1ENC2XS(1) Perl Programmers Reference Guide ENC2XS(1)
2
6 enc2xs -- Perl Encode Module Generator
7
9 enc2xs -[options]
10 enc2xs -M ModName mapfiles...
11 enc2xs -C
12
14 enc2xs builds a Perl extension for use by Encode from either Unicode
15 Character Mapping files (.ucm) or Tcl Encoding Files (.enc). Besides
16 being used internally during the build process of the Encode module,
17 you can use enc2xs to add your own encoding to perl. No knowledge of
18 XS is necessary.
19
21 If you want to know as little about Perl as possible but need to add a
22 new encoding, just read this chapter and forget the rest.
23 0. Have a .ucm file ready. You can get it from somewhere or you can
25 write your own from scratch or you can grab one from the Encode
26 distribution and customize it. For the UCM format, see the next
27 Chapter. In the example below, I'll call my theoretical encoding
28 myascii, defined in my.ucm. "$" is a shell prompt.
29 $ ls -F
31 my.ucm
32 1. Issue a command as follows;
34 $ enc2xs -M My my.ucm
36 generating Makefile.PL
37 generating My.pm
38 generating README
39 generating Changes
40 Now take a look at your current directory. It should look like
42 this.
43 $ ls -F
45 Makefile.PL My.pm my.ucm t/
46 The following files were created.
48 Makefile.PL - MakeMaker script
50 My.pm - Encode submodule
51 t/My.t - test file
52 1.1.
54 If you want *.ucm installed together with the modules, do as
55 follows;
56 $ mkdir Encode
58 $ mv *.ucm Encode
59 $ enc2xs -M My Encode/*ucm
60 2. Edit the files generated. You don't have to if you have no time
62 AND no intention to give it to someone else. But it is a good idea
63 to edit the pod and to add more tests.
64 3. Now issue a command all Perl Mongers love:
66 $ perl Makefile.PL
68 Writing Makefile for Encode::My
69 4. Now all you have to do is make.
71 $ make
73 cp My.pm blib/lib/Encode/My.pm
74 /usr/local/bin/perl /usr/local/bin/enc2xs -Q -O \
75 -o encode_t.c -f encode_t.fnm
76 Reading myascii (myascii)
77 Writing compiled form
78 128 bytes in string tables
79 384 bytes (75%) saved spotting duplicates
80 1 bytes (0.775%) saved using substrings
81 ....
82 chmod 644 blib/arch/auto/Encode/My/My.bs
83 $
84 The time it takes varies depending on how fast your machine is and
86 how large your encoding is. Unless you are working on something
87 big like euc-tw, it won't take too long.
88 5. You can "make install" already but you should test first.
90 $ make test
92 PERL_DL_NONLAZY=1 /usr/local/bin/perl -Iblib/arch -Iblib/lib \
93 -e 'use Test::Harness qw(&runtests $verbose); \
94 $verbose=0; runtests @ARGV;' t/*.t
95 t/My....ok
96 All tests successful.
97 Files=1, Tests=2, 0 wallclock secs
98 ( 0.09 cusr + 0.01 csys = 0.09 CPU)
99 6. If you are content with the test result, just "make install"
101 7. If you want to add your encoding to Encode's demand-loading list
103 (so you don't have to "use Encode::YourEncoding"), run
104 enc2xs -C
106 to update Encode::ConfigLocal, a module that controls local set‐
108 tings. After that, "use Encode;" is enough to load your encodings
109 on demand.
110
112 Encode uses the Unicode Character Map (UCM) format for source character
113 mappings. This format is used by IBM's ICU package and was adopted by
114 Nick Ing-Simmons for use with the Encode module. Since UCM is more
115 flexible than Tcl's Encoding Map and far more user-friendly, this is
116 the recommended formet for Encode now.
117 A UCM file looks like this.
119 #
121 # Comments
122 #
123 <code_set_name> "US-ascii" # Required
124 <code_set_alias> "ascii" # Optional
125 <mb_cur_min> 1 # Required; usually 1
126 <mb_cur_max> 1 # Max. # of bytes/char
127 <subchar> \x3F # Substitution char
128 #
129 CHARMAP
130 <U0000> \x00 ⎪0 # <control>
131 <U0001> \x01 ⎪0 # <control>
132 <U0002> \x02 ⎪0 # <control>
133 ....
134 <U007C> \x7C ⎪0 # VERTICAL LINE
135 <U007D> \x7D ⎪0 # RIGHT CURLY BRACKET
136 <U007E> \x7E ⎪0 # TILDE
137 <U007F> \x7F ⎪0 # <control>
138 END CHARMAP
139 · Anything that follows "#" is treated as a comment.
141 · The header section continues until a line containing the word
143 CHARMAP. This section has a form of <keyword> value, one pair per
144 line. Strings used as values must be quoted. Barewords are treated
145 as numbers. \xXX represents a byte.
146 Most of the keywords are self-explanatory. subchar means substitu‐
148 tion character, not subcharacter. When you decode a Unicode
149 sequence to this encoding but no matching character is found, the
150 byte sequence defined here will be used. For most cases, the value
151 here is \x3F; in ASCII, this is a question mark.
152 · CHARMAP starts the character map section. Each line has a form as
154 follows:
155 <UXXXX> \xXX.. ⎪0 # comment
157 ^ ^ ^
158 ⎪ ⎪ +- Fallback flag
159 ⎪ +-------- Encoded byte sequence
160 +-------------- Unicode Character ID in hex
161 The format is roughly the same as a header section except for the
163 fallback flag: ⎪ followed by 0..3. The meaning of the possible
164 values is as follows:
165 ⎪0 Round trip safe. A character decoded to Unicode encodes back
167 to the same byte sequence. Most characters have this flag.
168 ⎪1 Fallback for unicode -> encoding. When seen, enc2xs adds this
170 character for the encode map only.
171 ⎪2 Skip sub-char mapping should there be no code point.
173 ⎪3 Fallback for encoding -> unicode. When seen, enc2xs adds this
175 character for the decode map only.
176 · And finally, END OF CHARMAP ends the section.
178 When you are manually creating a UCM file, you should copy ascii.ucm or
180 an existing encoding which is close to yours, rather than write your
181 own from scratch.
182 When you do so, make sure you leave at least U0000 to U0020 as is,
184 unless your environment is EBCDIC.
185 CAVEAT: not all features in UCM are implemented. For example,
187 icu:state is not used. Because of that, you need to write a perl mod‐
188 ule if you want to support algorithmical encodings, notably the
189 ISO-2022 series. Such modules include Encode::JP::2022_JP,
190 Encode::KR::2022_KR, and Encode::TW::HZ.
191 Coping with duplicate mappings
193 When you create a map, you SHOULD make your mappings round-trip safe.
195 That is, "encode('your-encoding', decode('your-encoding', $data)) eq
196 $data" stands for all characters that are marked as "⎪0". Here is how
197 to make sure:
198 · Sort your map in Unicode order.
200 · When you have a duplicate entry, mark either one with '⎪1' or '⎪3'.
202 · And make sure the '⎪1' or '⎪3' entry FOLLOWS the '⎪0' entry.
204 Here is an example from big5-eten.
206 <U2550> \xF9\xF9 ⎪0
208 <U2550> \xA2\xA4 ⎪3
209 Internally Encoding -> Unicode and Unicode -> Encoding Map looks like
211 this;
212 E to U U to E
214 --------------------------------------
215 \xF9\xF9 => U2550 U2550 => \xF9\xF9
216 \xA2\xA4 => U2550
217 So it is round-trip safe for \xF9\xF9. But if the line above is upside
219 down, here is what happens.
220 E to U U to E
222 --------------------------------------
223 \xA2\xA4 => U2550 U2550 => \xF9\xF9
224 (\xF9\xF9 => U2550 is now overwritten!)
225 The Encode package comes with ucmlint, a crude but sufficient utility
227 to check the integrity of a UCM file. Check under the Encode/bin
228 directory for this.
229 When in doubt, you can use ucmsort, yet another utility under
231 Encode/bin directory.
232
234 · ICU Home Page <http://oss.software.ibm.com/icu/>
235 · ICU Character Mapping Tables <http://oss.soft‐
237 ware.ibm.com/icu/charset/>
238 · ICU:Conversion Data <http://oss.software.ibm.com/icu/userguide/con‐
240 version-data.html>
241
243 Encode, perlmod, perlpod
244perl v5.8.8 2008-05-05 ENC2XS(1)