting the CPAN module Unicode::Unihan handle them. Prior to
414# version 5.2, this database was in a single file, Unihan.txt. In 5.2 the
415# database was split into 8 different files, all beginning with the letters
416# 'Unihan'. This program will read those file(s) if present, but it needs to
417# know which of the many properties in the file(s) should have tables created
418# for them. It will create tables for any properties listed in
419# PropertyAliases.txt and PropValueAliases.txt, plus any listed in the
420# @cjk_properties array and the @cjk_property_values array. Thus, if a
421# property you want is not in those files of the release you are building
422# against, you must add it to those two arrays. Starting in 4.0, the
423# Unicode_Radical_Stroke was listed in those files, so if the Unihan database
424# is present in the directory, a table will be generated for that property.
425# In 5.2, several more properties were added. For your convenience, the two
5f7264c7 426# arrays are initialized with all the 6.0 listed properties that are also in
99870f4d
KW
427# earlier releases. But these are commented out. You can just uncomment the
428# ones you want, or use them as a template for adding entries for other
429# properties.
430#
431# You may need to adjust the entries to suit your purposes. setup_unihan(),
432# and filter_unihan_line() are the functions where this is done. This program
433# already does some adjusting to make the lines look more like the rest of the
434# Unicode DB; You can see what that is in filter_unihan_line()
435#
436# There is a bug in the 3.2 data file in which some values for the
437# kPrimaryNumeric property have commas and an unexpected comment. A filter
438# could be added for these; or for a particular installation, the Unihan.txt
439# file could be edited to fix them.
99870f4d 440#
678f13d5
KW
441# HOW TO ADD A FILE TO BE PROCESSED
442#
443# A new file from Unicode needs to have an object constructed for it in
444# @input_file_objects, probably at the end or at the end of the extracted
445# ones. The program should warn you if its name will clash with others on
446# restrictive file systems, like DOS. If so, figure out a better name, and
447# add lines to the README.perl file giving that. If the file is a character
448# property, it should be in the format that Unicode has by default
449# standardized for such files for the more recently introduced ones.
450# If so, the Input_file constructor for @input_file_objects can just be the
451# file name and release it first appeared in. If not, then it should be
452# possible to construct an each_line_handler() to massage the line into the
453# standardized form.
454#
455# For non-character properties, more code will be needed. You can look at
456# the existing entries for clues.
457#
458# UNICODE VERSIONS NOTES
459#
460# The Unicode UCD has had a number of errors in it over the versions. And
461# these remain, by policy, in the standard for that version. Therefore it is
462# risky to correct them, because code may be expecting the error. So this
463# program doesn't generally make changes, unless the error breaks the Perl
464# core. As an example, some versions of 2.1.x Jamo.txt have the wrong value
465# for U+1105, which causes real problems for the algorithms for Jamo
466# calculations, so it is changed here.
467#
468# But it isn't so clear cut as to what to do about concepts that are
469# introduced in a later release; should they extend back to earlier releases
470# where the concept just didn't exist? It was easier to do this than to not,
471# so that's what was done. For example, the default value for code points not
472# in the files for various properties was probably undefined until changed by
473# some version. No_Block for blocks is such an example. This program will
474# assign No_Block even in Unicode versions that didn't have it. This has the
475# benefit that code being written doesn't have to special case earlier
476# versions; and the detriment that it doesn't match the Standard precisely for
477# the affected versions.
478#
479# Here are some observations about some of the issues in early versions:
480#
6426c51b 481# The number of code points in \p{alpha} halved in 2.1.9. It turns out that
678f13d5
KW
482# the reason is that the CJK block starting at 4E00 was removed from PropList,
483# and was not put back in until 3.1.0
484#
485# Unicode introduced the synonym Space for White_Space in 4.1. Perl has
486# always had a \p{Space}. In release 3.2 only, they are not synonymous. The
487# reason is that 3.2 introduced U+205F=medium math space, which was not
488# classed as white space, but Perl figured out that it should have been. 4.0
489# reclassified it correctly.
490#
491# Another change between 3.2 and 4.0 is the CCC property value ATBL. In 3.2
492# this was erroneously a synonym for 202. In 4.0, ATB became 202, and ATBL
493# was left with no code points, as all the ones that mapped to 202 stayed
494# mapped to 202. Thus if your program used the numeric name for the class,
495# it would not have been affected, but if it used the mnemonic, it would have
496# been.
497#
498# \p{Script=Hrkt} (Katakana_Or_Hiragana) came in 4.0.1. Before that code
499# points which eventually came to have this script property value, instead
500# mapped to "Unknown". But in the next release all these code points were
501# moved to \p{sc=common} instead.
99870f4d
KW
502#
503# The default for missing code points for BidiClass is complicated. Starting
504# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
505# tries to do the best it can for earlier releases. It is done in
506# process_PropertyAliases()
507#
508##############################################################################
509
510my $UNDEF = ':UNDEF:'; # String to print out for undefined values in tracing
511 # and errors
512my $MAX_LINE_WIDTH = 78;
513
514# Debugging aid to skip most files so as to not be distracted by them when
515# concentrating on the ones being debugged. Add
516# non_skip => 1,
517# to the constructor for those files you want processed when you set this.
518# Files with a first version number of 0 are special: they are always
519# processed regardless of the state of this flag.
520my $debug_skip = 0;
521
522# Set to 1 to enable tracing.
523our $to_trace = 0;
524
525{ # Closure for trace: debugging aid
526 my $print_caller = 1; # ? Include calling subroutine name
527 my $main_with_colon = 'main::';
528 my $main_colon_length = length($main_with_colon);
529
530 sub trace {
531 return unless $to_trace; # Do nothing if global flag not set
532
533 my @input = @_;
534
535 local $DB::trace = 0;
536 $DB::trace = 0; # Quiet 'used only once' message
537
538 my $line_number;
539
540 # Loop looking up the stack to get the first non-trace caller
541 my $caller_line;
542 my $caller_name;
543 my $i = 0;
544 do {
545 $line_number = $caller_line;
546 (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
547 $caller = $main_with_colon unless defined $caller;
548
549 $caller_name = $caller;
550
551 # get rid of pkg
552 $caller_name =~ s/.*:://;
553 if (substr($caller_name, 0, $main_colon_length)
554 eq $main_with_colon)
555 {
556 $caller_name = substr($caller_name, $main_colon_length);
557 }
558
559 } until ($caller_name ne 'trace');
560
561 # If the stack was empty, we were called from the top level
562 $caller_name = 'main' if ($caller_name eq ""
563 || $caller_name eq 'trace');
564
565 my $output = "";
566 foreach my $string (@input) {
567 #print STDERR __LINE__, ": ", join ", ", @input, "\n";
568 if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
569 $output .= simple_dumper($string);
570 }
571 else {
572 $string = "$string" if ref $string;
573 $string = $UNDEF unless defined $string;
574 chomp $string;
575 $string = '""' if $string eq "";
576 $output .= " " if $output ne ""
577 && $string ne ""
578 && substr($output, -1, 1) ne " "
579 && substr($string, 0, 1) ne " ";
580 $output .= $string;
581 }
582 }
583
99f78760
KW
584 print STDERR sprintf "%4d: ", $line_number if defined $line_number;
585 print STDERR "$caller_name: " if $print_caller;
99870f4d
KW
586 print STDERR $output, "\n";
587 return;
588 }
589}
590
591# This is for a rarely used development feature that allows you to compare two
592# versions of the Unicode standard without having to deal with changes caused
1c2e8cca 593# by the code points introduced in the later version. Change the 0 to a SINGLE
99870f4d
KW
594# dotted Unicode release number (e.g. 2.1). Only code points introduced in
595# that release and earlier will be used; later ones are thrown away. You use
596# the version number of the earliest one you want to compare; then run this
597# program on directory structures containing each release, and compare the
598# outputs. These outputs will therefore include only the code points common
599# to both releases, and you can see the changes caused just by the underlying
600# release semantic changes. For versions earlier than 3.2, you must copy a
601# version of DAge.txt into the directory.
602my $string_compare_versions = DEBUG && 0; # e.g., v2.1;
603my $compare_versions = DEBUG
604 && $string_compare_versions
605 && pack "C*", split /\./, $string_compare_versions;
606
607sub uniques {
608 # Returns non-duplicated input values. From "Perl Best Practices:
609 # Encapsulated Cleverness". p. 455 in first edition.
610
611 my %seen;
0e407844
NC
612 # Arguably this breaks encapsulation, if the goal is to permit multiple
613 # distinct objects to stringify to the same value, and be interchangeable.
614 # However, for this program, no two objects stringify identically, and all
615 # lists passed to this function are either objects or strings. So this
616 # doesn't affect correctness, but it does give a couple of percent speedup.
617 no overloading;
99870f4d
KW
618 return grep { ! $seen{$_}++ } @_;
619}
620
621$0 = File::Spec->canonpath($0);
622
623my $make_test_script = 0; # ? Should we output a test script
624my $write_unchanged_files = 0; # ? Should we update the output files even if
625 # we don't think they have changed
626my $use_directory = ""; # ? Should we chdir somewhere.
627my $pod_directory; # input directory to store the pod file.
628my $pod_file = 'perluniprops';
629my $t_path; # Path to the .t test file
630my $file_list = 'mktables.lst'; # File to store input and output file names.
631 # This is used to speed up the build, by not
632 # executing the main body of the program if
633 # nothing on the list has changed since the
634 # previous build
635my $make_list = 1; # ? Should we write $file_list. Set to always
636 # make a list so that when the pumpking is
637 # preparing a release, s/he won't have to do
638 # special things
639my $glob_list = 0; # ? Should we try to include unknown .txt files
640 # in the input.
bd9ebcfd
KW
641my $output_range_counts = $debugging_build; # ? Should we include the number
642 # of code points in ranges in
643 # the output
558712cf 644my $annotate = 0; # ? Should character names be in the output
9ef2b94f 645
99870f4d
KW
646# Verbosity levels; 0 is quiet
647my $NORMAL_VERBOSITY = 1;
648my $PROGRESS = 2;
649my $VERBOSE = 3;
650
651my $verbosity = $NORMAL_VERBOSITY;
652
653# Process arguments
654while (@ARGV) {
cf25bb62
JH
655 my $arg = shift @ARGV;
656 if ($arg eq '-v') {
99870f4d
KW
657 $verbosity = $VERBOSE;
658 }
659 elsif ($arg eq '-p') {
660 $verbosity = $PROGRESS;
661 $| = 1; # Flush buffers as we go.
662 }
663 elsif ($arg eq '-q') {
664 $verbosity = 0;
665 }
666 elsif ($arg eq '-w') {
667 $write_unchanged_files = 1; # update the files even if havent changed
668 }
669 elsif ($arg eq '-check') {
6ae7e459
YO
670 my $this = shift @ARGV;
671 my $ok = shift @ARGV;
672 if ($this ne $ok) {
673 print "Skipping as check params are not the same.\n";
674 exit(0);
675 }
00a8df5c 676 }
99870f4d
KW
677 elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
678 -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
679 }
3df51b85
KW
680 elsif ($arg eq '-maketest' || ($arg eq '-T' && defined ($t_path = shift)))
681 {
99870f4d 682 $make_test_script = 1;
99870f4d
KW
683 }
684 elsif ($arg eq '-makelist') {
685 $make_list = 1;
686 }
687 elsif ($arg eq '-C' && defined ($use_directory = shift)) {
688 -d $use_directory or croak "Unknown directory '$use_directory'";
689 }
690 elsif ($arg eq '-L') {
691
692 # Existence not tested until have chdir'd
693 $file_list = shift;
694 }
695 elsif ($arg eq '-globlist') {
696 $glob_list = 1;
697 }
698 elsif ($arg eq '-c') {
699 $output_range_counts = ! $output_range_counts
700 }
b4a0206c 701 elsif ($arg eq '-annotate') {
558712cf 702 $annotate = 1;
bd9ebcfd
KW
703 $debugging_build = 1;
704 $output_range_counts = 1;
9ef2b94f 705 }
99870f4d
KW
706 else {
707 my $with_c = 'with';
708 $with_c .= 'out' if $output_range_counts; # Complements the state
709 croak <<END;
710usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
711 [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
712 [-check A B ]
713 -c : Output comments $with_c number of code points in ranges
714 -q : Quiet Mode: Only output serious warnings.
715 -p : Set verbosity level to normal plus show progress.
716 -v : Set Verbosity level high: Show progress and non-serious
717 warnings
718 -w : Write files regardless
719 -C dir : Change to this directory before proceeding. All relative paths
720 except those specified by the -P and -T options will be done
721 with respect to this directory.
722 -P dir : Output $pod_file file to directory 'dir'.
3df51b85 723 -T path : Create a test script as 'path'; overrides -maketest
99870f4d
KW
724 -L filelist : Use alternate 'filelist' instead of standard one
725 -globlist : Take as input all non-Test *.txt files in current and sub
726 directories
3df51b85
KW
727 -maketest : Make test script 'TestProp.pl' in current (or -C directory),
728 overrides -T
99870f4d 729 -makelist : Rewrite the file list $file_list based on current setup
b4a0206c 730 -annotate : Output an annotation for each character in the table files;
c4019d52
KW
731 useful for debugging mktables, looking at diffs; but is slow,
732 memory intensive; resulting tables are usable but slow and
733 very large.
99870f4d
KW
734 -check A B : Executes $0 only if A and B are the same
735END
736 }
737}
738
739# Stores the most-recently changed file. If none have changed, can skip the
740# build
aeab6150 741my $most_recent = (stat $0)[9]; # Do this before the chdir!
99870f4d
KW
742
743# Change directories now, because need to read 'version' early.
744if ($use_directory) {
3df51b85 745 if ($pod_directory && ! File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d
KW
746 $pod_directory = File::Spec->rel2abs($pod_directory);
747 }
3df51b85 748 if ($t_path && ! File::Spec->file_name_is_absolute($t_path)) {
99870f4d 749 $t_path = File::Spec->rel2abs($t_path);
00a8df5c 750 }
99870f4d 751 chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
3df51b85 752 if ($pod_directory && File::Spec->file_name_is_absolute($pod_directory)) {
99870f4d 753 $pod_directory = File::Spec->abs2rel($pod_directory);
02b1aeec 754 }
3df51b85 755 if ($t_path && File::Spec->file_name_is_absolute($t_path)) {
99870f4d 756 $t_path = File::Spec->abs2rel($t_path);
02b1aeec 757 }
00a8df5c
YO
758}
759
99870f4d
KW
760# Get Unicode version into regular and v-string. This is done now because
761# various tables below get populated based on it. These tables are populated
762# here to be near the top of the file, and so easily seeable by those needing
763# to modify things.
764open my $VERSION, "<", "version"
765 or croak "$0: can't open required file 'version': $!\n";
766my $string_version = <$VERSION>;
767close $VERSION;
768chomp $string_version;
769my $v_version = pack "C*", split /\./, $string_version; # v string
770
771# The following are the complete names of properties with property values that
772# are known to not match any code points in some versions of Unicode, but that
773# may change in the future so they should be matchable, hence an empty file is
774# generated for them.
775my @tables_that_may_be_empty = (
776 'Joining_Type=Left_Joining',
777 );
778push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
779push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
780push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
781 if $v_version ge v4.1.0;
82aed44a
KW
782push @tables_that_may_be_empty, 'Script_Extensions=Katakana_Or_Hiragana'
783 if $v_version ge v6.0.0;
99870f4d
KW
784
785# The lists below are hashes, so the key is the item in the list, and the
786# value is the reason why it is in the list. This makes generation of
787# documentation easier.
788
789my %why_suppressed; # No file generated for these.
790
791# Files aren't generated for empty extraneous properties. This is arguable.
792# Extraneous properties generally come about because a property is no longer
793# used in a newer version of Unicode. If we generated a file without code
794# points, programs that used to work on that property will still execute
795# without errors. It just won't ever match (or will always match, with \P{}).
796# This means that the logic is now likely wrong. I (khw) think its better to
797# find this out by getting an error message. Just move them to the table
798# above to change this behavior
799my %why_suppress_if_empty_warn_if_not = (
800
801 # It is the only property that has ever officially been removed from the
802 # Standard. The database never contained any code points for it.
803 'Special_Case_Condition' => 'Obsolete',
804
805 # Apparently never official, but there were code points in some versions of
806 # old-style PropList.txt
807 'Non_Break' => 'Obsolete',
808);
809
810# These would normally go in the warn table just above, but they were changed
811# a long time before this program was written, so warnings about them are
812# moot.
813if ($v_version gt v3.2.0) {
814 push @tables_that_may_be_empty,
815 'Canonical_Combining_Class=Attached_Below_Left'
816}
817
5f7264c7 818# These are listed in the Property aliases file in 6.0, but Unihan is ignored
99870f4d
KW
819# unless explicitly added.
820if ($v_version ge v5.2.0) {
821 my $unihan = 'Unihan; remove from list if using Unihan';
ea25a9b2 822 foreach my $table (qw (
99870f4d
KW
823 kAccountingNumeric
824 kOtherNumeric
825 kPrimaryNumeric
826 kCompatibilityVariant
827 kIICore
828 kIRG_GSource
829 kIRG_HSource
830 kIRG_JSource
831 kIRG_KPSource
832 kIRG_MSource
833 kIRG_KSource
834 kIRG_TSource
835 kIRG_USource
836 kIRG_VSource
837 kRSUnicode
ea25a9b2 838 ))
99870f4d
KW
839 {
840 $why_suppress_if_empty_warn_if_not{$table} = $unihan;
841 }
ca12659b
NC
842}
843
272501f6
KW
844# Enum values for to_output_map() method in the Map_Table package.
845my $EXTERNAL_MAP = 1;
846my $INTERNAL_MAP = 2;
847
fcf1973c
KW
848# To override computed values for writing the map tables for these properties.
849# The default for enum map tables is to write them out, so that the Unicode
850# .txt files can be removed, but all the data to compute any property value
851# for any code point is available in a more compact form.
852my %global_to_output_map = (
853 # Needed by UCD.pm, but don't want to publicize that it exists, so won't
854 # get stuck supporting it if things change. Sinc it is a STRING property,
855 # it normally would be listed in the pod, but INTERNAL_MAP suppresses
856 # that.
857 Unicode_1_Name => $INTERNAL_MAP,
858
859 Present_In => 0, # Suppress, as easily computed from Age
fcf1973c
KW
860 Block => 0, # Suppress, as Blocks.txt is retained.
861);
862
99870f4d
KW
863# Properties that this program ignores.
864my @unimplemented_properties = (
865'Unicode_Radical_Stroke' # Remove if changing to handle this one.
866);
d73e5302 867
99870f4d
KW
868# There are several types of obsolete properties defined by Unicode. These
869# must be hand-edited for every new Unicode release.
870my %why_deprecated; # Generates a deprecated warning message if used.
871my %why_stabilized; # Documentation only
872my %why_obsolete; # Documentation only
873
874{ # Closure
875 my $simple = 'Perl uses the more complete version of this property';
876 my $unihan = 'Unihan properties are by default not enabled in the Perl core. Instead use CPAN: Unicode::Unihan';
877
878 my $other_properties = 'other properties';
879 my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
5d294d41 880 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
881
882 %why_deprecated = (
5f7264c7 883 'Grapheme_Link' => 'Deprecated by Unicode: Duplicates ccc=vr (Canonical_Combining_Class=Virama)',
99870f4d
KW
884 'Jamo_Short_Name' => $contributory,
885 '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',
886 'Other_Alphabetic' => $contributory,
887 'Other_Default_Ignorable_Code_Point' => $contributory,
888 'Other_Grapheme_Extend' => $contributory,
889 'Other_ID_Continue' => $contributory,
890 'Other_ID_Start' => $contributory,
891 'Other_Lowercase' => $contributory,
892 'Other_Math' => $contributory,
893 'Other_Uppercase' => $contributory,
894 );
895
896 %why_suppressed = (
5f7264c7 897 # There is a lib/unicore/Decomposition.pl (used by Normalize.pm) which
99870f4d
KW
898 # contains the same information, but without the algorithmically
899 # determinable Hangul syllables'. This file is not published, so it's
900 # existence is not noted in the comment.
901 'Decomposition_Mapping' => 'Accessible via Unicode::Normalize',
902
903 'ISO_Comment' => 'Apparently no demand for it, but can access it through Unicode::UCD::charinfo. Obsoleted, and code points for it removed in Unicode 5.2',
99870f4d
KW
904
905 'Simple_Case_Folding' => "$simple. Can access this through Unicode::UCD::casefold",
906 'Simple_Lowercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
907 'Simple_Titlecase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
908 'Simple_Uppercase_Mapping' => "$simple. Can access this through Unicode::UCD::charinfo",
909
910 'Name' => "Accessible via 'use charnames;'",
911 'Name_Alias' => "Accessible via 'use charnames;'",
912
5f7264c7 913 FC_NFKC_Closure => 'Supplanted in usage by NFKC_Casefold; otherwise not useful',
99870f4d
KW
914 Expands_On_NFC => $why_no_expand,
915 Expands_On_NFD => $why_no_expand,
916 Expands_On_NFKC => $why_no_expand,
917 Expands_On_NFKD => $why_no_expand,
918 );
919
920 # The following are suppressed because they were made contributory or
921 # deprecated by Unicode before Perl ever thought about supporting them.
922 foreach my $property ('Jamo_Short_Name', 'Grapheme_Link') {
923 $why_suppressed{$property} = $why_deprecated{$property};
924 }
cf25bb62 925
99870f4d
KW
926 # Customize the message for all the 'Other_' properties
927 foreach my $property (keys %why_deprecated) {
928 next if (my $main_property = $property) !~ s/^Other_//;
929 $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
930 }
931}
932
933if ($v_version ge 4.0.0) {
934 $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
5f7264c7
KW
935 if ($v_version ge 6.0.0) {
936 $why_deprecated{'Hyphen'} = 'Supplanted by Line_Break property values; see www.unicode.org/reports/tr14';
937 }
99870f4d 938}
5f7264c7 939if ($v_version ge 5.2.0 && $v_version lt 6.0.0) {
99870f4d 940 $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
5f7264c7
KW
941 if ($v_version ge 6.0.0) {
942 $why_deprecated{'ISO_Comment'} = 'No longer needed for chart generation; otherwise not useful, and code points for it have been removed';
943 }
99870f4d
KW
944}
945
946# Probably obsolete forever
947if ($v_version ge v4.1.0) {
82aed44a
KW
948 $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete. All code points previously matched by this have been moved to "Script=Common".';
949}
950if ($v_version ge v6.0.0) {
2b352efd
KW
951 $why_suppressed{'Script=Katakana_Or_Hiragana'} .= ' Consider instead using "Script_Extensions=Katakana" or "Script_Extensions=Hiragana (or both)"';
952 $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
953}
954
955# This program can create files for enumerated-like properties, such as
956# 'Numeric_Type'. This file would be the same format as for a string
957# property, with a mapping from code point to its value, so you could look up,
958# for example, the script a code point is in. But no one so far wants this
959# mapping, or they have found another way to get it since this is a new
960# feature. So no file is generated except if it is in this list.
961my @output_mapped_properties = split "\n", <<END;
962END
963
964# If you are using the Unihan database, you need to add the properties that
965# you want to extract from it to this table. For your convenience, the
5f7264c7 966# properties in the 6.0 PropertyAliases.txt file are listed, commented out
99870f4d
KW
967my @cjk_properties = split "\n", <<'END';
968#cjkAccountingNumeric; kAccountingNumeric
969#cjkOtherNumeric; kOtherNumeric
970#cjkPrimaryNumeric; kPrimaryNumeric
971#cjkCompatibilityVariant; kCompatibilityVariant
972#cjkIICore ; kIICore
973#cjkIRG_GSource; kIRG_GSource
974#cjkIRG_HSource; kIRG_HSource
975#cjkIRG_JSource; kIRG_JSource
976#cjkIRG_KPSource; kIRG_KPSource
977#cjkIRG_KSource; kIRG_KSource
978#cjkIRG_TSource; kIRG_TSource
979#cjkIRG_USource; kIRG_USource
980#cjkIRG_VSource; kIRG_VSource
981#cjkRSUnicode; kRSUnicode ; Unicode_Radical_Stroke; URS
982END
983
984# Similarly for the property values. For your convenience, the lines in the
5f7264c7 985# 6.0 PropertyAliases.txt file are listed. Just remove the first BUT NOT both
99870f4d
KW
986# '#' marks
987my @cjk_property_values = split "\n", <<'END';
988## @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
989## @missing: 0000..10FFFF; cjkCompatibilityVariant; <code point>
990## @missing: 0000..10FFFF; cjkIICore; <none>
991## @missing: 0000..10FFFF; cjkIRG_GSource; <none>
992## @missing: 0000..10FFFF; cjkIRG_HSource; <none>
993## @missing: 0000..10FFFF; cjkIRG_JSource; <none>
994## @missing: 0000..10FFFF; cjkIRG_KPSource; <none>
995## @missing: 0000..10FFFF; cjkIRG_KSource; <none>
996## @missing: 0000..10FFFF; cjkIRG_TSource; <none>
997## @missing: 0000..10FFFF; cjkIRG_USource; <none>
998## @missing: 0000..10FFFF; cjkIRG_VSource; <none>
999## @missing: 0000..10FFFF; cjkOtherNumeric; NaN
1000## @missing: 0000..10FFFF; cjkPrimaryNumeric; NaN
1001## @missing: 0000..10FFFF; cjkRSUnicode; <none>
1002END
1003
1004# The input files don't list every code point. Those not listed are to be
1005# defaulted to some value. Below are hard-coded what those values are for
1006# non-binary properties as of 5.1. Starting in 5.0, there are
1007# machine-parsable comment lines in the files the give the defaults; so this
1008# list shouldn't have to be extended. The claim is that all missing entries
1009# for binary properties will default to 'N'. Unicode tried to change that in
1010# 5.2, but the beta period produced enough protest that they backed off.
1011#
1012# The defaults for the fields that appear in UnicodeData.txt in this hash must
1013# be in the form that it expects. The others may be synonyms.
1014my $CODE_POINT = '<code point>';
1015my %default_mapping = (
1016 Age => "Unassigned",
1017 # Bidi_Class => Complicated; set in code
1018 Bidi_Mirroring_Glyph => "",
1019 Block => 'No_Block',
1020 Canonical_Combining_Class => 0,
1021 Case_Folding => $CODE_POINT,
1022 Decomposition_Mapping => $CODE_POINT,
1023 Decomposition_Type => 'None',
1024 East_Asian_Width => "Neutral",
1025 FC_NFKC_Closure => $CODE_POINT,
1026 General_Category => 'Cn',
1027 Grapheme_Cluster_Break => 'Other',
1028 Hangul_Syllable_Type => 'NA',
1029 ISO_Comment => "",
1030 Jamo_Short_Name => "",
1031 Joining_Group => "No_Joining_Group",
1032 # Joining_Type => Complicated; set in code
1033 kIICore => 'N', # Is converted to binary
1034 #Line_Break => Complicated; set in code
1035 Lowercase_Mapping => $CODE_POINT,
1036 Name => "",
1037 Name_Alias => "",
1038 NFC_QC => 'Yes',
1039 NFD_QC => 'Yes',
1040 NFKC_QC => 'Yes',
1041 NFKD_QC => 'Yes',
1042 Numeric_Type => 'None',
1043 Numeric_Value => 'NaN',
1044 Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
1045 Sentence_Break => 'Other',
1046 Simple_Case_Folding => $CODE_POINT,
1047 Simple_Lowercase_Mapping => $CODE_POINT,
1048 Simple_Titlecase_Mapping => $CODE_POINT,
1049 Simple_Uppercase_Mapping => $CODE_POINT,
1050 Titlecase_Mapping => $CODE_POINT,
1051 Unicode_1_Name => "",
1052 Unicode_Radical_Stroke => "",
1053 Uppercase_Mapping => $CODE_POINT,
1054 Word_Break => 'Other',
1055);
1056
1057# Below are files that Unicode furnishes, but this program ignores, and why
1058my %ignored_files = (
1059 'CJKRadicals.txt' => 'Unihan data',
1060 'Index.txt' => 'An index, not actual data',
1061 'NamedSqProv.txt' => 'Not officially part of the Unicode standard; Append it to NamedSequences.txt if you want to process the contents.',
1062 'NamesList.txt' => 'Just adds commentary',
1063 'NormalizationCorrections.txt' => 'Data is already in other files.',
1064 'Props.txt' => 'Adds nothing to PropList.txt; only in very early releases',
1065 'ReadMe.txt' => 'Just comments',
1066 'README.TXT' => 'Just comments',
1067 'StandardizedVariants.txt' => 'Only for glyph changes, not a Unicode character property. Does not fit into current scheme where one code point is mapped',
5f7264c7
KW
1068 'EmojiSources.txt' => 'Not of general utility: for Japanese legacy cell-phone applications',
1069 'IndicMatraCategory.txt' => 'Provisional',
1070 'IndicSyllabicCategory.txt' => 'Provisional',
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
99870f4d
KW
1098my $LAST_UNICODE_CODEPOINT_STRING = "10FFFF";
1099my $LAST_UNICODE_CODEPOINT = hex $LAST_UNICODE_CODEPOINT_STRING;
1100my $MAX_UNICODE_CODEPOINTS = $LAST_UNICODE_CODEPOINT + 1;
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
8c32d378
KW
1104# to decrease backtracking. The first one allows the code point to be at the
1105# end of a word, but to work properly, the word shouldn't end with a valid hex
1106# character. The second one won't match a code point at the end of a word,
1107# and doesn't have the run-on issue
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 $missing_defaults_prefix =
1117 qr/^#\s+\@missing:\s+0000\.\.$LAST_UNICODE_CODEPOINT_STRING\s*;/;
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;
1124my $ENUM = 3; # Include catalog
1125my $STRING = 4; # Anything else: string or misc
1126
1127# Some input files have lines that give default values for code points not
1128# contained in the file. Sometimes these should be ignored.
1129my $NO_DEFAULTS = 0; # Must evaluate to false
1130my $NOT_IGNORED = 1;
1131my $IGNORED = 2;
1132
1133# Range types. Each range has a type. Most ranges are type 0, for normal,
1134# and will appear in the main body of the tables in the output files, but
1135# there are other types of ranges as well, listed below, that are specially
1136# handled. There are pseudo-types as well that will never be stored as a
1137# type, but will affect the calculation of the type.
1138
1139# 0 is for normal, non-specials
1140my $MULTI_CP = 1; # Sequence of more than code point
1141my $HANGUL_SYLLABLE = 2;
1142my $CP_IN_NAME = 3; # The NAME contains the code point appended to it.
1143my $NULL = 4; # The map is to the null string; utf8.c can't
1144 # handle these, nor is there an accepted syntax
1145 # for them in \p{} constructs
f86864ac 1146my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; means that ranges that would
99870f4d
KW
1147 # otherwise be $MULTI_CP type are instead type 0
1148
1149# process_generic_property_file() can accept certain overrides in its input.
1150# Each of these must begin AND end with $CMD_DELIM.
1151my $CMD_DELIM = "\a";
1152my $REPLACE_CMD = 'replace'; # Override the Replace
1153my $MAP_TYPE_CMD = 'map_type'; # Override the Type
1154
1155my $NO = 0;
1156my $YES = 1;
1157
1158# Values for the Replace argument to add_range.
1159# $NO # Don't replace; add only the code points not
1160 # already present.
1161my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
1162 # the comments at the subroutine definition.
1163my $UNCONDITIONALLY = 2; # Replace without conditions.
1164my $MULTIPLE = 4; # Don't replace, but add a duplicate record if
1165 # already there
56343c78 1166my $CROAK = 5; # Die with an error if is already there
99870f4d
KW
1167
1168# Flags to give property statuses. The phrases are to remind maintainers that
1169# if the flag is changed, the indefinite article referring to it in the
1170# documentation may need to be as well.
1171my $NORMAL = "";
1172my $SUPPRESSED = 'z'; # The character should never actually be seen, since
1173 # it is suppressed
37e2e78e 1174my $PLACEHOLDER = 'P'; # Implies no pod entry generated
99870f4d
KW
1175my $DEPRECATED = 'D';
1176my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
1177my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
1178my $DISCOURAGED = 'X';
1179my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
1180my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
1181my $STRICTER = 'T';
1182my $a_bold_stricter = "a 'B<$STRICTER>'";
1183my $A_bold_stricter = "A 'B<$STRICTER>'";
1184my $STABILIZED = 'S';
1185my $a_bold_stabilized = "an 'B<$STABILIZED>'";
1186my $A_bold_stabilized = "An 'B<$STABILIZED>'";
1187my $OBSOLETE = 'O';
1188my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
1189my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
1190
1191my %status_past_participles = (
1192 $DISCOURAGED => 'discouraged',
1193 $SUPPRESSED => 'should never be generated',
1194 $STABILIZED => 'stabilized',
1195 $OBSOLETE => 'obsolete',
37e2e78e 1196 $DEPRECATED => 'deprecated',
99870f4d
KW
1197);
1198
f5817e0a
KW
1199# The format of the values of the tables:
1200my $EMPTY_FORMAT = "";
99870f4d
KW
1201my $BINARY_FORMAT = 'b';
1202my $DECIMAL_FORMAT = 'd';
1203my $FLOAT_FORMAT = 'f';
1204my $INTEGER_FORMAT = 'i';
1205my $HEX_FORMAT = 'x';
1206my $RATIONAL_FORMAT = 'r';
1207my $STRING_FORMAT = 's';
a14f3cb1 1208my $DECOMP_STRING_FORMAT = 'c';
99870f4d
KW
1209
1210my %map_table_formats = (
1211 $BINARY_FORMAT => 'binary',
1212 $DECIMAL_FORMAT => 'single decimal digit',
1213 $FLOAT_FORMAT => 'floating point number',
1214 $INTEGER_FORMAT => 'integer',
add63c13 1215 $HEX_FORMAT => 'non-negative hex whole number; a code point',
99870f4d 1216 $RATIONAL_FORMAT => 'rational: an integer or a fraction',
1a9d544b 1217 $STRING_FORMAT => 'string',
92f9d56c 1218 $DECOMP_STRING_FORMAT => 'Perl\'s internal (Normalize.pm) decomposition mapping',
99870f4d
KW
1219);
1220
1221# Unicode didn't put such derived files in a separate directory at first.
1222my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
1223my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
1224my $AUXILIARY = 'auxiliary';
1225
1226# Hashes that will eventually go into Heavy.pl for the use of utf8_heavy.pl
1227my %loose_to_file_of; # loosely maps table names to their respective
1228 # files
1229my %stricter_to_file_of; # same; but for stricter mapping.
1230my %nv_floating_to_rational; # maps numeric values floating point numbers to
1231 # their rational equivalent
1232my %loose_property_name_of; # Loosely maps property names to standard form
1233
d867ccfb
KW
1234# Most properties are immune to caseless matching, otherwise you would get
1235# nonsensical results, as properties are a function of a code point, not
1236# everything that is caselessly equivalent to that code point. For example,
1237# Changes_When_Case_Folded('s') should be false, whereas caselessly it would
1238# be true because 's' and 'S' are equivalent caselessly. However,
1239# traditionally, [:upper:] and [:lower:] are equivalent caselessly, so we
1240# extend that concept to those very few properties that are like this. Each
1241# such property will match the full range caselessly. They are hard-coded in
1242# the program; it's not worth trying to make it general as it's extremely
1243# unlikely that they will ever change.
1244my %caseless_equivalent_to;
1245
99870f4d
KW
1246# These constants names and values were taken from the Unicode standard,
1247# version 5.1, section 3.12. They are used in conjunction with Hangul
6e5a209b
KW
1248# syllables. The '_string' versions are so generated tables can retain the
1249# hex format, which is the more familiar value
1250my $SBase_string = "0xAC00";
1251my $SBase = CORE::hex $SBase_string;
1252my $LBase_string = "0x1100";
1253my $LBase = CORE::hex $LBase_string;
1254my $VBase_string = "0x1161";
1255my $VBase = CORE::hex $VBase_string;
1256my $TBase_string = "0x11A7";
1257my $TBase = CORE::hex $TBase_string;
99870f4d
KW
1258my $SCount = 11172;
1259my $LCount = 19;
1260my $VCount = 21;
1261my $TCount = 28;
1262my $NCount = $VCount * $TCount;
1263
1264# For Hangul syllables; These store the numbers from Jamo.txt in conjunction
1265# with the above published constants.
1266my %Jamo;
1267my %Jamo_L; # Leading consonants
1268my %Jamo_V; # Vowels
1269my %Jamo_T; # Trailing consonants
1270
37e2e78e 1271my @backslash_X_tests; # List of tests read in for testing \X
99870f4d
KW
1272my @unhandled_properties; # Will contain a list of properties found in
1273 # the input that we didn't process.
f86864ac 1274my @match_properties; # Properties that have match tables, to be
99870f4d
KW
1275 # listed in the pod
1276my @map_properties; # Properties that get map files written
1277my @named_sequences; # NamedSequences.txt contents.
1278my %potential_files; # Generated list of all .txt files in the directory
1279 # structure so we can warn if something is being
1280 # ignored.
1281my @files_actually_output; # List of files we generated.
1282my @more_Names; # Some code point names are compound; this is used
1283 # to store the extra components of them.
1284my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
1285 # the minimum before we consider it equivalent to a
1286 # candidate rational
1287my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
1288
1289# These store references to certain commonly used property objects
1290my $gc;
1291my $perl;
1292my $block;
3e20195b
KW
1293my $perl_charname;
1294my $print;
7fc6cb55 1295my $Any;
99870f4d
KW
1296
1297# Are there conflicting names because of beginning with 'In_', or 'Is_'
1298my $has_In_conflicts = 0;
1299my $has_Is_conflicts = 0;
1300
1301sub internal_file_to_platform ($) {
1302 # Convert our file paths which have '/' separators to those of the
1303 # platform.
1304
1305 my $file = shift;
1306 return undef unless defined $file;
1307
1308 return File::Spec->join(split '/', $file);
d07a55ed 1309}
5beb625e 1310
99870f4d
KW
1311sub file_exists ($) { # platform independent '-e'. This program internally
1312 # uses slash as a path separator.
1313 my $file = shift;
1314 return 0 if ! defined $file;
1315 return -e internal_file_to_platform($file);
1316}
5beb625e 1317
99870f4d 1318sub objaddr($) {
23e33b60
KW
1319 # Returns the address of the blessed input object.
1320 # It doesn't check for blessedness because that would do a string eval
1321 # every call, and the program is structured so that this is never called
1322 # for a non-blessed object.
99870f4d 1323
23e33b60 1324 no overloading; # If overloaded, numifying below won't work.
99870f4d
KW
1325
1326 # Numifying a ref gives its address.
051df77b 1327 return pack 'J', $_[0];
99870f4d
KW
1328}
1329
558712cf 1330# These are used only if $annotate is true.
c4019d52
KW
1331# The entire range of Unicode characters is examined to populate these
1332# after all the input has been processed. But most can be skipped, as they
1333# have the same descriptive phrases, such as being unassigned
1334my @viacode; # Contains the 1 million character names
1335my @printable; # boolean: And are those characters printable?
1336my @annotate_char_type; # Contains a type of those characters, specifically
1337 # for the purposes of annotation.
1338my $annotate_ranges; # A map of ranges of code points that have the same
98dc9551 1339 # name for the purposes of annotation. They map to the
c4019d52
KW
1340 # upper edge of the range, so that the end point can
1341 # be immediately found. This is used to skip ahead to
1342 # the end of a range, and avoid processing each
1343 # individual code point in it.
1344my $unassigned_sans_noncharacters; # A Range_List of the unassigned
1345 # characters, but excluding those which are
1346 # also noncharacter code points
1347
1348# The annotation types are an extension of the regular range types, though
1349# some of the latter are folded into one. Make the new types negative to
1350# avoid conflicting with the regular types
1351my $SURROGATE_TYPE = -1;
1352my $UNASSIGNED_TYPE = -2;
1353my $PRIVATE_USE_TYPE = -3;
1354my $NONCHARACTER_TYPE = -4;
1355my $CONTROL_TYPE = -5;
1356my $UNKNOWN_TYPE = -6; # Used only if there is a bug in this program
1357
1358sub populate_char_info ($) {
558712cf 1359 # Used only with the $annotate option. Populates the arrays with the
c4019d52
KW
1360 # input code point's info that are needed for outputting more detailed
1361 # comments. If calling context wants a return, it is the end point of
1362 # any contiguous range of characters that share essentially the same info
1363
1364 my $i = shift;
1365 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1366
1367 $viacode[$i] = $perl_charname->value_of($i) || "";
1368
1369 # A character is generally printable if Unicode says it is,
1370 # but below we make sure that most Unicode general category 'C' types
1371 # aren't.
1372 $printable[$i] = $print->contains($i);
1373
1374 $annotate_char_type[$i] = $perl_charname->type_of($i) || 0;
1375
1376 # Only these two regular types are treated specially for annotations
1377 # purposes
1378 $annotate_char_type[$i] = 0 if $annotate_char_type[$i] != $CP_IN_NAME
1379 && $annotate_char_type[$i] != $HANGUL_SYLLABLE;
1380
1381 # Give a generic name to all code points that don't have a real name.
1382 # We output ranges, if applicable, for these. Also calculate the end
1383 # point of the range.
1384 my $end;
1385 if (! $viacode[$i]) {
1386 if ($gc-> table('Surrogate')->contains($i)) {
1387 $viacode[$i] = 'Surrogate';
1388 $annotate_char_type[$i] = $SURROGATE_TYPE;
1389 $printable[$i] = 0;
1390 $end = $gc->table('Surrogate')->containing_range($i)->end;
1391 }
1392 elsif ($gc-> table('Private_use')->contains($i)) {
1393 $viacode[$i] = 'Private Use';
1394 $annotate_char_type[$i] = $PRIVATE_USE_TYPE;
1395 $printable[$i] = 0;
1396 $end = $gc->table('Private_Use')->containing_range($i)->end;
1397 }
1398 elsif (Property::property_ref('Noncharacter_Code_Point')-> table('Y')->
1399 contains($i))
1400 {
1401 $viacode[$i] = 'Noncharacter';
1402 $annotate_char_type[$i] = $NONCHARACTER_TYPE;
1403 $printable[$i] = 0;
1404 $end = property_ref('Noncharacter_Code_Point')->table('Y')->
1405 containing_range($i)->end;
1406 }
1407 elsif ($gc-> table('Control')->contains($i)) {
1408 $viacode[$i] = 'Control';
1409 $annotate_char_type[$i] = $CONTROL_TYPE;
1410 $printable[$i] = 0;
1411 $end = 0x81 if $i == 0x80; # Hard-code this one known case
1412 }
1413 elsif ($gc-> table('Unassigned')->contains($i)) {
1414 $viacode[$i] = 'Unassigned, block=' . $block-> value_of($i);
1415 $annotate_char_type[$i] = $UNASSIGNED_TYPE;
1416 $printable[$i] = 0;
1417
1418 # Because we name the unassigned by the blocks they are in, it
1419 # can't go past the end of that block, and it also can't go past
1420 # the unassigned range it is in. The special table makes sure
1421 # that the non-characters, which are unassigned, are separated
1422 # out.
1423 $end = min($block->containing_range($i)->end,
1424 $unassigned_sans_noncharacters-> containing_range($i)->
1425 end);
13ca76ff
KW
1426 }
1427 else {
1428 Carp::my_carp_bug("Can't figure out how to annotate "
1429 . sprintf("U+%04X", $i)
1430 . ". Proceeding anyway.");
c4019d52
KW
1431 $viacode[$i] = 'UNKNOWN';
1432 $annotate_char_type[$i] = $UNKNOWN_TYPE;
1433 $printable[$i] = 0;
1434 }
1435 }
1436
1437 # Here, has a name, but if it's one in which the code point number is
1438 # appended to the name, do that.
1439 elsif ($annotate_char_type[$i] == $CP_IN_NAME) {
1440 $viacode[$i] .= sprintf("-%04X", $i);
1441 $end = $perl_charname->containing_range($i)->end;
1442 }
1443
1444 # And here, has a name, but if it's a hangul syllable one, replace it with
1445 # the correct name from the Unicode algorithm
1446 elsif ($annotate_char_type[$i] == $HANGUL_SYLLABLE) {
1447 use integer;
1448 my $SIndex = $i - $SBase;
1449 my $L = $LBase + $SIndex / $NCount;
1450 my $V = $VBase + ($SIndex % $NCount) / $TCount;
1451 my $T = $TBase + $SIndex % $TCount;
1452 $viacode[$i] = "HANGUL SYLLABLE $Jamo{$L}$Jamo{$V}";
1453 $viacode[$i] .= $Jamo{$T} if $T != $TBase;
1454 $end = $perl_charname->containing_range($i)->end;
1455 }
1456
1457 return if ! defined wantarray;
1458 return $i if ! defined $end; # If not a range, return the input
1459
1460 # Save this whole range so can find the end point quickly
1461 $annotate_ranges->add_map($i, $end, $end);
1462
1463 return $end;
1464}
1465
23e33b60
KW
1466# Commented code below should work on Perl 5.8.
1467## This 'require' doesn't necessarily work in miniperl, and even if it does,
1468## the native perl version of it (which is what would operate under miniperl)
1469## is extremely slow, as it does a string eval every call.
1470#my $has_fast_scalar_util = $\18 !~ /miniperl/
1471# && defined eval "require Scalar::Util";
1472#
1473#sub objaddr($) {
1474# # Returns the address of the blessed input object. Uses the XS version if
1475# # available. It doesn't check for blessedness because that would do a
1476# # string eval every call, and the program is structured so that this is
1477# # never called for a non-blessed object.
1478#
1479# return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
1480#
1481# # Check at least that is a ref.
1482# my $pkg = ref($_[0]) or return undef;
1483#
1484# # Change to a fake package to defeat any overloaded stringify
1485# bless $_[0], 'main::Fake';
1486#
1487# # Numifying a ref gives its address.
051df77b 1488# my $addr = pack 'J', $_[0];
23e33b60
KW
1489#
1490# # Return to original class
1491# bless $_[0], $pkg;
1492# return $addr;
1493#}
1494
99870f4d
KW
1495sub max ($$) {
1496 my $a = shift;
1497 my $b = shift;
1498 return $a if $a >= $b;
1499 return $b;
1500}
1501
1502sub min ($$) {
1503 my $a = shift;
1504 my $b = shift;
1505 return $a if $a <= $b;
1506 return $b;
1507}
1508
1509sub clarify_number ($) {
1510 # This returns the input number with underscores inserted every 3 digits
1511 # in large (5 digits or more) numbers. Input must be entirely digits, not
1512 # checked.
1513
1514 my $number = shift;
1515 my $pos = length($number) - 3;
1516 return $number if $pos <= 1;
1517 while ($pos > 0) {
1518 substr($number, $pos, 0) = '_';
1519 $pos -= 3;
5beb625e 1520 }
99870f4d 1521 return $number;
99598c8c
JH
1522}
1523
12ac2576 1524
99870f4d 1525package Carp;
7ebf06b3 1526
99870f4d
KW
1527# These routines give a uniform treatment of messages in this program. They
1528# are placed in the Carp package to cause the stack trace to not include them,
1529# although an alternative would be to use another package and set @CARP_NOT
1530# for it.
12ac2576 1531
99870f4d 1532our $Verbose = 1 if main::DEBUG; # Useful info when debugging
12ac2576 1533
99f78760
KW
1534# This is a work-around suggested by Nicholas Clark to fix a problem with Carp
1535# and overload trying to load Scalar:Util under miniperl. See
1536# http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2009-11/msg01057.html
1537undef $overload::VERSION;
1538
99870f4d
KW
1539sub my_carp {
1540 my $message = shift || "";
1541 my $nofold = shift || 0;
7ebf06b3 1542
99870f4d
KW
1543 if ($message) {
1544 $message = main::join_lines($message);
1545 $message =~ s/^$0: *//; # Remove initial program name
1546 $message =~ s/[.;,]+$//; # Remove certain ending punctuation
1547 $message = "\n$0: $message;";
12ac2576 1548
99870f4d
KW
1549 # Fold the message with program name, semi-colon end punctuation
1550 # (which looks good with the message that carp appends to it), and a
1551 # hanging indent for continuation lines.
1552 $message = main::simple_fold($message, "", 4) unless $nofold;
1553 $message =~ s/\n$//; # Remove the trailing nl so what carp
1554 # appends is to the same line
1555 }
12ac2576 1556
99870f4d 1557 return $message if defined wantarray; # If a caller just wants the msg
12ac2576 1558
99870f4d
KW
1559 carp $message;
1560 return;
1561}
7ebf06b3 1562
99870f4d
KW
1563sub my_carp_bug {
1564 # This is called when it is clear that the problem is caused by a bug in
1565 # this program.
7ebf06b3 1566
99870f4d
KW
1567 my $message = shift;
1568 $message =~ s/^$0: *//;
1569 $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");
1570 carp $message;
1571 return;
1572}
7ebf06b3 1573
99870f4d
KW
1574sub carp_too_few_args {
1575 if (@_ != 2) {
1576 my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'. No action taken.");
1577 return;
12ac2576 1578 }
7ebf06b3 1579
99870f4d
KW
1580 my $args_ref = shift;
1581 my $count = shift;
7ebf06b3 1582
99870f4d
KW
1583 my_carp_bug("Need at least $count arguments to "
1584 . (caller 1)[3]
1585 . ". Instead got: '"
1586 . join ', ', @$args_ref
1587 . "'. No action taken.");
1588 return;
12ac2576
JP
1589}
1590
99870f4d
KW
1591sub carp_extra_args {
1592 my $args_ref = shift;
1593 my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . "); Extras ignored.") if @_;
12ac2576 1594
99870f4d
KW
1595 unless (ref $args_ref) {
1596 my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref. Not checking arguments.");
1597 return;
1598 }
1599 my ($package, $file, $line) = caller;
1600 my $subroutine = (caller 1)[3];
cf25bb62 1601
99870f4d
KW
1602 my $list;
1603 if (ref $args_ref eq 'HASH') {
1604 foreach my $key (keys %$args_ref) {
1605 $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
cf25bb62 1606 }
99870f4d 1607 $list = join ', ', each %{$args_ref};
cf25bb62 1608 }
99870f4d
KW
1609 elsif (ref $args_ref eq 'ARRAY') {
1610 foreach my $arg (@$args_ref) {
1611 $arg = $UNDEF unless defined $arg;
1612 }
1613 $list = join ', ', @$args_ref;
1614 }
1615 else {
1616 my_carp_bug("Can't cope with ref "
1617 . ref($args_ref)
1618 . " . argument to 'carp_extra_args'. Not checking arguments.");
1619 return;
1620 }
1621
1622 my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine. Skipped.");
1623 return;
d73e5302
JH
1624}
1625
99870f4d
KW
1626package main;
1627
1628{ # Closure
1629
1630 # This program uses the inside-out method for objects, as recommended in
1631 # "Perl Best Practices". This closure aids in generating those. There
1632 # are two routines. setup_package() is called once per package to set
1633 # things up, and then set_access() is called for each hash representing a
1634 # field in the object. These routines arrange for the object to be
1635 # properly destroyed when no longer used, and for standard accessor
1636 # functions to be generated. If you need more complex accessors, just
1637 # write your own and leave those accesses out of the call to set_access().
1638 # More details below.
1639
1640 my %constructor_fields; # fields that are to be used in constructors; see
1641 # below
1642
1643 # The values of this hash will be the package names as keys to other
1644 # hashes containing the name of each field in the package as keys, and
1645 # references to their respective hashes as values.
1646 my %package_fields;
1647
1648 sub setup_package {
1649 # Sets up the package, creating standard DESTROY and dump methods
1650 # (unless already defined). The dump method is used in debugging by
1651 # simple_dumper().
1652 # The optional parameters are:
1653 # a) a reference to a hash, that gets populated by later
1654 # set_access() calls with one of the accesses being
1655 # 'constructor'. The caller can then refer to this, but it is
1656 # not otherwise used by these two routines.
1657 # b) a reference to a callback routine to call during destruction
1658 # of the object, before any fields are actually destroyed
1659
1660 my %args = @_;
1661 my $constructor_ref = delete $args{'Constructor_Fields'};
1662 my $destroy_callback = delete $args{'Destroy_Callback'};
1663 Carp::carp_extra_args(\@_) if main::DEBUG && %args;
1664
1665 my %fields;
1666 my $package = (caller)[0];
1667
1668 $package_fields{$package} = \%fields;
1669 $constructor_fields{$package} = $constructor_ref;
1670
1671 unless ($package->can('DESTROY')) {
1672 my $destroy_name = "${package}::DESTROY";
1673 no strict "refs";
1674
1675 # Use typeglob to give the anonymous subroutine the name we want
1676 *$destroy_name = sub {
1677 my $self = shift;
ffe43484 1678 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1679
1680 $self->$destroy_callback if $destroy_callback;
1681 foreach my $field (keys %{$package_fields{$package}}) {
1682 #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
1683 delete $package_fields{$package}{$field}{$addr};
1684 }
1685 return;
1686 }
1687 }
1688
1689 unless ($package->can('dump')) {
1690 my $dump_name = "${package}::dump";
1691 no strict "refs";
1692 *$dump_name = sub {
1693 my $self = shift;
1694 return dump_inside_out($self, $package_fields{$package}, @_);
1695 }
1696 }
1697 return;
1698 }
1699
1700 sub set_access {
1701 # Arrange for the input field to be garbage collected when no longer
1702 # needed. Also, creates standard accessor functions for the field
1703 # based on the optional parameters-- none if none of these parameters:
1704 # 'addable' creates an 'add_NAME()' accessor function.
1705 # 'readable' or 'readable_array' creates a 'NAME()' accessor
1706 # function.
1707 # 'settable' creates a 'set_NAME()' accessor function.
1708 # 'constructor' doesn't create an accessor function, but adds the
1709 # field to the hash that was previously passed to
1710 # setup_package();
1711 # Any of the accesses can be abbreviated down, so that 'a', 'ad',
1712 # 'add' etc. all mean 'addable'.
1713 # The read accessor function will work on both array and scalar
1714 # values. If another accessor in the parameter list is 'a', the read
1715 # access assumes an array. You can also force it to be array access
1716 # by specifying 'readable_array' instead of 'readable'
1717 #
1718 # A sort-of 'protected' access can be set-up by preceding the addable,
1719 # readable or settable with some initial portion of 'protected_' (but,
1720 # the underscore is required), like 'p_a', 'pro_set', etc. The
1721 # "protection" is only by convention. All that happens is that the
1722 # accessor functions' names begin with an underscore. So instead of
1723 # calling set_foo, the call is _set_foo. (Real protection could be
c1739a4a 1724 # accomplished by having a new subroutine, end_package, called at the
99870f4d
KW
1725 # end of each package, and then storing the __LINE__ ranges and
1726 # checking them on every accessor. But that is way overkill.)
1727
1728 # We create anonymous subroutines as the accessors and then use
1729 # typeglobs to assign them to the proper package and name
1730
1731 my $name = shift; # Name of the field
1732 my $field = shift; # Reference to the inside-out hash containing the
1733 # field
1734
1735 my $package = (caller)[0];
1736
1737 if (! exists $package_fields{$package}) {
1738 croak "$0: Must call 'setup_package' before 'set_access'";
1739 }
d73e5302 1740
99870f4d
KW
1741 # Stash the field so DESTROY can get it.
1742 $package_fields{$package}{$name} = $field;
cf25bb62 1743
99870f4d
KW
1744 # Remaining arguments are the accessors. For each...
1745 foreach my $access (@_) {
1746 my $access = lc $access;
cf25bb62 1747
99870f4d 1748 my $protected = "";
cf25bb62 1749
99870f4d
KW
1750 # Match the input as far as it goes.
1751 if ($access =~ /^(p[^_]*)_/) {
1752 $protected = $1;
1753 if (substr('protected_', 0, length $protected)
1754 eq $protected)
1755 {
1756
1757 # Add 1 for the underscore not included in $protected
1758 $access = substr($access, length($protected) + 1);
1759 $protected = '_';
1760 }
1761 else {
1762 $protected = "";
1763 }
1764 }
1765
1766 if (substr('addable', 0, length $access) eq $access) {
1767 my $subname = "${package}::${protected}add_$name";
1768 no strict "refs";
1769
1770 # add_ accessor. Don't add if already there, which we
1771 # determine using 'eq' for scalars and '==' otherwise.
1772 *$subname = sub {
1773 use strict "refs";
1774 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
1775 my $self = shift;
1776 my $value = shift;
ffe43484 1777 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
1778 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
1779 if (ref $value) {
f998e60c 1780 return if grep { $value == $_ } @{$field->{$addr}};
99870f4d
KW
1781 }
1782 else {
f998e60c 1783 return if grep { $value eq $_ } @{$field->{$addr}};
99870f4d 1784 }
f998e60c 1785 push @{$field->{$addr}}, $value;
99870f4d
KW
1786 return;
1787 }
1788 }
1789 elsif (substr('constructor', 0, length $access) eq $access) {
1790 if ($protected) {
1791 Carp::my_carp_bug("Can't set-up 'protected' constructors")
1792 }
1793 else {
1794 $constructor_fields{$package}{$name} = $field;
1795 }
1796 }
1797 elsif (substr('readable_array', 0, length $access) eq $access) {
1798
1799 # Here has read access. If one of the other parameters for
1800 # access is array, or this one specifies array (by being more
1801 # than just 'readable_'), then create a subroutine that
1802 # assumes the data is an array. Otherwise just a scalar
1803 my $subname = "${package}::${protected}$name";
1804 if (grep { /^a/i } @_
1805 or length($access) > length('readable_'))
1806 {
1807 no strict "refs";
1808 *$subname = sub {
1809 use strict "refs";
23e33b60 1810 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
ffe43484 1811 my $addr = do { no overloading; pack 'J', $_[0]; };
99870f4d
KW
1812 if (ref $field->{$addr} ne 'ARRAY') {
1813 my $type = ref $field->{$addr};
1814 $type = 'scalar' unless $type;
1815 Carp::my_carp_bug("Trying to read $name as an array when it is a $type. Big problems.");
1816 return;
1817 }
1818 return scalar @{$field->{$addr}} unless wantarray;
1819
1820 # Make a copy; had problems with caller modifying the
1821 # original otherwise
1822 my @return = @{$field->{$addr}};
1823 return @return;
1824 }
1825 }
1826 else {
1827
1828 # Here not an array value, a simpler function.
1829 no strict "refs";
1830 *$subname = sub {
1831 use strict "refs";
23e33b60 1832 Carp::carp_extra_args(\@_) if main::DEBUG && @_ > 1;
f998e60c 1833 no overloading;
051df77b 1834 return $field->{pack 'J', $_[0]};
99870f4d
KW
1835 }
1836 }
1837 }
1838 elsif (substr('settable', 0, length $access) eq $access) {
1839 my $subname = "${package}::${protected}set_$name";
1840 no strict "refs";
1841 *$subname = sub {
1842 use strict "refs";
23e33b60
KW
1843 if (main::DEBUG) {
1844 return Carp::carp_too_few_args(\@_, 2) if @_ < 2;
1845 Carp::carp_extra_args(\@_) if @_ > 2;
1846 }
1847 # $self is $_[0]; $value is $_[1]
f998e60c 1848 no overloading;
051df77b 1849 $field->{pack 'J', $_[0]} = $_[1];
99870f4d
KW
1850 return;
1851 }
1852 }
1853 else {
1854 Carp::my_carp_bug("Unknown accessor type $access. No accessor set.");
1855 }
cf25bb62 1856 }
99870f4d 1857 return;
cf25bb62 1858 }
99870f4d
KW
1859}
1860
1861package Input_file;
1862
1863# All input files use this object, which stores various attributes about them,
1864# and provides for convenient, uniform handling. The run method wraps the
1865# processing. It handles all the bookkeeping of opening, reading, and closing
1866# the file, returning only significant input lines.
1867#
1868# Each object gets a handler which processes the body of the file, and is
1869# called by run(). Most should use the generic, default handler, which has
1870# code scrubbed to handle things you might not expect. A handler should
1871# basically be a while(next_line()) {...} loop.
1872#
1873# You can also set up handlers to
1874# 1) call before the first line is read for pre processing
1875# 2) call to adjust each line of the input before the main handler gets them
1876# 3) call upon EOF before the main handler exits its loop
1877# 4) call at the end for post processing
1878#
1879# $_ is used to store the input line, and is to be filtered by the
1880# each_line_handler()s. So, if the format of the line is not in the desired
1881# format for the main handler, these are used to do that adjusting. They can
1882# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
1883# so the $_ output of one is used as the input to the next. None of the other
1884# handlers are stackable, but could easily be changed to be so.
1885#
1886# Most of the handlers can call insert_lines() or insert_adjusted_lines()
1887# which insert the parameters as lines to be processed before the next input
1888# file line is read. This allows the EOF handler to flush buffers, for
1889# example. The difference between the two routines is that the lines inserted
1890# by insert_lines() are subjected to the each_line_handler()s. (So if you
1891# called it from such a handler, you would get infinite recursion.) Lines
1892# inserted by insert_adjusted_lines() go directly to the main handler without
1893# any adjustments. If the post-processing handler calls any of these, there
1894# will be no effect. Some error checking for these conditions could be added,
1895# but it hasn't been done.
1896#
1897# carp_bad_line() should be called to warn of bad input lines, which clears $_
1898# to prevent further processing of the line. This routine will output the
1899# message as a warning once, and then keep a count of the lines that have the
1900# same message, and output that count at the end of the file's processing.
1901# This keeps the number of messages down to a manageable amount.
1902#
1903# get_missings() should be called to retrieve any @missing input lines.
1904# Messages will be raised if this isn't done if the options aren't to ignore
1905# missings.
1906
1907sub trace { return main::trace(@_); }
1908
99870f4d
KW
1909{ # Closure
1910 # Keep track of fields that are to be put into the constructor.
1911 my %constructor_fields;
1912
1913 main::setup_package(Constructor_Fields => \%constructor_fields);
1914
1915 my %file; # Input file name, required
1916 main::set_access('file', \%file, qw{ c r });
1917
1918 my %first_released; # Unicode version file was first released in, required
1919 main::set_access('first_released', \%first_released, qw{ c r });
1920
1921 my %handler; # Subroutine to process the input file, defaults to
1922 # 'process_generic_property_file'
1923 main::set_access('handler', \%handler, qw{ c });
1924
1925 my %property;
1926 # name of property this file is for. defaults to none, meaning not
1927 # applicable, or is otherwise determinable, for example, from each line.
1928 main::set_access('property', \%property, qw{ c });
1929
1930 my %optional;
1931 # If this is true, the file is optional. If not present, no warning is
1932 # output. If it is present, the string given by this parameter is
1933 # evaluated, and if false the file is not processed.
1934 main::set_access('optional', \%optional, 'c', 'r');
1935
1936 my %non_skip;
1937 # This is used for debugging, to skip processing of all but a few input
1938 # files. Add 'non_skip => 1' to the constructor for those files you want
1939 # processed when you set the $debug_skip global.
1940 main::set_access('non_skip', \%non_skip, 'c');
1941
37e2e78e
KW
1942 my %skip;
1943 # This is used to skip processing of this input file semi-permanently.
1944 # It is used for files that we aren't planning to process anytime soon,
1945 # but want to allow to be in the directory and not raise a message that we
1946 # are not handling. Mostly for test files. This is in contrast to the
1947 # non_skip element, which is supposed to be used very temporarily for
1948 # debugging. Sets 'optional' to 1
1949 main::set_access('skip', \%skip, 'c');
1950
99870f4d
KW
1951 my %each_line_handler;
1952 # list of subroutines to look at and filter each non-comment line in the
1953 # file. defaults to none. The subroutines are called in order, each is
1954 # to adjust $_ for the next one, and the final one adjusts it for
1955 # 'handler'
1956 main::set_access('each_line_handler', \%each_line_handler, 'c');
1957
1958 my %has_missings_defaults;
1959 # ? Are there lines in the file giving default values for code points
1960 # missing from it?. Defaults to NO_DEFAULTS. Otherwise NOT_IGNORED is
1961 # the norm, but IGNORED means it has such lines, but the handler doesn't
1962 # use them. Having these three states allows us to catch changes to the
1963 # UCD that this program should track
1964 main::set_access('has_missings_defaults',
1965 \%has_missings_defaults, qw{ c r });
1966
1967 my %pre_handler;
1968 # Subroutine to call before doing anything else in the file. If undef, no
1969 # such handler is called.
1970 main::set_access('pre_handler', \%pre_handler, qw{ c });
1971
1972 my %eof_handler;
1973 # Subroutine to call upon getting an EOF on the input file, but before
1974 # that is returned to the main handler. This is to allow buffers to be
1975 # flushed. The handler is expected to call insert_lines() or
1976 # insert_adjusted() with the buffered material
1977 main::set_access('eof_handler', \%eof_handler, qw{ c r });
1978
1979 my %post_handler;
1980 # Subroutine to call after all the lines of the file are read in and
1981 # processed. If undef, no such handler is called.
1982 main::set_access('post_handler', \%post_handler, qw{ c });
1983
1984 my %progress_message;
1985 # Message to print to display progress in lieu of the standard one
1986 main::set_access('progress_message', \%progress_message, qw{ c });
1987
1988 my %handle;
1989 # cache open file handle, internal. Is undef if file hasn't been
1990 # processed at all, empty if has;
1991 main::set_access('handle', \%handle);
1992
1993 my %added_lines;
1994 # cache of lines added virtually to the file, internal
1995 main::set_access('added_lines', \%added_lines);
1996
1997 my %errors;
1998 # cache of errors found, internal
1999 main::set_access('errors', \%errors);
2000
2001 my %missings;
2002 # storage of '@missing' defaults lines
2003 main::set_access('missings', \%missings);
2004
2005 sub new {
2006 my $class = shift;
2007
2008 my $self = bless \do{ my $anonymous_scalar }, $class;
ffe43484 2009 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2010
2011 # Set defaults
2012 $handler{$addr} = \&main::process_generic_property_file;
2013 $non_skip{$addr} = 0;
37e2e78e 2014 $skip{$addr} = 0;
99870f4d
KW
2015 $has_missings_defaults{$addr} = $NO_DEFAULTS;
2016 $handle{$addr} = undef;
2017 $added_lines{$addr} = [ ];
2018 $each_line_handler{$addr} = [ ];
2019 $errors{$addr} = { };
2020 $missings{$addr} = [ ];
2021
2022 # Two positional parameters.
99f78760 2023 return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
99870f4d
KW
2024 $file{$addr} = main::internal_file_to_platform(shift);
2025 $first_released{$addr} = shift;
2026
2027 # The rest of the arguments are key => value pairs
2028 # %constructor_fields has been set up earlier to list all possible
2029 # ones. Either set or push, depending on how the default has been set
2030 # up just above.
2031 my %args = @_;
2032 foreach my $key (keys %args) {
2033 my $argument = $args{$key};
2034
2035 # Note that the fields are the lower case of the constructor keys
2036 my $hash = $constructor_fields{lc $key};
2037 if (! defined $hash) {
2038 Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self. Skipped");
2039 next;
2040 }
2041 if (ref $hash->{$addr} eq 'ARRAY') {
2042 if (ref $argument eq 'ARRAY') {
2043 foreach my $argument (@{$argument}) {
2044 next if ! defined $argument;
2045 push @{$hash->{$addr}}, $argument;
2046 }
2047 }
2048 else {
2049 push @{$hash->{$addr}}, $argument if defined $argument;
2050 }
2051 }
2052 else {
2053 $hash->{$addr} = $argument;
2054 }
2055 delete $args{$key};
2056 };
2057
2058 # If the file has a property for it, it means that the property is not
2059 # listed in the file's entries. So add a handler to the list of line
2060 # handlers to insert the property name into the lines, to provide a
2061 # uniform interface to the final processing subroutine.
2062 # the final code doesn't have to worry about that.
2063 if ($property{$addr}) {
2064 push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
2065 }
2066
2067 if ($non_skip{$addr} && ! $debug_skip && $verbosity) {
2068 print "Warning: " . __PACKAGE__ . " constructor for $file{$addr} has useless 'non_skip' in it\n";
a3a8c5f0 2069 }
99870f4d 2070
37e2e78e
KW
2071 $optional{$addr} = 1 if $skip{$addr};
2072
99870f4d 2073 return $self;
d73e5302
JH
2074 }
2075
cf25bb62 2076
99870f4d
KW
2077 use overload
2078 fallback => 0,
2079 qw("") => "_operator_stringify",
2080 "." => \&main::_operator_dot,
2081 ;
cf25bb62 2082
99870f4d
KW
2083 sub _operator_stringify {
2084 my $self = shift;
cf25bb62 2085
99870f4d 2086 return __PACKAGE__ . " object for " . $self->file;
d73e5302 2087 }
d73e5302 2088
99870f4d
KW
2089 # flag to make sure extracted files are processed early
2090 my $seen_non_extracted_non_age = 0;
d73e5302 2091
99870f4d
KW
2092 sub run {
2093 # Process the input object $self. This opens and closes the file and
2094 # calls all the handlers for it. Currently, this can only be called
2095 # once per file, as it destroy's the EOF handler
d73e5302 2096
99870f4d
KW
2097 my $self = shift;
2098 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
b6922eda 2099
ffe43484 2100 my $addr = do { no overloading; pack 'J', $self; };
b6922eda 2101
99870f4d 2102 my $file = $file{$addr};
d73e5302 2103
99870f4d
KW
2104 # Don't process if not expecting this file (because released later
2105 # than this Unicode version), and isn't there. This means if someone
2106 # copies it into an earlier version's directory, we will go ahead and
2107 # process it.
2108 return if $first_released{$addr} gt $v_version && ! -e $file;
2109
2110 # If in debugging mode and this file doesn't have the non-skip
2111 # flag set, and isn't one of the critical files, skip it.
2112 if ($debug_skip
2113 && $first_released{$addr} ne v0
2114 && ! $non_skip{$addr})
2115 {
2116 print "Skipping $file in debugging\n" if $verbosity;
2117 return;
2118 }
2119
2120 # File could be optional
37e2e78e 2121 if ($optional{$addr}) {
99870f4d
KW
2122 return unless -e $file;
2123 my $result = eval $optional{$addr};
2124 if (! defined $result) {
2125 Carp::my_carp_bug("Got '$@' when tried to eval $optional{$addr}. $file Skipped.");
2126 return;
2127 }
2128 if (! $result) {
2129 if ($verbosity) {
2130 print STDERR "Skipping processing input file '$file' because '$optional{$addr}' is not true\n";
2131 }
2132 return;
2133 }
2134 }
2135
2136 if (! defined $file || ! -e $file) {
2137
2138 # If the file doesn't exist, see if have internal data for it
2139 # (based on first_released being 0).
2140 if ($first_released{$addr} eq v0) {
2141 $handle{$addr} = 'pretend_is_open';
2142 }
2143 else {
2144 if (! $optional{$addr} # File could be optional
2145 && $v_version ge $first_released{$addr})
2146 {
2147 print STDERR "Skipping processing input file '$file' because not found\n" if $v_version ge $first_released{$addr};
2148 }
2149 return;
2150 }
2151 }
2152 else {
2153
37e2e78e
KW
2154 # Here, the file exists. Some platforms may change the case of
2155 # its name
99870f4d 2156 if ($seen_non_extracted_non_age) {
517956bf 2157 if ($file =~ /$EXTRACTED/i) {
99870f4d 2158 Carp::my_carp_bug(join_lines(<<END
99f78760 2159$file should be processed just after the 'Prop...Alias' files, and before
99870f4d
KW
2160anything not in the $EXTRACTED_DIR directory. Proceeding, but the results may
2161have subtle problems
2162END
2163 ));
2164 }
2165 }
2166 elsif ($EXTRACTED_DIR
2167 && $first_released{$addr} ne v0
517956bf
CB
2168 && $file !~ /$EXTRACTED/i
2169 && lc($file) ne 'dage.txt')
99870f4d
KW
2170 {
2171 # We don't set this (by the 'if' above) if we have no
2172 # extracted directory, so if running on an early version,
2173 # this test won't work. Not worth worrying about.
2174 $seen_non_extracted_non_age = 1;
2175 }
2176
2177 # And mark the file as having being processed, and warn if it
2178 # isn't a file we are expecting. As we process the files,
2179 # they are deleted from the hash, so any that remain at the
2180 # end of the program are files that we didn't process.
517956bf
CB
2181 my $fkey = File::Spec->rel2abs($file);
2182 my $expecting = delete $potential_files{$fkey};
2183 $expecting = delete $potential_files{lc($fkey)} unless defined $expecting;
678f13d5
KW
2184 Carp::my_carp("Was not expecting '$file'.") if
2185 ! $expecting
99870f4d
KW
2186 && ! defined $handle{$addr};
2187
37e2e78e
KW
2188 # Having deleted from expected files, we can quit if not to do
2189 # anything. Don't print progress unless really want verbosity
2190 if ($skip{$addr}) {
2191 print "Skipping $file.\n" if $verbosity >= $VERBOSE;
2192 return;
2193 }
2194
99870f4d
KW
2195 # Open the file, converting the slashes used in this program
2196 # into the proper form for the OS
2197 my $file_handle;
2198 if (not open $file_handle, "<", $file) {
2199 Carp::my_carp("Can't open $file. Skipping: $!");
2200 return 0;
2201 }
2202 $handle{$addr} = $file_handle; # Cache the open file handle
2203 }
2204
2205 if ($verbosity >= $PROGRESS) {
2206 if ($progress_message{$addr}) {
2207 print "$progress_message{$addr}\n";
2208 }
2209 else {
2210 # If using a virtual file, say so.
2211 print "Processing ", (-e $file)
2212 ? $file
2213 : "substitute $file",
2214 "\n";
2215 }
2216 }
2217
2218
2219 # Call any special handler for before the file.
2220 &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
2221
2222 # Then the main handler
2223 &{$handler{$addr}}($self);
2224
2225 # Then any special post-file handler.
2226 &{$post_handler{$addr}}($self) if $post_handler{$addr};
2227
2228 # If any errors have been accumulated, output the counts (as the first
2229 # error message in each class was output when it was encountered).
2230 if ($errors{$addr}) {
2231 my $total = 0;
2232 my $types = 0;
2233 foreach my $error (keys %{$errors{$addr}}) {
2234 $total += $errors{$addr}->{$error};
2235 delete $errors{$addr}->{$error};
2236 $types++;
2237 }
2238 if ($total > 1) {
2239 my $message
2240 = "A total of $total lines had errors in $file. ";
2241
2242 $message .= ($types == 1)
2243 ? '(Only the first one was displayed.)'
2244 : '(Only the first of each type was displayed.)';
2245 Carp::my_carp($message);
2246 }
2247 }
2248
2249 if (@{$missings{$addr}}) {
2250 Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines. Generated tables likely are wrong");
2251 }
2252
2253 # If a real file handle, close it.
2254 close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
2255 ref $handle{$addr};
2256 $handle{$addr} = ""; # Uses empty to indicate that has already seen
2257 # the file, as opposed to undef
2258 return;
2259 }
2260
2261 sub next_line {
2262 # Sets $_ to be the next logical input line, if any. Returns non-zero
2263 # if such a line exists. 'logical' means that any lines that have
2264 # been added via insert_lines() will be returned in $_ before the file
2265 # is read again.
2266
2267 my $self = shift;
2268 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2269
ffe43484 2270 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2271
2272 # Here the file is open (or if the handle is not a ref, is an open
2273 # 'virtual' file). Get the next line; any inserted lines get priority
2274 # over the file itself.
2275 my $adjusted;
2276
2277 LINE:
2278 while (1) { # Loop until find non-comment, non-empty line
2279 #local $to_trace = 1 if main::DEBUG;
2280 my $inserted_ref = shift @{$added_lines{$addr}};
2281 if (defined $inserted_ref) {
2282 ($adjusted, $_) = @{$inserted_ref};
2283 trace $adjusted, $_ if main::DEBUG && $to_trace;
2284 return 1 if $adjusted;
2285 }
2286 else {
2287 last if ! ref $handle{$addr}; # Don't read unless is real file
2288 last if ! defined ($_ = readline $handle{$addr});
2289 }
2290 chomp;
2291 trace $_ if main::DEBUG && $to_trace;
2292
2293 # See if this line is the comment line that defines what property
2294 # value that code points that are not listed in the file should
2295 # have. The format or existence of these lines is not guaranteed
2296 # by Unicode since they are comments, but the documentation says
2297 # that this was added for machine-readability, so probably won't
2298 # change. This works starting in Unicode Version 5.0. They look
2299 # like:
2300 #
2301 # @missing: 0000..10FFFF; Not_Reordered
2302 # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
2303 # @missing: 0000..10FFFF; ; NaN
2304 #
2305 # Save the line for a later get_missings() call.
2306 if (/$missing_defaults_prefix/) {
2307 if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
2308 $self->carp_bad_line("Unexpected \@missing line. Assuming no missing entries");
2309 }
2310 elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
2311 my @defaults = split /\s* ; \s*/x, $_;
2312
2313 # The first field is the @missing, which ends in a
2314 # semi-colon, so can safely shift.
2315 shift @defaults;
2316
2317 # Some of these lines may have empty field placeholders
2318 # which get in the way. An example is:
2319 # @missing: 0000..10FFFF; ; NaN
2320 # Remove them. Process starting from the top so the
2321 # splice doesn't affect things still to be looked at.
2322 for (my $i = @defaults - 1; $i >= 0; $i--) {
2323 next if $defaults[$i] ne "";
2324 splice @defaults, $i, 1;
2325 }
2326
2327 # What's left should be just the property (maybe) and the
2328 # default. Having only one element means it doesn't have
2329 # the property.
2330 my $default;
2331 my $property;
2332 if (@defaults >= 1) {
2333 if (@defaults == 1) {
2334 $default = $defaults[0];
2335 }
2336 else {
2337 $property = $defaults[0];
2338 $default = $defaults[1];
2339 }
2340 }
2341
2342 if (@defaults < 1
2343 || @defaults > 2
2344 || ($default =~ /^</
2345 && $default !~ /^<code *point>$/i
2346 && $default !~ /^<none>$/i))
2347 {
2348 $self->carp_bad_line("Unrecognized \@missing line: $_. Assuming no missing entries");
2349 }
2350 else {
2351
2352 # If the property is missing from the line, it should
2353 # be the one for the whole file
2354 $property = $property{$addr} if ! defined $property;
2355
2356 # Change <none> to the null string, which is what it
2357 # really means. If the default is the code point
2358 # itself, set it to <code point>, which is what
2359 # Unicode uses (but sometimes they've forgotten the
2360 # space)
2361 if ($default =~ /^<none>$/i) {
2362 $default = "";
2363 }
2364 elsif ($default =~ /^<code *point>$/i) {
2365 $default = $CODE_POINT;
2366 }
2367
2368 # Store them as a sub-arrays with both components.
2369 push @{$missings{$addr}}, [ $default, $property ];
2370 }
2371 }
2372
2373 # There is nothing for the caller to process on this comment
2374 # line.
2375 next;
2376 }
2377
2378 # Remove comments and trailing space, and skip this line if the
2379 # result is empty
2380 s/#.*//;
2381 s/\s+$//;
2382 next if /^$/;
2383
2384 # Call any handlers for this line, and skip further processing of
2385 # the line if the handler sets the line to null.
2386 foreach my $sub_ref (@{$each_line_handler{$addr}}) {
2387 &{$sub_ref}($self);
2388 next LINE if /^$/;
2389 }
2390
2391 # Here the line is ok. return success.
2392 return 1;
2393 } # End of looping through lines.
2394
2395 # If there is an EOF handler, call it (only once) and if it generates
2396 # more lines to process go back in the loop to handle them.
2397 if ($eof_handler{$addr}) {
2398 &{$eof_handler{$addr}}($self);
2399 $eof_handler{$addr} = ""; # Currently only get one shot at it.
2400 goto LINE if $added_lines{$addr};
2401 }
2402
2403 # Return failure -- no more lines.
2404 return 0;
2405
2406 }
2407
2408# Not currently used, not fully tested.
2409# sub peek {
2410# # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
2411# # record. Not callable from an each_line_handler(), nor does it call
2412# # an each_line_handler() on the line.
2413#
2414# my $self = shift;
ffe43484 2415# my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2416#
2417# foreach my $inserted_ref (@{$added_lines{$addr}}) {
2418# my ($adjusted, $line) = @{$inserted_ref};
2419# next if $adjusted;
2420#
2421# # Remove comments and trailing space, and return a non-empty
2422# # resulting line
2423# $line =~ s/#.*//;
2424# $line =~ s/\s+$//;
2425# return $line if $line ne "";
2426# }
2427#
2428# return if ! ref $handle{$addr}; # Don't read unless is real file
2429# while (1) { # Loop until find non-comment, non-empty line
2430# local $to_trace = 1 if main::DEBUG;
2431# trace $_ if main::DEBUG && $to_trace;
2432# return if ! defined (my $line = readline $handle{$addr});
2433# chomp $line;
2434# push @{$added_lines{$addr}}, [ 0, $line ];
2435#
2436# $line =~ s/#.*//;
2437# $line =~ s/\s+$//;
2438# return $line if $line ne "";
2439# }
2440#
2441# return;
2442# }
2443
2444
2445 sub insert_lines {
2446 # Lines can be inserted so that it looks like they were in the input
2447 # file at the place it was when this routine is called. See also
2448 # insert_adjusted_lines(). Lines inserted via this routine go through
2449 # any each_line_handler()
2450
2451 my $self = shift;
2452
2453 # Each inserted line is an array, with the first element being 0 to
2454 # indicate that this line hasn't been adjusted, and needs to be
2455 # processed.
f998e60c 2456 no overloading;
051df77b 2457 push @{$added_lines{pack 'J', $self}}, map { [ 0, $_ ] } @_;
99870f4d
KW
2458 return;
2459 }
2460
2461 sub insert_adjusted_lines {
2462 # Lines can be inserted so that it looks like they were in the input
2463 # file at the place it was when this routine is called. See also
2464 # insert_lines(). Lines inserted via this routine are already fully
2465 # adjusted, ready to be processed; each_line_handler()s handlers will
2466 # not be called. This means this is not a completely general
2467 # facility, as only the last each_line_handler on the stack should
2468 # call this. It could be made more general, by passing to each of the
2469 # line_handlers their position on the stack, which they would pass on
2470 # to this routine, and that would replace the boolean first element in
2471 # the anonymous array pushed here, so that the next_line routine could
2472 # use that to call only those handlers whose index is after it on the
2473 # stack. But this is overkill for what is needed now.
2474
2475 my $self = shift;
2476 trace $_[0] if main::DEBUG && $to_trace;
2477
2478 # Each inserted line is an array, with the first element being 1 to
2479 # indicate that this line has been adjusted
f998e60c 2480 no overloading;
051df77b 2481 push @{$added_lines{pack 'J', $self}}, map { [ 1, $_ ] } @_;
99870f4d
KW
2482 return;
2483 }
2484
2485 sub get_missings {
2486 # Returns the stored up @missings lines' values, and clears the list.
2487 # The values are in an array, consisting of the default in the first
2488 # element, and the property in the 2nd. However, since these lines
2489 # can be stacked up, the return is an array of all these arrays.
2490
2491 my $self = shift;
2492 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2493
ffe43484 2494 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2495
2496 # If not accepting a list return, just return the first one.
2497 return shift @{$missings{$addr}} unless wantarray;
2498
2499 my @return = @{$missings{$addr}};
2500 undef @{$missings{$addr}};
2501 return @return;
2502 }
2503
2504 sub _insert_property_into_line {
2505 # Add a property field to $_, if this file requires it.
2506
f998e60c 2507 my $self = shift;
ffe43484 2508 my $addr = do { no overloading; pack 'J', $self; };
f998e60c 2509 my $property = $property{$addr};
99870f4d
KW
2510 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2511
2512 $_ =~ s/(;|$)/; $property$1/;
2513 return;
2514 }
2515
2516 sub carp_bad_line {
2517 # Output consistent error messages, using either a generic one, or the
2518 # one given by the optional parameter. To avoid gazillions of the
2519 # same message in case the syntax of a file is way off, this routine
2520 # only outputs the first instance of each message, incrementing a
2521 # count so the totals can be output at the end of the file.
2522
2523 my $self = shift;
2524 my $message = shift;
2525 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2526
ffe43484 2527 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2528
2529 $message = 'Unexpected line' unless $message;
2530
2531 # No trailing punctuation so as to fit with our addenda.
2532 $message =~ s/[.:;,]$//;
2533
2534 # If haven't seen this exact message before, output it now. Otherwise
2535 # increment the count of how many times it has occurred
2536 unless ($errors{$addr}->{$message}) {
2537 Carp::my_carp("$message in '$_' in "
f998e60c 2538 . $file{$addr}
99870f4d
KW
2539 . " at line $.. Skipping this line;");
2540 $errors{$addr}->{$message} = 1;
2541 }
2542 else {
2543 $errors{$addr}->{$message}++;
2544 }
2545
2546 # Clear the line to prevent any further (meaningful) processing of it.
2547 $_ = "";
2548
2549 return;
2550 }
2551} # End closure
2552
2553package Multi_Default;
2554
2555# Certain properties in early versions of Unicode had more than one possible
2556# default for code points missing from the files. In these cases, one
2557# default applies to everything left over after all the others are applied,
2558# and for each of the others, there is a description of which class of code
2559# points applies to it. This object helps implement this by storing the
2560# defaults, and for all but that final default, an eval string that generates
2561# the class that it applies to.
2562
2563
2564{ # Closure
2565
2566 main::setup_package();
2567
2568 my %class_defaults;
2569 # The defaults structure for the classes
2570 main::set_access('class_defaults', \%class_defaults);
2571
2572 my %other_default;
2573 # The default that applies to everything left over.
2574 main::set_access('other_default', \%other_default, 'r');
2575
2576
2577 sub new {
2578 # The constructor is called with default => eval pairs, terminated by
2579 # the left-over default. e.g.
2580 # Multi_Default->new(
2581 # 'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
2582 # - 0x200D',
2583 # 'R' => 'some other expression that evaluates to code points',
2584 # .
2585 # .
2586 # .
2587 # 'U'));
2588
2589 my $class = shift;
2590
2591 my $self = bless \do{my $anonymous_scalar}, $class;
ffe43484 2592 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2593
2594 while (@_ > 1) {
2595 my $default = shift;
2596 my $eval = shift;
2597 $class_defaults{$addr}->{$default} = $eval;
2598 }
2599
2600 $other_default{$addr} = shift;
2601
2602 return $self;
2603 }
2604
2605 sub get_next_defaults {
2606 # Iterates and returns the next class of defaults.
2607 my $self = shift;
2608 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2609
ffe43484 2610 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2611
2612 return each %{$class_defaults{$addr}};
2613 }
2614}
2615
2616package Alias;
2617
2618# An alias is one of the names that a table goes by. This class defines them
2619# including some attributes. Everything is currently setup in the
2620# constructor.
2621
2622
2623{ # Closure
2624
2625 main::setup_package();
2626
2627 my %name;
2628 main::set_access('name', \%name, 'r');
2629
2630 my %loose_match;
2631 # Determined by the constructor code if this name should match loosely or
2632 # not. The constructor parameters can override this, but it isn't fully
2633 # implemented, as should have ability to override Unicode one's via
2634 # something like a set_loose_match()
2635 main::set_access('loose_match', \%loose_match, 'r');
2636
2637 my %make_pod_entry;
2638 # Some aliases should not get their own entries because they are covered
2639 # by a wild-card, and some we want to discourage use of. Binary
2640 main::set_access('make_pod_entry', \%make_pod_entry, 'r');
2641
2642 my %status;
2643 # Aliases have a status, like deprecated, or even suppressed (which means
2644 # they don't appear in documentation). Enum
2645 main::set_access('status', \%status, 'r');
2646
2647 my %externally_ok;
2648 # Similarly, some aliases should not be considered as usable ones for
2649 # external use, such as file names, or we don't want documentation to
2650 # recommend them. Boolean
2651 main::set_access('externally_ok', \%externally_ok, 'r');
2652
2653 sub new {
2654 my $class = shift;
2655
2656 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2657 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2658
2659 $name{$addr} = shift;
2660 $loose_match{$addr} = shift;
2661 $make_pod_entry{$addr} = shift;
2662 $externally_ok{$addr} = shift;
2663 $status{$addr} = shift;
2664
2665 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2666
2667 # Null names are never ok externally
2668 $externally_ok{$addr} = 0 if $name{$addr} eq "";
2669
2670 return $self;
2671 }
2672}
2673
2674package Range;
2675
2676# A range is the basic unit for storing code points, and is described in the
2677# comments at the beginning of the program. Each range has a starting code
2678# point; an ending code point (not less than the starting one); a value
2679# that applies to every code point in between the two end-points, inclusive;
2680# and an enum type that applies to the value. The type is for the user's
2681# convenience, and has no meaning here, except that a non-zero type is
2682# considered to not obey the normal Unicode rules for having standard forms.
2683#
2684# The same structure is used for both map and match tables, even though in the
2685# latter, the value (and hence type) is irrelevant and could be used as a
2686# comment. In map tables, the value is what all the code points in the range
2687# map to. Type 0 values have the standardized version of the value stored as
2688# well, so as to not have to recalculate it a lot.
2689
2690sub trace { return main::trace(@_); }
2691
2692{ # Closure
2693
2694 main::setup_package();
2695
2696 my %start;
2697 main::set_access('start', \%start, 'r', 's');
2698
2699 my %end;
2700 main::set_access('end', \%end, 'r', 's');
2701
2702 my %value;
2703 main::set_access('value', \%value, 'r');
2704
2705 my %type;
2706 main::set_access('type', \%type, 'r');
2707
2708 my %standard_form;
2709 # The value in internal standard form. Defined only if the type is 0.
2710 main::set_access('standard_form', \%standard_form);
2711
2712 # Note that if these fields change, the dump() method should as well
2713
2714 sub new {
2715 return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
2716 my $class = shift;
2717
2718 my $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2719 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2720
2721 $start{$addr} = shift;
2722 $end{$addr} = shift;
2723
2724 my %args = @_;
2725
2726 my $value = delete $args{'Value'}; # Can be 0
2727 $value = "" unless defined $value;
2728 $value{$addr} = $value;
2729
2730 $type{$addr} = delete $args{'Type'} || 0;
2731
2732 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2733
2734 if (! $type{$addr}) {
2735 $standard_form{$addr} = main::standardize($value);
2736 }
2737
2738 return $self;
2739 }
2740
2741 use overload
2742 fallback => 0,
2743 qw("") => "_operator_stringify",
2744 "." => \&main::_operator_dot,
2745 ;
2746
2747 sub _operator_stringify {
2748 my $self = shift;
ffe43484 2749 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2750
2751 # Output it like '0041..0065 (value)'
2752 my $return = sprintf("%04X", $start{$addr})
2753 . '..'
2754 . sprintf("%04X", $end{$addr});
2755 my $value = $value{$addr};
2756 my $type = $type{$addr};
2757 $return .= ' (';
2758 $return .= "$value";
2759 $return .= ", Type=$type" if $type != 0;
2760 $return .= ')';
2761
2762 return $return;
2763 }
2764
2765 sub standard_form {
2766 # The standard form is the value itself if the standard form is
2767 # undefined (that is if the value is special)
2768
2769 my $self = shift;
2770 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2771
ffe43484 2772 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2773
2774 return $standard_form{$addr} if defined $standard_form{$addr};
2775 return $value{$addr};
2776 }
2777
2778 sub dump {
2779 # Human, not machine readable. For machine readable, comment out this
2780 # entire routine and let the standard one take effect.
2781 my $self = shift;
2782 my $indent = shift;
2783 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
2784
ffe43484 2785 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2786
2787 my $return = $indent
2788 . sprintf("%04X", $start{$addr})
2789 . '..'
2790 . sprintf("%04X", $end{$addr})
2791 . " '$value{$addr}';";
2792 if (! defined $standard_form{$addr}) {
2793 $return .= "(type=$type{$addr})";
2794 }
2795 elsif ($standard_form{$addr} ne $value{$addr}) {
2796 $return .= "(standard '$standard_form{$addr}')";
2797 }
2798 return $return;
2799 }
2800} # End closure
2801
2802package _Range_List_Base;
2803
2804# Base class for range lists. A range list is simply an ordered list of
2805# ranges, so that the ranges with the lowest starting numbers are first in it.
2806#
2807# When a new range is added that is adjacent to an existing range that has the
2808# same value and type, it merges with it to form a larger range.
2809#
2810# Ranges generally do not overlap, except that there can be multiple entries
2811# of single code point ranges. This is because of NameAliases.txt.
2812#
2813# In this program, there is a standard value such that if two different
2814# values, have the same standard value, they are considered equivalent. This
2815# value was chosen so that it gives correct results on Unicode data
2816
2817# There are a number of methods to manipulate range lists, and some operators
2818# are overloaded to handle them.
2819
99870f4d
KW
2820sub trace { return main::trace(@_); }
2821
2822{ # Closure
2823
2824 our $addr;
2825
2826 main::setup_package();
2827
2828 my %ranges;
2829 # The list of ranges
2830 main::set_access('ranges', \%ranges, 'readable_array');
2831
2832 my %max;
2833 # The highest code point in the list. This was originally a method, but
2834 # actual measurements said it was used a lot.
2835 main::set_access('max', \%max, 'r');
2836
2837 my %each_range_iterator;
2838 # Iterator position for each_range()
2839 main::set_access('each_range_iterator', \%each_range_iterator);
2840
2841 my %owner_name_of;
2842 # Name of parent this is attached to, if any. Solely for better error
2843 # messages.
2844 main::set_access('owner_name_of', \%owner_name_of, 'p_r');
2845
2846 my %_search_ranges_cache;
2847 # A cache of the previous result from _search_ranges(), for better
2848 # performance
2849 main::set_access('_search_ranges_cache', \%_search_ranges_cache);
2850
2851 sub new {
2852 my $class = shift;
2853 my %args = @_;
2854
2855 # Optional initialization data for the range list.
2856 my $initialize = delete $args{'Initialize'};
2857
2858 my $self;
2859
2860 # Use _union() to initialize. _union() returns an object of this
2861 # class, which means that it will call this constructor recursively.
2862 # But it won't have this $initialize parameter so that it won't
2863 # infinitely loop on this.
2864 return _union($class, $initialize, %args) if defined $initialize;
2865
2866 $self = bless \do { my $anonymous_scalar }, $class;
ffe43484 2867 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2868
2869 # Optional parent object, only for debug info.
2870 $owner_name_of{$addr} = delete $args{'Owner'};
2871 $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
2872
2873 # Stringify, in case it is an object.
2874 $owner_name_of{$addr} = "$owner_name_of{$addr}";
2875
2876 # This is used only for error messages, and so a colon is added
2877 $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
2878
2879 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
2880
2881 # Max is initialized to a negative value that isn't adjacent to 0,
2882 # for simpler tests
2883 $max{$addr} = -2;
2884
2885 $_search_ranges_cache{$addr} = 0;
2886 $ranges{$addr} = [];
2887
2888 return $self;
2889 }
2890
2891 use overload
2892 fallback => 0,
2893 qw("") => "_operator_stringify",
2894 "." => \&main::_operator_dot,
2895 ;
2896
2897 sub _operator_stringify {
2898 my $self = shift;
ffe43484 2899 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
2900
2901 return "Range_List attached to '$owner_name_of{$addr}'"
2902 if $owner_name_of{$addr};
2903 return "anonymous Range_List " . \$self;
2904 }
2905
2906 sub _union {
2907 # Returns the union of the input code points. It can be called as
2908 # either a constructor or a method. If called as a method, the result
2909 # will be a new() instance of the calling object, containing the union
2910 # of that object with the other parameter's code points; if called as
2911 # a constructor, the first parameter gives the class the new object
2912 # should be, and the second parameter gives the code points to go into
2913 # it.
2914 # In either case, there are two parameters looked at by this routine;
2915 # any additional parameters are passed to the new() constructor.
2916 #
2917 # The code points can come in the form of some object that contains
2918 # ranges, and has a conventionally named method to access them; or
2919 # they can be an array of individual code points (as integers); or
2920 # just a single code point.
2921 #
2922 # If they are ranges, this routine doesn't make any effort to preserve
2923 # the range values of one input over the other. Therefore this base
2924 # class should not allow _union to be called from other than
2925 # initialization code, so as to prevent two tables from being added
2926 # together where the range values matter. The general form of this
2927 # routine therefore belongs in a derived class, but it was moved here
2928 # to avoid duplication of code. The failure to overload this in this
2929 # class keeps it safe.
2930 #
2931
2932 my $self;
2933 my @args; # Arguments to pass to the constructor
2934
2935 my $class = shift;
2936
2937 # If a method call, will start the union with the object itself, and
2938 # the class of the new object will be the same as self.
2939 if (ref $class) {
2940 $self = $class;
2941 $class = ref $self;
2942 push @args, $self;
2943 }
2944
2945 # Add the other required parameter.
2946 push @args, shift;
2947 # Rest of parameters are passed on to the constructor
2948
2949 # Accumulate all records from both lists.
2950 my @records;
2951 for my $arg (@args) {
2952 #local $to_trace = 0 if main::DEBUG;
2953 trace "argument = $arg" if main::DEBUG && $to_trace;
2954 if (! defined $arg) {
2955 my $message = "";
2956 if (defined $self) {
f998e60c 2957 no overloading;
051df77b 2958 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2959 }
2960 Carp::my_carp_bug($message .= "Undefined argument to _union. No union done.");
2961 return;
2962 }
2963 $arg = [ $arg ] if ! ref $arg;
2964 my $type = ref $arg;
2965 if ($type eq 'ARRAY') {
2966 foreach my $element (@$arg) {
2967 push @records, Range->new($element, $element);
2968 }
2969 }
2970 elsif ($arg->isa('Range')) {
2971 push @records, $arg;
2972 }
2973 elsif ($arg->can('ranges')) {
2974 push @records, $arg->ranges;
2975 }
2976 else {
2977 my $message = "";
2978 if (defined $self) {
f998e60c 2979 no overloading;
051df77b 2980 $message .= $owner_name_of{pack 'J', $self};
99870f4d
KW
2981 }
2982 Carp::my_carp_bug($message . "Cannot take the union of a $type. No union done.");
2983 return;
2984 }
2985 }
2986
2987 # Sort with the range containing the lowest ordinal first, but if
2988 # two ranges start at the same code point, sort with the bigger range
2989 # of the two first, because it takes fewer cycles.
2990 @records = sort { ($a->start <=> $b->start)
2991 or
2992 # if b is shorter than a, b->end will be
2993 # less than a->end, and we want to select
2994 # a, so want to return -1
2995 ($b->end <=> $a->end)
2996 } @records;
2997
2998 my $new = $class->new(@_);
2999
3000 # Fold in records so long as they add new information.
3001 for my $set (@records) {
3002 my $start = $set->start;
3003 my $end = $set->end;
3004 my $value = $set->value;
3005 if ($start > $new->max) {
3006 $new->_add_delete('+', $start, $end, $value);
3007 }
3008 elsif ($end > $new->max) {
3009 $new->_add_delete('+', $new->max +1, $end, $value);
3010 }
3011 }
3012
3013 return $new;
3014 }
3015
3016 sub range_count { # Return the number of ranges in the range list
3017 my $self = shift;
3018 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3019
f998e60c 3020 no overloading;
051df77b 3021 return scalar @{$ranges{pack 'J', $self}};
99870f4d
KW
3022 }
3023
3024 sub min {
3025 # Returns the minimum code point currently in the range list, or if
3026 # the range list is empty, 2 beyond the max possible. This is a
3027 # method because used so rarely, that not worth saving between calls,
3028 # and having to worry about changing it as ranges are added and
3029 # deleted.
3030
3031 my $self = shift;
3032 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3033
ffe43484 3034 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3035
3036 # If the range list is empty, return a large value that isn't adjacent
3037 # to any that could be in the range list, for simpler tests
3038 return $LAST_UNICODE_CODEPOINT + 2 unless scalar @{$ranges{$addr}};
3039 return $ranges{$addr}->[0]->start;
3040 }
3041
3042 sub contains {
3043 # Boolean: Is argument in the range list? If so returns $i such that:
3044 # range[$i]->end < $codepoint <= range[$i+1]->end
3045 # which is one beyond what you want; this is so that the 0th range
3046 # doesn't return false
3047 my $self = shift;
3048 my $codepoint = shift;
3049 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3050
99870f4d
KW
3051 my $i = $self->_search_ranges($codepoint);
3052 return 0 unless defined $i;
3053
3054 # The search returns $i, such that
3055 # range[$i-1]->end < $codepoint <= range[$i]->end
3056 # So is in the table if and only iff it is at least the start position
3057 # of range $i.
f998e60c 3058 no overloading;
051df77b 3059 return 0 if $ranges{pack 'J', $self}->[$i]->start > $codepoint;
99870f4d
KW
3060 return $i + 1;
3061 }
3062
2f7a8815
KW
3063 sub containing_range {
3064 # Returns the range object that contains the code point, undef if none
3065
3066 my $self = shift;
3067 my $codepoint = shift;
3068 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3069
3070 my $i = $self->contains($codepoint);
3071 return unless $i;
3072
3073 # contains() returns 1 beyond where we should look
3074 no overloading;
3075 return $ranges{pack 'J', $self}->[$i-1];
3076 }
3077
99870f4d
KW
3078 sub value_of {
3079 # Returns the value associated with the code point, undef if none
3080
3081 my $self = shift;
3082 my $codepoint = shift;
3083 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3084
d69c231b
KW
3085 my $range = $self->containing_range($codepoint);
3086 return unless defined $range;
99870f4d 3087
d69c231b 3088 return $range->value;
99870f4d
KW
3089 }
3090
0a9dbafc
KW
3091 sub type_of {
3092 # Returns the type of the range containing the code point, undef if
3093 # the code point is not in the table
3094
3095 my $self = shift;
3096 my $codepoint = shift;
3097 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3098
3099 my $range = $self->containing_range($codepoint);
3100 return unless defined $range;
3101
3102 return $range->type;
3103 }
3104
99870f4d
KW
3105 sub _search_ranges {
3106 # Find the range in the list which contains a code point, or where it
3107 # should go if were to add it. That is, it returns $i, such that:
3108 # range[$i-1]->end < $codepoint <= range[$i]->end
3109 # Returns undef if no such $i is possible (e.g. at end of table), or
3110 # if there is an error.
3111
3112 my $self = shift;
3113 my $code_point = shift;
3114 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3115
ffe43484 3116 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3117
3118 return if $code_point > $max{$addr};
3119 my $r = $ranges{$addr}; # The current list of ranges
3120 my $range_list_size = scalar @$r;
3121 my $i;
3122
3123 use integer; # want integer division
3124
3125 # Use the cached result as the starting guess for this one, because,
3126 # an experiment on 5.1 showed that 90% of the time the cache was the
3127 # same as the result on the next call (and 7% it was one less).
3128 $i = $_search_ranges_cache{$addr};
3129 $i = 0 if $i >= $range_list_size; # Reset if no longer valid (prob.
3130 # from an intervening deletion
3131 #local $to_trace = 1 if main::DEBUG;
3132 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);
3133 return $i if $code_point <= $r->[$i]->end
3134 && ($i == 0 || $r->[$i-1]->end < $code_point);
3135
3136 # Here the cache doesn't yield the correct $i. Try adding 1.
3137 if ($i < $range_list_size - 1
3138 && $r->[$i]->end < $code_point &&
3139 $code_point <= $r->[$i+1]->end)
3140 {
3141 $i++;
3142 trace "next \$i is correct: $i" if main::DEBUG && $to_trace;
3143 $_search_ranges_cache{$addr} = $i;
3144 return $i;
3145 }
3146
3147 # Here, adding 1 also didn't work. We do a binary search to
3148 # find the correct position, starting with current $i
3149 my $lower = 0;
3150 my $upper = $range_list_size - 1;
3151 while (1) {
3152 trace "top of loop i=$i:", sprintf("%04X", $r->[$lower]->start), "[$lower] .. ", sprintf("%04X", $r->[$i]->start), "[$i] .. ", sprintf("%04X", $r->[$upper]->start), "[$upper]" if main::DEBUG && $to_trace;
3153
3154 if ($code_point <= $r->[$i]->end) {
3155
3156 # Here we have met the upper constraint. We can quit if we
3157 # also meet the lower one.
3158 last if $i == 0 || $r->[$i-1]->end < $code_point;
3159
3160 $upper = $i; # Still too high.
3161
3162 }
3163 else {
3164
3165 # Here, $r[$i]->end < $code_point, so look higher up.
3166 $lower = $i;
3167 }
3168
3169 # Split search domain in half to try again.
3170 my $temp = ($upper + $lower) / 2;
3171
3172 # No point in continuing unless $i changes for next time
3173 # in the loop.
3174 if ($temp == $i) {
3175
3176 # We can't reach the highest element because of the averaging.
3177 # So if one below the upper edge, force it there and try one
3178 # more time.
3179 if ($i == $range_list_size - 2) {
3180
3181 trace "Forcing to upper edge" if main::DEBUG && $to_trace;
3182 $i = $range_list_size - 1;
3183
3184 # Change $lower as well so if fails next time through,
3185 # taking the average will yield the same $i, and we will
3186 # quit with the error message just below.
3187 $lower = $i;
3188 next;
3189 }
3190 Carp::my_carp_bug("$owner_name_of{$addr}Can't find where the range ought to go. No action taken.");
3191 return;
3192 }
3193 $i = $temp;
3194 } # End of while loop
3195
3196 if (main::DEBUG && $to_trace) {
3197 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i;
3198 trace "i= [ $i ]", $r->[$i];
3199 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < $range_list_size - 1;
3200 }
3201
3202 # Here we have found the offset. Cache it as a starting point for the
3203 # next call.
3204 $_search_ranges_cache{$addr} = $i;
3205 return $i;
3206 }
3207
3208 sub _add_delete {
3209 # Add, replace or delete ranges to or from a list. The $type
3210 # parameter gives which:
3211 # '+' => insert or replace a range, returning a list of any changed
3212 # ranges.
3213 # '-' => delete a range, returning a list of any deleted ranges.
3214 #
3215 # The next three parameters give respectively the start, end, and
3216 # value associated with the range. 'value' should be null unless the
3217 # operation is '+';
3218 #
3219 # The range list is kept sorted so that the range with the lowest
3220 # starting position is first in the list, and generally, adjacent
c1739a4a 3221 # ranges with the same values are merged into a single larger one (see
99870f4d
KW
3222 # exceptions below).
3223 #
c1739a4a 3224 # There are more parameters; all are key => value pairs:
99870f4d
KW
3225 # Type gives the type of the value. It is only valid for '+'.
3226 # All ranges have types; if this parameter is omitted, 0 is
3227 # assumed. Ranges with type 0 are assumed to obey the
3228 # Unicode rules for casing, etc; ranges with other types are
3229 # not. Otherwise, the type is arbitrary, for the caller's
3230 # convenience, and looked at only by this routine to keep
3231 # adjacent ranges of different types from being merged into
3232 # a single larger range, and when Replace =>
3233 # $IF_NOT_EQUIVALENT is specified (see just below).
3234 # Replace determines what to do if the range list already contains
3235 # ranges which coincide with all or portions of the input
3236 # range. It is only valid for '+':
3237 # => $NO means that the new value is not to replace
3238 # any existing ones, but any empty gaps of the
3239 # range list coinciding with the input range
3240 # will be filled in with the new value.
3241 # => $UNCONDITIONALLY means to replace the existing values with
3242 # this one unconditionally. However, if the
3243 # new and old values are identical, the
3244 # replacement is skipped to save cycles
3245 # => $IF_NOT_EQUIVALENT means to replace the existing values
3246 # with this one if they are not equivalent.
3247 # Ranges are equivalent if their types are the
c1739a4a 3248 # same, and they are the same string; or if
99870f4d
KW
3249 # both are type 0 ranges, if their Unicode
3250 # standard forms are identical. In this last
3251 # case, the routine chooses the more "modern"
3252 # one to use. This is because some of the
3253 # older files are formatted with values that
3254 # are, for example, ALL CAPs, whereas the
3255 # derived files have a more modern style,
3256 # which looks better. By looking for this
3257 # style when the pre-existing and replacement
3258 # standard forms are the same, we can move to
3259 # the modern style
3260 # => $MULTIPLE means that if this range duplicates an
3261 # existing one, but has a different value,
3262 # don't replace the existing one, but insert
3263 # this, one so that the same range can occur
53d84487
KW
3264 # multiple times. They are stored LIFO, so
3265 # that the final one inserted is the first one
3266 # returned in an ordered search of the table.
99870f4d
KW
3267 # => anything else is the same as => $IF_NOT_EQUIVALENT
3268 #
c1739a4a
KW
3269 # "same value" means identical for non-type-0 ranges, and it means
3270 # having the same standard forms for type-0 ranges.
99870f4d
KW
3271
3272 return Carp::carp_too_few_args(\@_, 5) if main::DEBUG && @_ < 5;
3273
3274 my $self = shift;
3275 my $operation = shift; # '+' for add/replace; '-' for delete;
3276 my $start = shift;
3277 my $end = shift;
3278 my $value = shift;
3279
3280 my %args = @_;
3281
3282 $value = "" if not defined $value; # warning: $value can be "0"
3283
3284 my $replace = delete $args{'Replace'};
3285 $replace = $IF_NOT_EQUIVALENT unless defined $replace;
3286
3287 my $type = delete $args{'Type'};
3288 $type = 0 unless defined $type;
3289
3290 Carp::carp_extra_args(\%args) if main::DEBUG && %args;
3291
ffe43484 3292 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3293
3294 if ($operation ne '+' && $operation ne '-') {
3295 Carp::my_carp_bug("$owner_name_of{$addr}First parameter to _add_delete must be '+' or '-'. No action taken.");
3296 return;
3297 }
3298 unless (defined $start && defined $end) {
3299 Carp::my_carp_bug("$owner_name_of{$addr}Undefined start and/or end to _add_delete. No action taken.");
3300 return;
3301 }
3302 unless ($end >= $start) {
3303 Carp::my_carp_bug("$owner_name_of{$addr}End of range (" . sprintf("%04X", $end) . ") must not be before start (" . sprintf("%04X", $start) . "). No action taken.");
3304 return;
3305 }
3306 #local $to_trace = 1 if main::DEBUG;
3307
3308 if ($operation eq '-') {
3309 if ($replace != $IF_NOT_EQUIVALENT) {
3310 Carp::my_carp_bug("$owner_name_of{$addr}Replace => \$IF_NOT_EQUIVALENT is required when deleting a range from a range list. Assuming Replace => \$IF_NOT_EQUIVALENT.");
3311 $replace = $IF_NOT_EQUIVALENT;
3312 }
3313 if ($type) {
3314 Carp::my_carp_bug("$owner_name_of{$addr}Type => 0 is required when deleting a range from a range list. Assuming Type => 0.");
3315 $type = 0;
3316 }
3317 if ($value ne "") {
3318 Carp::my_carp_bug("$owner_name_of{$addr}Value => \"\" is required when deleting a range from a range list. Assuming Value => \"\".");
3319 $value = "";
3320 }
3321 }
3322
3323 my $r = $ranges{$addr}; # The current list of ranges
3324 my $range_list_size = scalar @$r; # And its size
3325 my $max = $max{$addr}; # The current high code point in
3326 # the list of ranges
3327
3328 # Do a special case requiring fewer machine cycles when the new range
3329 # starts after the current highest point. The Unicode input data is
3330 # structured so this is common.
3331 if ($start > $max) {
3332
3333 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) type=$type" if main::DEBUG && $to_trace;
3334 return if $operation eq '-'; # Deleting a non-existing range is a
3335 # no-op
3336
3337 # If the new range doesn't logically extend the current final one
3338 # in the range list, create a new range at the end of the range
3339 # list. (max cleverly is initialized to a negative number not
3340 # adjacent to 0 if the range list is empty, so even adding a range
3341 # to an empty range list starting at 0 will have this 'if'
3342 # succeed.)
3343 if ($start > $max + 1 # non-adjacent means can't extend.
3344 || @{$r}[-1]->value ne $value # values differ, can't extend.
3345 || @{$r}[-1]->type != $type # types differ, can't extend.
3346 ) {
3347 push @$r, Range->new($start, $end,
3348 Value => $value,
3349 Type => $type);
3350 }
3351 else {
3352
3353 # Here, the new range starts just after the current highest in
3354 # the range list, and they have the same type and value.
3355 # Extend the current range to incorporate the new one.
3356 @{$r}[-1]->set_end($end);
3357 }
3358
3359 # This becomes the new maximum.
3360 $max{$addr} = $end;
3361
3362 return;
3363 }
3364 #local $to_trace = 0 if main::DEBUG;
3365
3366 trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) replace=$replace" if main::DEBUG && $to_trace;
3367
3368 # Here, the input range isn't after the whole rest of the range list.
3369 # Most likely 'splice' will be needed. The rest of the routine finds
3370 # the needed splice parameters, and if necessary, does the splice.
3371 # First, find the offset parameter needed by the splice function for
3372 # the input range. Note that the input range may span multiple
3373 # existing ones, but we'll worry about that later. For now, just find
3374 # the beginning. If the input range is to be inserted starting in a
3375 # position not currently in the range list, it must (obviously) come
3376 # just after the range below it, and just before the range above it.
3377 # Slightly less obviously, it will occupy the position currently
3378 # occupied by the range that is to come after it. More formally, we
3379 # are looking for the position, $i, in the array of ranges, such that:
3380 #
3381 # r[$i-1]->start <= r[$i-1]->end < $start < r[$i]->start <= r[$i]->end
3382 #
3383 # (The ordered relationships within existing ranges are also shown in
3384 # the equation above). However, if the start of the input range is
3385 # within an existing range, the splice offset should point to that
3386 # existing range's position in the list; that is $i satisfies a
3387 # somewhat different equation, namely:
3388 #
3389 #r[$i-1]->start <= r[$i-1]->end < r[$i]->start <= $start <= r[$i]->end
3390 #
3391 # More briefly, $start can come before or after r[$i]->start, and at
3392 # this point, we don't know which it will be. However, these
3393 # two equations share these constraints:
3394 #
3395 # r[$i-1]->end < $start <= r[$i]->end
3396 #
3397 # And that is good enough to find $i.
3398
3399 my $i = $self->_search_ranges($start);
3400 if (! defined $i) {
3401 Carp::my_carp_bug("Searching $self for range beginning with $start unexpectedly returned undefined. Operation '$operation' not performed");
3402 return;
3403 }
3404
3405 # The search function returns $i such that:
3406 #
3407 # r[$i-1]->end < $start <= r[$i]->end
3408 #
3409 # That means that $i points to the first range in the range list
3410 # that could possibly be affected by this operation. We still don't
3411 # know if the start of the input range is within r[$i], or if it
3412 # points to empty space between r[$i-1] and r[$i].
3413 trace "[$i] is the beginning splice point. Existing range there is ", $r->[$i] if main::DEBUG && $to_trace;
3414
3415 # Special case the insertion of data that is not to replace any
3416 # existing data.
3417 if ($replace == $NO) { # If $NO, has to be operation '+'
3418 #local $to_trace = 1 if main::DEBUG;
3419 trace "Doesn't replace" if main::DEBUG && $to_trace;
3420
3421 # Here, the new range is to take effect only on those code points
3422 # that aren't already in an existing range. This can be done by
3423 # looking through the existing range list and finding the gaps in
3424 # the ranges that this new range affects, and then calling this
3425 # function recursively on each of those gaps, leaving untouched
3426 # anything already in the list. Gather up a list of the changed
3427 # gaps first so that changes to the internal state as new ranges
3428 # are added won't be a problem.
3429 my @gap_list;
3430
3431 # First, if the starting point of the input range is outside an
3432 # existing one, there is a gap from there to the beginning of the
3433 # existing range -- add a span to fill the part that this new
3434 # range occupies
3435 if ($start < $r->[$i]->start) {
3436 push @gap_list, Range->new($start,
3437 main::min($end,
3438 $r->[$i]->start - 1),
3439 Type => $type);
3440 trace "gap before $r->[$i] [$i], will add", $gap_list[-1] if main::DEBUG && $to_trace;
3441 }
3442
3443 # Then look through the range list for other gaps until we reach
3444 # the highest range affected by the input one.
3445 my $j;
3446 for ($j = $i+1; $j < $range_list_size; $j++) {
3447 trace "j=[$j]", $r->[$j] if main::DEBUG && $to_trace;
3448 last if $end < $r->[$j]->start;
3449
3450 # If there is a gap between when this range starts and the
3451 # previous one ends, add a span to fill it. Note that just
3452 # because there are two ranges doesn't mean there is a
3453 # non-zero gap between them. It could be that they have
3454 # different values or types
3455 if ($r->[$j-1]->end + 1 != $r->[$j]->start) {
3456 push @gap_list,
3457 Range->new($r->[$j-1]->end + 1,
3458 $r->[$j]->start - 1,
3459 Type => $type);
3460 trace "gap between $r->[$j-1] and $r->[$j] [$j], will add: $gap_list[-1]" if main::DEBUG && $to_trace;
3461 }
3462 }
3463
3464 # Here, we have either found an existing range in the range list,
3465 # beyond the area affected by the input one, or we fell off the
3466 # end of the loop because the input range affects the whole rest
3467 # of the range list. In either case, $j is 1 higher than the
3468 # highest affected range. If $j == $i, it means that there are no
3469 # affected ranges, that the entire insertion is in the gap between
3470 # r[$i-1], and r[$i], which we already have taken care of before
3471 # the loop.
3472 # On the other hand, if there are affected ranges, it might be
3473 # that there is a gap that needs filling after the final such
3474 # range to the end of the input range
3475 if ($r->[$j-1]->end < $end) {
3476 push @gap_list, Range->new(main::max($start,
3477 $r->[$j-1]->end + 1),
3478 $end,
3479 Type => $type);
3480 trace "gap after $r->[$j-1], will add $gap_list[-1]" if main::DEBUG && $to_trace;
3481 }
3482
3483 # Call recursively to fill in all the gaps.
3484 foreach my $gap (@gap_list) {
3485 $self->_add_delete($operation,
3486 $gap->start,
3487 $gap->end,
3488 $value,
3489 Type => $type);
3490 }
3491
3492 return;
3493 }
3494
53d84487
KW
3495 # Here, we have taken care of the case where $replace is $NO.
3496 # Remember that here, r[$i-1]->end < $start <= r[$i]->end
3497 # If inserting a multiple record, this is where it goes, before the
3498 # first (if any) existing one. This implies an insertion, and no
3499 # change to any existing ranges. Note that $i can be -1 if this new
3500 # range doesn't actually duplicate any existing, and comes at the
3501 # beginning of the list.
3502 if ($replace == $MULTIPLE) {
3503
3504 if ($start != $end) {
3505 Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple record when the range ($start..$end) contains more than one code point. No action taken.");
3506 return;
3507 }
3508
3509 # Don't add an exact duplicate, as it isn't really a multiple
3510 if ($end >= $r->[$i]->start) {
1f6798c4
KW
3511 my $existing_value = $r->[$i]->value;
3512 my $existing_type = $r->[$i]->type;
3513 return if $value eq $existing_value && $type eq $existing_type;
3514
3515 # If the multiple value is part of an existing range, we want
3516 # to split up that range, so that only the single code point
3517 # is affected. To do this, we first call ourselves
3518 # recursively to delete that code point from the table, having
3519 # preserved its current data above. Then we call ourselves
3520 # recursively again to add the new multiple, which we know by
3521 # the test just above is different than the current code
3522 # point's value, so it will become a range containing a single
3523 # code point: just itself. Finally, we add back in the
3524 # pre-existing code point, which will again be a single code
3525 # point range. Because 'i' likely will have changed as a
3526 # result of these operations, we can't just continue on, but
3527 # do this operation recursively as well.
53d84487 3528 if ($r->[$i]->start != $r->[$i]->end) {
1f6798c4
KW
3529 $self->_add_delete('-', $start, $end, "");
3530 $self->_add_delete('+', $start, $end, $value, Type => $type);
3531 return $self->_add_delete('+', $start, $end, $existing_value, Type => $existing_type, Replace => $MULTIPLE);
53d84487 3532 }
53d84487
KW
3533 }
3534
3535 trace "Adding multiple record at $i with $start..$end, $value" if main::DEBUG && $to_trace;
3536 my @return = splice @$r,
3537 $i,
3538 0,
3539 Range->new($start,
3540 $end,
3541 Value => $value,
3542 Type => $type);
3543 if (main::DEBUG && $to_trace) {
3544 trace "After splice:";
3545 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3546 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3547 trace "i =[", $i, "]", $r->[$i] if $i >= 0;
3548 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3549 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3550 trace 'i+3=[', $i+3, ']', $r->[$i+3] if $i < @$r - 3;
3551 }
3552 return @return;
3553 }
3554
3555 # Here, we have taken care of $NO and $MULTIPLE replaces. This leaves
3556 # delete, insert, and replace either unconditionally or if not
3557 # equivalent. $i still points to the first potential affected range.
3558 # Now find the highest range affected, which will determine the length
3559 # parameter to splice. (The input range can span multiple existing
3560 # ones.) If this isn't a deletion, while we are looking through the
3561 # range list, see also if this is a replacement rather than a clean
3562 # insertion; that is if it will change the values of at least one
3563 # existing range. Start off assuming it is an insert, until find it
3564 # isn't.
3565 my $clean_insert = $operation eq '+';
99870f4d
KW
3566 my $j; # This will point to the highest affected range
3567
3568 # For non-zero types, the standard form is the value itself;
3569 my $standard_form = ($type) ? $value : main::standardize($value);
3570
3571 for ($j = $i; $j < $range_list_size; $j++) {
3572 trace "Looking for highest affected range; the one at $j is ", $r->[$j] if main::DEBUG && $to_trace;
3573
3574 # If find a range that it doesn't overlap into, we can stop
3575 # searching
3576 last if $end < $r->[$j]->start;
3577
969a34cc
KW
3578 # Here, overlaps the range at $j. If the values don't match,
3579 # and so far we think this is a clean insertion, it becomes a
3580 # non-clean insertion, i.e., a 'change' or 'replace' instead.
3581 if ($clean_insert) {
99870f4d 3582 if ($r->[$j]->standard_form ne $standard_form) {
969a34cc 3583 $clean_insert = 0;
56343c78
KW
3584 if ($replace == $CROAK) {
3585 main::croak("The range to add "
3586 . sprintf("%04X", $start)
3587 . '-'
3588 . sprintf("%04X", $end)
3589 . " with value '$value' overlaps an existing range $r->[$j]");
3590 }
99870f4d
KW
3591 }
3592 else {
3593
3594 # Here, the two values are essentially the same. If the
3595 # two are actually identical, replacing wouldn't change
3596 # anything so skip it.
3597 my $pre_existing = $r->[$j]->value;
3598 if ($pre_existing ne $value) {
3599
3600 # Here the new and old standardized values are the
3601 # same, but the non-standardized values aren't. If
3602 # replacing unconditionally, then replace
3603 if( $replace == $UNCONDITIONALLY) {
969a34cc 3604 $clean_insert = 0;
99870f4d
KW
3605 }
3606 else {
3607
3608 # Here, are replacing conditionally. Decide to
3609 # replace or not based on which appears to look
3610 # the "nicest". If one is mixed case and the
3611 # other isn't, choose the mixed case one.
3612 my $new_mixed = $value =~ /[A-Z]/
3613 && $value =~ /[a-z]/;
3614 my $old_mixed = $pre_existing =~ /[A-Z]/
3615 && $pre_existing =~ /[a-z]/;
3616
3617 if ($old_mixed != $new_mixed) {
969a34cc 3618 $clean_insert = 0 if $new_mixed;
99870f4d 3619 if (main::DEBUG && $to_trace) {
969a34cc
KW
3620 if ($clean_insert) {
3621 trace "Retaining $pre_existing over $value";
99870f4d
KW
3622 }
3623 else {
969a34cc 3624 trace "Replacing $pre_existing with $value";
99870f4d
KW
3625 }
3626 }
3627 }
3628 else {
3629
3630 # Here casing wasn't different between the two.
3631 # If one has hyphens or underscores and the
3632 # other doesn't, choose the one with the
3633 # punctuation.
3634 my $new_punct = $value =~ /[-_]/;
3635 my $old_punct = $pre_existing =~ /[-_]/;
3636
3637 if ($old_punct != $new_punct) {
969a34cc 3638 $clean_insert = 0 if $new_punct;
99870f4d 3639 if (main::DEBUG && $to_trace) {
969a34cc
KW
3640 if ($clean_insert) {
3641 trace "Retaining $pre_existing over $value";
99870f4d
KW
3642 }
3643 else {
969a34cc 3644 trace "Replacing $pre_existing with $value";
99870f4d
KW
3645 }
3646 }
3647 } # else existing one is just as "good";
3648 # retain it to save cycles.
3649 }
3650 }
3651 }
3652 }
3653 }
3654 } # End of loop looking for highest affected range.
3655
3656 # Here, $j points to one beyond the highest range that this insertion
3657 # affects (hence to beyond the range list if that range is the final
3658 # one in the range list).
3659
3660 # The splice length is all the affected ranges. Get it before
3661 # subtracting, for efficiency, so we don't have to later add 1.
3662 my $length = $j - $i;
3663
3664 $j--; # $j now points to the highest affected range.
3665 trace "Final affected range is $j: $r->[$j]" if main::DEBUG && $to_trace;
3666
99870f4d
KW
3667 # Here, have taken care of $NO and $MULTIPLE replaces.
3668 # $j points to the highest affected range. But it can be < $i or even
3669 # -1. These happen only if the insertion is entirely in the gap
3670 # between r[$i-1] and r[$i]. Here's why: j < i means that the j loop
3671 # above exited first time through with $end < $r->[$i]->start. (And
3672 # then we subtracted one from j) This implies also that $start <
3673 # $r->[$i]->start, but we know from above that $r->[$i-1]->end <
3674 # $start, so the entire input range is in the gap.
3675 if ($j < $i) {
3676
3677 # Here the entire input range is in the gap before $i.
3678
3679 if (main::DEBUG && $to_trace) {
3680 if ($i) {
3681 trace "Entire range is between $r->[$i-1] and $r->[$i]";
3682 }
3683 else {
3684 trace "Entire range is before $r->[$i]";
3685 }
3686 }
3687 return if $operation ne '+'; # Deletion of a non-existent range is
3688 # a no-op
3689 }
3690 else {
3691
969a34cc
KW
3692 # Here part of the input range is not in the gap before $i. Thus,
3693 # there is at least one affected one, and $j points to the highest
3694 # such one.
99870f4d
KW
3695
3696 # At this point, here is the situation:
3697 # This is not an insertion of a multiple, nor of tentative ($NO)
3698 # data.
3699 # $i points to the first element in the current range list that
3700 # may be affected by this operation. In fact, we know
3701 # that the range at $i is affected because we are in
3702 # the else branch of this 'if'
3703 # $j points to the highest affected range.
3704 # In other words,
3705 # r[$i-1]->end < $start <= r[$i]->end
3706 # And:
3707 # r[$i-1]->end < $start <= $end <= r[$j]->end
3708 #
3709 # Also:
969a34cc
KW
3710 # $clean_insert is a boolean which is set true if and only if
3711 # this is a "clean insertion", i.e., not a change nor a
3712 # deletion (multiple was handled above).
99870f4d
KW
3713
3714 # We now have enough information to decide if this call is a no-op
969a34cc
KW
3715 # or not. It is a no-op if this is an insertion of already
3716 # existing data.
99870f4d 3717
969a34cc 3718 if (main::DEBUG && $to_trace && $clean_insert
99870f4d
KW
3719 && $i == $j
3720 && $start >= $r->[$i]->start)
3721 {
3722 trace "no-op";
3723 }
969a34cc 3724 return if $clean_insert
99870f4d
KW
3725 && $i == $j # more than one affected range => not no-op
3726
3727 # Here, r[$i-1]->end < $start <= $end <= r[$i]->end
3728 # Further, $start and/or $end is >= r[$i]->start
3729 # The test below hence guarantees that
3730 # r[$i]->start < $start <= $end <= r[$i]->end
3731 # This means the input range is contained entirely in
3732 # the one at $i, so is a no-op
3733 && $start >= $r->[$i]->start;
3734 }
3735
3736 # Here, we know that some action will have to be taken. We have
3737 # calculated the offset and length (though adjustments may be needed)
3738 # for the splice. Now start constructing the replacement list.
3739 my @replacement;
3740 my $splice_start = $i;
3741
3742 my $extends_below;
3743 my $extends_above;
3744
3745 # See if should extend any adjacent ranges.
3746 if ($operation eq '-') { # Don't extend deletions
3747 $extends_below = $extends_above = 0;
3748 }
3749 else { # Here, should extend any adjacent ranges. See if there are
3750 # any.
3751 $extends_below = ($i > 0
3752 # can't extend unless adjacent
3753 && $r->[$i-1]->end == $start -1
3754 # can't extend unless are same standard value
3755 && $r->[$i-1]->standard_form eq $standard_form
3756 # can't extend unless share type
3757 && $r->[$i-1]->type == $type);
3758 $extends_above = ($j+1 < $range_list_size
3759 && $r->[$j+1]->start == $end +1
3760 && $r->[$j+1]->standard_form eq $standard_form
23822bda 3761 && $r->[$j+1]->type == $type);
99870f4d
KW
3762 }
3763 if ($extends_below && $extends_above) { # Adds to both
3764 $splice_start--; # start replace at element below
3765 $length += 2; # will replace on both sides
3766 trace "Extends both below and above ranges" if main::DEBUG && $to_trace;
3767
3768 # The result will fill in any gap, replacing both sides, and
3769 # create one large range.
3770 @replacement = Range->new($r->[$i-1]->start,
3771 $r->[$j+1]->end,
3772 Value => $value,
3773 Type => $type);
3774 }
3775 else {
3776
3777 # Here we know that the result won't just be the conglomeration of
3778 # a new range with both its adjacent neighbors. But it could
3779 # extend one of them.
3780
3781 if ($extends_below) {
3782
3783 # Here the new element adds to the one below, but not to the
3784 # one above. If inserting, and only to that one range, can
3785 # just change its ending to include the new one.
969a34cc 3786 if ($length == 0 && $clean_insert) {
99870f4d
KW
3787 $r->[$i-1]->set_end($end);
3788 trace "inserted range extends range to below so it is now $r->[$i-1]" if main::DEBUG && $to_trace;
3789 return;
3790 }
3791 else {
3792 trace "Changing inserted range to start at ", sprintf("%04X", $r->[$i-1]->start), " instead of ", sprintf("%04X", $start) if main::DEBUG && $to_trace;
3793 $splice_start--; # start replace at element below
3794 $length++; # will replace the element below
3795 $start = $r->[$i-1]->start;
3796 }
3797 }
3798 elsif ($extends_above) {
3799
3800 # Here the new element adds to the one above, but not below.
3801 # Mirror the code above
969a34cc 3802 if ($length == 0 && $clean_insert) {
99870f4d
KW
3803 $r->[$j+1]->set_start($start);
3804 trace "inserted range extends range to above so it is now $r->[$j+1]" if main::DEBUG && $to_trace;
3805 return;
3806 }
3807 else {
3808 trace "Changing inserted range to end at ", sprintf("%04X", $r->[$j+1]->end), " instead of ", sprintf("%04X", $end) if main::DEBUG && $to_trace;
3809 $length++; # will replace the element above
3810 $end = $r->[$j+1]->end;
3811 }
3812 }
3813
3814 trace "Range at $i is $r->[$i]" if main::DEBUG && $to_trace;
3815
3816 # Finally, here we know there will have to be a splice.
3817 # If the change or delete affects only the highest portion of the
3818 # first affected range, the range will have to be split. The
3819 # splice will remove the whole range, but will replace it by a new
3820 # range containing just the unaffected part. So, in this case,
3821 # add to the replacement list just this unaffected portion.
3822 if (! $extends_below
3823 && $start > $r->[$i]->start && $start <= $r->[$i]->end)
3824 {
3825 push @replacement,
3826 Range->new($r->[$i]->start,
3827 $start - 1,
3828 Value => $r->[$i]->value,
3829 Type => $r->[$i]->type);
3830 }
3831
3832 # In the case of an insert or change, but not a delete, we have to
3833 # put in the new stuff; this comes next.
3834 if ($operation eq '+') {
3835 push @replacement, Range->new($start,
3836 $end,
3837 Value => $value,
3838 Type => $type);
3839 }
3840
3841 trace "Range at $j is $r->[$j]" if main::DEBUG && $to_trace && $j != $i;
3842 #trace "$end >=", $r->[$j]->start, " && $end <", $r->[$j]->end if main::DEBUG && $to_trace;
3843
3844 # And finally, if we're changing or deleting only a portion of the
3845 # highest affected range, it must be split, as the lowest one was.
3846 if (! $extends_above
3847 && $j >= 0 # Remember that j can be -1 if before first
3848 # current element
3849 && $end >= $r->[$j]->start
3850 && $end < $r->[$j]->end)
3851 {
3852 push @replacement,
3853 Range->new($end + 1,
3854 $r->[$j]->end,
3855 Value => $r->[$j]->value,
3856 Type => $r->[$j]->type);
3857 }
3858 }
3859
3860 # And do the splice, as calculated above
3861 if (main::DEBUG && $to_trace) {
3862 trace "replacing $length element(s) at $i with ";
3863 foreach my $replacement (@replacement) {
3864 trace " $replacement";
3865 }
3866 trace "Before splice:";
3867 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3868 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3869 trace "i =[", $i, "]", $r->[$i];
3870 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3871 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
3872 }
3873
3874 my @return = splice @$r, $splice_start, $length, @replacement;
3875
3876 if (main::DEBUG && $to_trace) {
3877 trace "After splice:";
3878 trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
3879 trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
3880 trace "i =[", $i, "]", $r->[$i];
3881 trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
3882 trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
e6451557 3883 trace "removed ", @return if @return;
99870f4d
KW
3884 }
3885
3886 # An actual deletion could have changed the maximum in the list.
3887 # There was no deletion if the splice didn't return something, but
3888 # otherwise recalculate it. This is done too rarely to worry about
3889 # performance.
3890 if ($operation eq '-' && @return) {
3891 $max{$addr} = $r->[-1]->end;
3892 }
3893 return @return;
3894 }
3895
3896 sub reset_each_range { # reset the iterator for each_range();
3897 my $self = shift;
3898 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3899
f998e60c 3900 no overloading;
051df77b 3901 undef $each_range_iterator{pack 'J', $self};
99870f4d
KW
3902 return;
3903 }
3904
3905 sub each_range {
3906 # Iterate over each range in a range list. Results are undefined if
3907 # the range list is changed during the iteration.
3908
3909 my $self = shift;
3910 Carp::carp_extra_args(\@_) if main::DEBUG && @_;
3911
ffe43484 3912 my $addr = do { no overloading; pack 'J', $self; };
99870f4d
KW
3913
3914 return if $self->is_empty;
3915
3916 $each_range_iterator{$addr} = -1
3917 if ! defined $each_range_iterator{$addr};