S_utf16_textfilter() was not returning EOF correctly in some situations.
[p5sagit/p5-mst-13.2.git] / lib / unicore / README.perl
1 The *.txt files were copied from
2
3         http://www.unicode.org/Public/5.1.0/ucd
4
5 and subdirectories 'extracted' and 'auxiliary' as of Unicode 5.1.0 (March 2008).
6
7 The big file, Unihan.txt (28 MB, 5.8 MB zip) was not included due to space
8 considerations.  Also NOT included were any *.html files and *Test.txt files.
9
10 To be 8.3 filesystem friendly, the names of some of the input files have been
11 changed from the values that are in the Unicode DB:
12
13 mv PropertyValueAliases.txt PropValueAliases.txt
14 mv NamedSequencesProv.txt NamedSqProv.txt
15 mv DerivedAge.txt DAge.txt
16 mv DerivedCoreProperties.txt DCoreProperties.txt
17 mv DerivedNormalizationProps.txt DNormalizationProps.txt
18 mv extracted/DerivedBidiClass.txt extracted/DBidiClass.txt
19 mv extracted/DerivedBinaryProperties.txt extracted/DBinaryProperties.txt
20 mv extracted/DerivedCombiningClass.txt extracted/DCombiningClass.txt
21 mv extracted/DerivedDecompositionType.txt extracted/DDecompositionType.txt
22 mv extracted/DerivedEastAsianWidth.txt extracted/DEastAsianWidth.txt
23 mv extracted/DerivedGeneralCategory.txt extracted/DGeneralCategory.txt
24 mv extracted/DerivedJoiningGroup.txt extracted/DJoinGroup.txt
25 mv extracted/DerivedJoiningType.txt extracted/DJoinType.txt
26 mv extracted/DerivedLineBreak.txt extracted/DLineBreak.txt
27 mv extracted/DerivedNumericType.txt extracted/DNumType.txt
28 mv extracted/DerivedNumericValues.txt extracted/DNumValues.txt
29
30 The names of files, such as test files, that are not used by mktables are not
31 changed, and will not work correctly on 8.3 filesystems.
32
33 The file 'version' should exist and be a single line with the Unicode version,
34 like
35 5.1.0
36
37 NOTE: If you modify the input file set you should also run
38  
39     mktables -makelist
40     
41 which will recreate the mktables.lst file which is used to speed up
42 the build process.    
43
44 FOR PUMPKINS
45
46 The files are inter-related.  If you take the latest UnicodeData.txt, for example,
47 but leave the older versions of other files, there can be subtle problems.
48
49 The *.pl files are generated from the *.txt files by the mktables script,
50 more recently done during the Perl build process, but if you want to try
51 the old manual way:
52         
53         cd lib/unicore
54         cp .../UnicodeOriginal/*.txt .
55         rm NormalizationTest.txt Unihan.txt Derived*.txt
56         p4 edit Properties *.pl */*.pl
57         perl ./mktables
58         p4 revert -a
59         cd ../..
60         perl Porting/manicheck
61
62 You need to update version by hand
63
64         p4 edit version
65         ...
66         
67 If any new (or deleted, unlikely but not impossible) *.pl files are indicated:
68
69         cd lib/unicore
70         p4 add ...
71         p4 delete ...
72         cd ../...
73         p4 edit MANIFEST
74         ...
75
76 And finally:
77
78         p4 submit
79
80 -- 
81 jhi@iki.fi; updated by nick@ccl4.org