organicmaps/icu
13
Fork 0
mirror of https://github.com/unicode-org/icu.git synced 2025-04-08 23:10:40 +00:00
Code Issues 1 Projects Releases Wiki Activity Actions 1

ICU-8972 preparseucd.py writes pnames_data.h with formatting nearly identical to genpname/preparse.pl; delete obsolete genpname/*

Browse source 
X-SVN-Rev: 31160
This commit is contained in:
Markus Scherer 2011-12-20 07:43:38 +00:00
parent 6bb13512c1
commit cc68c8c6e2
10 changed files with 172 additions and 2187 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

13
.gitignore vendored
View file

@ -1024,19 +1024,6 @@ tools/unicode/c/gennorm/gennorm.vcproj.*.*.user
tools/unicode/c/gennorm/release
tools/unicode/c/gennorm/x64
tools/unicode/c/gennorm/x86
tools/unicode/c/genpname/*.d
tools/unicode/c/genpname/*.o
tools/unicode/c/genpname/*.pdb
tools/unicode/c/genpname/*.plg
tools/unicode/c/genpname/Debug
tools/unicode/c/genpname/Makefile
tools/unicode/c/genpname/Release
tools/unicode/c/genpname/debug
tools/unicode/c/genpname/genpname
tools/unicode/c/genpname/genpname.vcproj.*.*.user
tools/unicode/c/genpname/release
tools/unicode/c/genpname/x64
tools/unicode/c/genpname/x86
tools/unicode/c/genprops/*.d
tools/unicode/c/genprops/*.ncb
tools/unicode/c/genprops/*.o

9
tools/unicode/c/genpname/CMakeLists.txt
View file

@ -1,9 +0,0 @@
# Copyright (C) 2010, International Business Machines
# Corporation and others. All Rights Reserved.
#
# created on: 2010jun03
# created by: Markus W. Scherer
# edited on: 2010jul20
# edited by: Stuart G. Gill
add_executable(genpname genpname.cpp)
target_link_libraries(genpname icuuc icutu)

97
tools/unicode/c/genpname/Makefile.in
View file

@ -1,97 +0,0 @@
## Makefile.in for ICU - tools/genpname
## Copyright (c) 1999-2005, International Business Machines Corporation and
## others. All Rights Reserved.
## Steven R. Loomis
## Source directory information
srcdir = @srcdir@
top_srcdir = @top_srcdir@
top_builddir = ../..
include $(top_builddir)/icudefs.mk
## Build directory information
subdir = tools/genpname
TARGET_STUB_NAME = genpname
SECTION = 8
MAN_FILES = $(TARGET_STUB_NAME).$(SECTION)
## Extra files to remove for 'make clean'
CLEANFILES = *~ $(DEPS)
## Target information
TARGET = $(BINDIR)/$(TARGET_STUB_NAME)$(EXEEXT)
ifneq ($(top_builddir),$(top_srcdir))
CPPFLAGS += -I$(top_builddir)/common
endif
CPPFLAGS += -I$(top_srcdir)/common -I$(srcdir)/../toolutil
LIBS = $(LIBICUTOOLUTIL) $(LIBICUI18N) $(LIBICUUC) $(DEFAULT_LIBS) $(LIB_M)
OBJECTS = genpname.o
DEPS = $(OBJECTS:.o=.d)
## List of phony targets
.PHONY : all all-local install install-local clean clean-local \
distclean distclean-local dist dist-local check check-local install-man
## Clear suffix list
.SUFFIXES :
## List of standard targets
all: all-local
install: install-local
clean: clean-local
distclean : distclean-local
dist: dist-local
check: all check-local
all-local: $(TARGET)
install-local: all-local
# $(MKINSTALLDIRS) $(DESTDIR)$(sbindir)
# $(INSTALL) $(TARGET) $(DESTDIR)$(sbindir)
install-man: $(MAN_FILES)
# $(MKINSTALLDIRS) $(DESTDIR)$(mandir)/man$(SECTION)
# $(INSTALL_DATA) $? $(DESTDIR)$(mandir)/man$(SECTION)
dist-local:
clean-local:
test -z "$(CLEANFILES)" || $(RMV) $(CLEANFILES)
$(RMV) $(TARGET) $(OBJECTS)
distclean-local: clean-local
$(RMV) Makefile
check-local: all-local
Makefile: $(srcdir)/Makefile.in $(top_builddir)/config.status
cd $(top_builddir) \
&& CONFIG_FILES=$(subdir)/$@ CONFIG_HEADERS= $(SHELL) ./config.status
$(TARGET) : $(OBJECTS)
$(LINK.cc) $(OUTOPT)$@ $^ $(LIBS)
$(POST_BUILD_STEP)
%.$(SECTION): $(srcdir)/%.$(SECTION).in
cd $(top_builddir) \
&& CONFIG_FILES=$(subdir)/$@ CONFIG_HEADERS= $(SHELL) ./config.status
ifeq (,$(MAKECMDGOALS))
-include $(DEPS)
else
ifneq ($(patsubst %clean,,$(MAKECMDGOALS)),)
-include $(DEPS)
endif
endif

52
tools/unicode/c/genpname/SyntheticPropertyAliases.txt
View file

@ -1,52 +0,0 @@
######################################################################
# Copyright (c) 2003-2011, International Business Machines
# Corporation and others. All Rights Reserved.
######################################################################
# Author: Alan Liu
# Created: February 20 2003
# Since: ICU 2.6
######################################################################
# This file follows the format of PropertyAliases.txt
# It contains synthetic property aliases not present
# in the UCD. Unlike PropertyAliases.txt, it should
# NOT contain a version number.
# ================================================
# ================================================
# Miscellaneous Properties
# ================================================
# ================================================
# Enumerated Properties
# ================================================
# lccc(c)=ccc(NFD(c)[0])
# tccc(c)=ccc(NFD(c)[last])
lccc; Lead_Canonical_Combining_Class
tccc; Trail_Canonical_Combining_Class
# ================================================
# Bitmask Properties
# ================================================
gcm ; General_Category_Mask
# ================================================
# Binary Properties
# ================================================
Sensitive ; Case_Sensitive
nfdinert; NFD_Inert
nfkdinert; NFKD_Inert
nfcinert; NFC_Inert
nfkcinert; NFKC_Inert
segstart; Segment_Starter
# C/POSIX character classes that do not have Unicode property [value] aliases
# see uchar.h
n/a; alnum
n/a; blank
n/a; graph
n/a; print
n/a; xdigit

85
tools/unicode/c/genpname/SyntheticPropertyValueAliases.txt
View file

@ -1,85 +0,0 @@
########################################################################
# Copyright (c) 2006-2011, International Business Machines
# Corporation and others. All Rights Reserved.
########################################################################
# file name: SyntheticPropertyValueAliases.txt
# encoding: US-ASCII
# tab size: 8 (not used)
# indentation: 4
# created by: gensvpa.pl
########################################################################
# This file follows the format of PropertyValueAliases.txt
# It contains synthetic property value aliases not present
# in the UCD. Unlike PropertyValueAliases.txt, it should
# NOT contain a version number.
########################################################################
# THIS FILE IS MACHINE-GENERATED, DON'T PLAY WITH IT IF YOU DON'T KNOW
# WHAT YOU ARE DOING, OTHERWISE VERY BAD THINGS WILL HAPPEN!
########################################################################
# set the same names as short and long names to fit the syntax without
# inventing names that we would have to support forever
# Script (sc)
sc ; Blis ; Blis
sc ; Cirt ; Cirt
sc ; Cyrs ; Cyrs
sc ; Egyd ; Egyd
sc ; Egyh ; Egyh
sc ; Geok ; Geok
sc ; Hans ; Hans
sc ; Hant ; Hant
sc ; Hmng ; Hmng
sc ; Hung ; Hung
sc ; Inds ; Inds
sc ; Jpan ; Jpan
sc ; Latf ; Latf
sc ; Latg ; Latg
sc ; Lina ; Lina
sc ; Maya ; Maya
sc ; Moon ; Moon
sc ; Perm ; Perm
sc ; Roro ; Roro
sc ; Sara ; Sara
sc ; Sgnw ; Sgnw
sc ; Syre ; Syre
sc ; Syrj ; Syrj
sc ; Syrn ; Syrn
sc ; Teng ; Teng
sc ; Visp ; Visp
sc ; Zxxx ; Zxxx
sc ; Kore ; Kore
sc ; Mani ; Mani
sc ; Phlp ; Phlp
sc ; Phlv ; Phlv
sc ; Zmth ; Zmth
sc ; Zsym ; Zsym
sc ; Nkgb ; Nkgb
sc ; Bass ; Bass
sc ; Dupl ; Dupl
sc ; Elba ; Elba
sc ; Gran ; Gran
sc ; Kpel ; Kpel
sc ; Loma ; Loma
sc ; Mend ; Mend
sc ; Narb ; Narb
sc ; Nbat ; Nbat
sc ; Palm ; Palm
sc ; Sind ; Sind
sc ; Wara ; Wara
sc ; Afak ; Afak
sc ; Jurc ; Jurc
sc ; Mroo ; Mroo
sc ; Nshu ; Nshu
sc ; Tang ; Tang
sc ; Wole ; Wole
sc ; Khoj ; Khoj
sc ; Tirh ; Tirh

403
tools/unicode/c/genpname/genpname.vcproj
View file

@ -1,403 +0,0 @@
<?xml version="1.0" encoding="Windows-1252"?>
<VisualStudioProject
ProjectType="Visual C++"
Version="9.00"
Name="genpname"
ProjectGUID="{DBC0AF0B-B9FF-4B23-905B-4D4CDC2A91CB}"
TargetFrameworkVersion="131072"
>
<Platforms>
<Platform
Name="Win32"
/>
<Platform
Name="x64"
/>
</Platforms>
<ToolFiles>
</ToolFiles>
<Configurations>
<Configuration
Name="Release|Win32"
OutputDirectory=".\x86\Release"
IntermediateDirectory=".\x86\Release"
ConfigurationType="1"
InheritedPropertySheets="$(VCInstallDir)VCProjectDefaults\UpgradeFromVC71.vsprops"
UseOfMFC="0"
ATLMinimizesCRunTimeLibraryUsage="false"
CharacterSet="2"
>
<Tool
Name="VCPreBuildEventTool"
/>
<Tool
Name="VCCustomBuildTool"
CommandLine="copy &quot;$(TargetPath)&quot; ..\..\..\bin&#x0D;&#x0A;"
Outputs="..\..\..\bin\$(TargetFileName)"
/>
<Tool
Name="VCXMLDataGeneratorTool"
/>
<Tool
Name="VCWebServiceProxyGeneratorTool"
/>
<Tool
Name="VCMIDLTool"
TypeLibraryName=".\x86\Release/genpname.tlb"
/>
<Tool
Name="VCCLCompilerTool"
AdditionalIncludeDirectories="..\..\common;..\toolutil"
PreprocessorDefinitions="WIN32;NDEBUG;_CRT_SECURE_NO_DEPRECATE"
StringPooling="true"
RuntimeLibrary="2"
EnableFunctionLevelLinking="true"
DisableLanguageExtensions="true"
TreatWChar_tAsBuiltInType="true"
PrecompiledHeaderFile=".\x86\Release/genpname.pch"
AssemblerListingLocation=".\x86\Release/"
ObjectFile=".\x86\Release/"
ProgramDataBaseFileName=".\x86\Release/"
WarningLevel="3"
SuppressStartupBanner="true"
CompileAs="0"
/>
<Tool
Name="VCManagedResourceCompilerTool"
/>
<Tool
Name="VCResourceCompilerTool"
PreprocessorDefinitions="NDEBUG"
Culture="1033"
/>
<Tool
Name="VCPreLinkEventTool"
/>
<Tool
Name="VCLinkerTool"
OutputFile=".\x86\Release/genpname.exe"
LinkIncremental="1"
SuppressStartupBanner="true"
ProgramDatabaseFile=".\x86\Release/genpname.pdb"
SubSystem="1"
RandomizedBaseAddress="1"
DataExecutionPrevention="0"
/>
<Tool
Name="VCALinkTool"
/>
<Tool
Name="VCManifestTool"
/>
<Tool
Name="VCXDCMakeTool"
/>
<Tool
Name="VCBscMakeTool"
/>
<Tool
Name="VCFxCopTool"
/>
<Tool
Name="VCAppVerifierTool"
/>
<Tool
Name="VCPostBuildEventTool"
/>
</Configuration>
<Configuration
Name="Debug|Win32"
OutputDirectory=".\x86\Debug"
IntermediateDirectory=".\x86\Debug"
ConfigurationType="1"
InheritedPropertySheets="$(VCInstallDir)VCProjectDefaults\UpgradeFromVC71.vsprops"
UseOfMFC="0"
ATLMinimizesCRunTimeLibraryUsage="false"
CharacterSet="2"
>
<Tool
Name="VCPreBuildEventTool"
/>
<Tool
Name="VCCustomBuildTool"
CommandLine="copy &quot;$(TargetPath)&quot; ..\..\..\bin&#x0D;&#x0A;"
Outputs="..\..\..\bin\$(TargetFileName)"
/>
<Tool
Name="VCXMLDataGeneratorTool"
/>
<Tool
Name="VCWebServiceProxyGeneratorTool"
/>
<Tool
Name="VCMIDLTool"
TypeLibraryName=".\x86\Debug/genpname.tlb"
/>
<Tool
Name="VCCLCompilerTool"
Optimization="0"
AdditionalIncludeDirectories="..\..\common;..\toolutil"
PreprocessorDefinitions="WIN32;_DEBUG;_CRT_SECURE_NO_DEPRECATE"
BasicRuntimeChecks="3"
RuntimeLibrary="3"
BufferSecurityCheck="true"
DisableLanguageExtensions="true"
TreatWChar_tAsBuiltInType="true"
PrecompiledHeaderFile=".\x86\Debug/genpname.pch"
AssemblerListingLocation=".\x86\Debug/"
ObjectFile=".\x86\Debug/"
ProgramDataBaseFileName=".\x86\Debug/"
BrowseInformation="1"
WarningLevel="3"
SuppressStartupBanner="true"
DebugInformationFormat="4"
CompileAs="0"
/>
<Tool
Name="VCManagedResourceCompilerTool"
/>
<Tool
Name="VCResourceCompilerTool"
PreprocessorDefinitions="_DEBUG"
Culture="1033"
/>
<Tool
Name="VCPreLinkEventTool"
/>
<Tool
Name="VCLinkerTool"
OutputFile=".\x86\Debug/genpname.exe"
LinkIncremental="2"
SuppressStartupBanner="true"
GenerateDebugInformation="true"
ProgramDatabaseFile=".\x86\Debug/genpname.pdb"
SubSystem="1"
RandomizedBaseAddress="1"
DataExecutionPrevention="0"
/>
<Tool
Name="VCALinkTool"
/>
<Tool
Name="VCManifestTool"
UseFAT32Workaround="true"
/>
<Tool
Name="VCXDCMakeTool"
/>
<Tool
Name="VCBscMakeTool"
/>
<Tool
Name="VCFxCopTool"
/>
<Tool
Name="VCAppVerifierTool"
/>
<Tool
Name="VCPostBuildEventTool"
/>
</Configuration>
<Configuration
Name="Release|x64"
OutputDirectory=".\x64\Release"
IntermediateDirectory=".\x64\Release"
ConfigurationType="1"
InheritedPropertySheets="$(VCInstallDir)VCProjectDefaults\UpgradeFromVC71.vsprops"
UseOfMFC="0"
ATLMinimizesCRunTimeLibraryUsage="false"
CharacterSet="2"
>
<Tool
Name="VCPreBuildEventTool"
/>
<Tool
Name="VCCustomBuildTool"
CommandLine="copy &quot;$(TargetPath)&quot; ..\..\..\bin64&#x0D;&#x0A;"
Outputs="..\..\..\bin64\$(TargetFileName)"
/>
<Tool
Name="VCXMLDataGeneratorTool"
/>
<Tool
Name="VCWebServiceProxyGeneratorTool"
/>
<Tool
Name="VCMIDLTool"
TargetEnvironment="3"
TypeLibraryName=".\x64\Release/genpname.tlb"
/>
<Tool
Name="VCCLCompilerTool"
AdditionalIncludeDirectories="..\..\common;..\toolutil"
PreprocessorDefinitions="WIN64;WIN32;NDEBUG;_CRT_SECURE_NO_DEPRECATE"
StringPooling="true"
RuntimeLibrary="2"
EnableFunctionLevelLinking="true"
DisableLanguageExtensions="true"
TreatWChar_tAsBuiltInType="true"
PrecompiledHeaderFile=".\x64\Release/genpname.pch"
AssemblerListingLocation=".\x64\Release/"
ObjectFile=".\x64\Release/"
ProgramDataBaseFileName=".\x64\Release/"
WarningLevel="3"
SuppressStartupBanner="true"
CompileAs="0"
/>
<Tool
Name="VCManagedResourceCompilerTool"
/>
<Tool
Name="VCResourceCompilerTool"
PreprocessorDefinitions="NDEBUG"
Culture="1033"
/>
<Tool
Name="VCPreLinkEventTool"
/>
<Tool
Name="VCLinkerTool"
OutputFile=".\x64\Release/genpname.exe"
LinkIncremental="1"
SuppressStartupBanner="true"
ProgramDatabaseFile=".\x64\Release/genpname.pdb"
SubSystem="1"
TargetMachine="17"
/>
<Tool
Name="VCALinkTool"
/>
<Tool
Name="VCManifestTool"
/>
<Tool
Name="VCXDCMakeTool"
/>
<Tool
Name="VCBscMakeTool"
/>
<Tool
Name="VCFxCopTool"
/>
<Tool
Name="VCAppVerifierTool"
/>
<Tool
Name="VCWebDeploymentTool"
/>
<Tool
Name="VCPostBuildEventTool"
/>
</Configuration>
<Configuration
Name="Debug|x64"
OutputDirectory=".\x64\Debug"
IntermediateDirectory=".\x64\Debug"
ConfigurationType="1"
InheritedPropertySheets="$(VCInstallDir)VCProjectDefaults\UpgradeFromVC71.vsprops"
UseOfMFC="0"
ATLMinimizesCRunTimeLibraryUsage="false"
CharacterSet="2"
>
<Tool
Name="VCPreBuildEventTool"
/>
<Tool
Name="VCCustomBuildTool"
CommandLine="copy &quot;$(TargetPath)&quot; ..\..\..\bin64&#x0D;&#x0A;"
Outputs="..\..\..\bin64\$(TargetFileName)"
/>
<Tool
Name="VCXMLDataGeneratorTool"
/>
<Tool
Name="VCWebServiceProxyGeneratorTool"
/>
<Tool
Name="VCMIDLTool"
TargetEnvironment="3"
TypeLibraryName=".\x64\Debug/genpname.tlb"
/>
<Tool
Name="VCCLCompilerTool"
Optimization="0"
AdditionalIncludeDirectories="..\..\common;..\toolutil"
PreprocessorDefinitions="WIN64;WIN32;_DEBUG;_CRT_SECURE_NO_DEPRECATE"
BasicRuntimeChecks="3"
RuntimeLibrary="3"
BufferSecurityCheck="true"
DisableLanguageExtensions="true"
TreatWChar_tAsBuiltInType="true"
PrecompiledHeaderFile=".\x64\Debug/genpname.pch"
AssemblerListingLocation=".\x64\Debug/"
ObjectFile=".\x64\Debug/"
ProgramDataBaseFileName=".\x64\Debug/"
BrowseInformation="1"
WarningLevel="3"
SuppressStartupBanner="true"
DebugInformationFormat="3"
CompileAs="0"
/>
<Tool
Name="VCManagedResourceCompilerTool"
/>
<Tool
Name="VCResourceCompilerTool"
PreprocessorDefinitions="_DEBUG"
Culture="1033"
/>
<Tool
Name="VCPreLinkEventTool"
/>
<Tool
Name="VCLinkerTool"
OutputFile=".\x64\Debug/genpname.exe"
LinkIncremental="2"
SuppressStartupBanner="true"
GenerateDebugInformation="true"
ProgramDatabaseFile=".\x64\Debug/genpname.pdb"
SubSystem="1"
TargetMachine="17"
/>
<Tool
Name="VCALinkTool"
/>
<Tool
Name="VCManifestTool"
UseFAT32Workaround="true"
/>
<Tool
Name="VCXDCMakeTool"
/>
<Tool
Name="VCBscMakeTool"
/>
<Tool
Name="VCFxCopTool"
/>
<Tool
Name="VCAppVerifierTool"
/>
<Tool
Name="VCWebDeploymentTool"
/>
<Tool
Name="VCPostBuildEventTool"
/>
</Configuration>
</Configurations>
<References>
</References>
<Files>
<File
RelativePath=".\data.h"
>
</File>
<File
RelativePath=".\genpname.cpp"
>
</File>
</Files>
<Globals>
</Globals>
</VisualStudioProject>

161
tools/unicode/c/genpname/gensvpa.pl
View file

@ -1,161 +0,0 @@
#!/usr/bin/perl
#*
#*******************************************************************************
#* Copyright (C) 2006, International Business Machines
#* Corporation and others. All Rights Reserved.
#*******************************************************************************
#*
#* file name: genspva.pl
#* encoding: US-ASCII
#* tab size: 8 (not used)
#* indentation:4
#*
#* Created by: Ram Viswanadha
#*
#* This file filters iso15924-utf8-<date>.txt
#*
use File::Find;
use File::Basename;
use IO::File;
use Cwd;
use File::Copy;
use Getopt::Long;
use File::Path;
use File::Copy;
#run the program
main();
#---------------------------------------------------------------------
# The main program
sub main(){
GetOptions(
"--destdir=s" => \$destdir,
"--iso15924=s" => \$iso,
"--prop=s" => \$prop,
"--code-start=s" => \$code,
);
usage() unless defined $destdir;
usage() unless defined $iso;
usage() unless defined $prop;
$outfile = "$destdir/SyntheticPropertyValueAliases.txt";
$propFH = IO::File->new($prop,"r")
or die "could not open the file $prop for reading: $! \n";
$isoFH = IO::File->new($iso,"r")
or die "could not open the file $iso for reading: $! \n";
$outFH = IO::File->new($outfile,"w")
or die "could not open the file $outfile for reading: $! \n";
my @propLines;
while (<$propFH>) {
next if(!($_ =~/sc ; /));
push(@propLines, $_);
}
printHeader($outFH);
if(defined $code){
print "Please add the following to UScriptCode enum in uscript.h.\n";
print "#ifndef U_HIDE_DRAFT_API\n";
}
while (<$isoFH>) {
next if($_=~/^#/);#skip if the line starts with a comment char
($script, $t, $name, $rest) = split(/;/,$_,4);
#sc ; Arab
$outstr = "sc ; $script";
$encoded = 0; #false
# seach the propLines to make sure that this scipt code is not
# encoded in Unicode
foreach $key (@propLines){
if($key =~ /$outstr/){
$encoded = 1;
}
}
next if($encoded == 1);
#ignore private use codes
next if($script =~ /Qa[ab][a-z]/);
#if($script eq "Qaaa"){
# $outstr = $outstr." ; Private_Use_Start\n";
#}elsif($script eq "Qabx"){
# $outstr = $outstr." ; Private_Use_End\n";
#}else{
# $outstr = $outstr." ; $script \n";
#}
$outstr = $outstr." ; $script \n";
print $outFH $outstr;
#print to console
if(defined $code){
if($name =~ /[(\s,\x80-\xFF]/){
$name = $script;
}
$name =~s/-/_/g;
$scriptcode = "USCRIPT_".uc($name);
print " $scriptcode = $code, /* $script */\n";
$code++;
}
}
if(defined $code){
print "#endif /* U_HIDE_DRAFT_API */\n";
}
for($i=0; $i<2; $i++){
}
close($isoFH);
close($propFH);
close($outFH);
}
#-----------------------------------------------------------------------
sub printHeader{
($outFH) = @_;
($DAY, $MONTH, $YEAR) = (localtime)[3,4,5];
$YEAR += 1900;
#We will print our copyright here + warnings
print $outFH <<END_HEADER_COMMENT;
########################################################################
# Copyright (c) 2006-$YEAR, International Business Machines
# Corporation and others. All Rights Reserved.
########################################################################
# file name: SyntheticPropertyValueAliases.txt
# encoding: US-ASCII
# tab size: 8 (not used)
# indentation: 4
# created by: gensvpa.pl
########################################################################
# This file follows the format of PropertyValueAliases.txt
# It contains synthetic property value aliases not present
# in the UCD. Unlike PropertyValueAliases.txt, it should
# NOT contain a version number.
########################################################################
# THIS FILE IS MACHINE-GENERATED, DON'T PLAY WITH IT IF YOU DON'T KNOW
# WHAT YOU ARE DOING, OTHERWISE VERY BAD THINGS WILL HAPPEN!
########################################################################
# set the same names as short and long names to fit the syntax without
# inventing names that we would have to support forever
# Script (sc)
END_HEADER_COMMENT
}
#-----------------------------------------------------------------------
sub usage {
print << "END";
Usage:
gensvpa.pl
Options:
--destdir=<directory>
--iso15924=<file name>
--prop=<PropertyValueAliases.txt>
--code-start=s
e.g.: gensvpa.pl --destdir=<icu>/source/tools/genpname --iso15924=iso15924-utf8-20041025.txt --prop=<icu>/source/data/unidata --code-start=60
END
exit(0);
}

1323
tools/unicode/c/genpname/preparse.pl
View file

File diff suppressed because it is too large Load diff

25
tools/unicode/c/genprops/pnames_data.h
View file

@ -2,18 +2,7 @@
* Copyright (C) 2002-2011, International Business Machines Corporation and
* others. All Rights Reserved.
*
* MACHINE GENERATED FILE. !!! Do not edit manually !!!
*
* Generated from
* uchar.h
* uscript.h
* Blocks.txt
* PropertyAliases.txt
* PropertyValueAliases.txt
*
* Date: Fri Dec 2 13:00:14 2011
* Unicode version: 6.1.0
* Script: preparse.pl
* machine-generated by: icu/tools/unicode/py/preparseucd.py
*/
/* Unicode version 6.1.0 */
@ -24,7 +13,6 @@ const uint8_t VERSION_3 = 0;
const int32_t STRING_COUNT = 1096;
/* to be sorted */
const AliasName STRING_TABLE[] = {
AliasName("", 0),
AliasName("A", 1),
@ -1124,9 +1112,6 @@ const AliasName STRING_TABLE[] = {
AliasName("xdigit", 1095),
};
/* to be filled in */
int32_t REMAP[1096];
const int32_t NAME_GROUP_COUNT = 1547;
int32_t NAME_GROUP[] = {
@ -2887,6 +2872,7 @@ const Property PROPERTY[] = {
Property((int32_t) UCHAR_WHITE_SPACE, 176, VALUES_binprop_COUNT, VALUES_binprop),
Property((int32_t) UCHAR_XID_CONTINUE, 179, VALUES_binprop_COUNT, VALUES_binprop),
Property((int32_t) UCHAR_XID_START, 181, VALUES_binprop_COUNT, VALUES_binprop),
Property((int32_t) UCHAR_NUMERIC_VALUE, 183, 0, NULL),
Property((int32_t) UCHAR_BIDI_CLASS, 185, VALUES_bc_COUNT, VALUES_bc),
Property((int32_t) UCHAR_BLOCK, 187, VALUES_blk_COUNT, VALUES_blk),
Property((int32_t) UCHAR_CANONICAL_COMBINING_CLASS, 189, VALUES_ccc_COUNT, VALUES_ccc),
@ -2908,6 +2894,8 @@ const Property PROPERTY[] = {
Property((int32_t) UCHAR_SENTENCE_BREAK, 221, VALUES_SB_COUNT, VALUES_SB),
Property((int32_t) UCHAR_TRAIL_CANONICAL_COMBINING_CLASS, 223, VALUES_tccc_COUNT, VALUES_tccc),
Property((int32_t) UCHAR_WORD_BREAK, 225, VALUES_WB_COUNT, VALUES_WB),
Property((int32_t) UCHAR_GENERAL_CATEGORY_MASK, 227, VALUES_gcm_COUNT, VALUES_gcm),
Property((int32_t) UCHAR_SCRIPT_EXTENSIONS, 229, 0, NULL),
Property((int32_t) UCHAR_AGE, 231, 0, NULL),
Property((int32_t) UCHAR_BIDI_MIRRORING_GLYPH, 233, 0, NULL),
Property((int32_t) UCHAR_CASE_FOLDING, 235, 0, NULL),
@ -2921,9 +2909,4 @@ const Property PROPERTY[] = {
Property((int32_t) UCHAR_TITLECASE_MAPPING, 252, 0, NULL),
Property((int32_t) UCHAR_UNICODE_1_NAME, 254, 0, NULL),
Property((int32_t) UCHAR_UPPERCASE_MAPPING, 256, 0, NULL),
Property((int32_t) UCHAR_NUMERIC_VALUE, 183, 0, NULL),
Property((int32_t) UCHAR_GENERAL_CATEGORY_MASK, 227, VALUES_gcm_COUNT, VALUES_gcm),
Property((int32_t) UCHAR_SCRIPT_EXTENSIONS, 229, 0, NULL),
};
/*eof*/

191
tools/unicode/py/preparseucd.py
View file

@ -28,6 +28,7 @@
import array
import bisect
import codecs
import datetime
import os
import os.path
import re
@ -177,7 +178,9 @@ def GetProperty(pname):
def GetShortPropertyName(pname):
if pname in _null_values: return pname # pname is already the short name.
prop = GetProperty(pname)
return prop[1][0] if prop else "" # "" for ignored properties.
if not prop: return "" # For ignored properties.
name = prop[1][0]
return name if name else prop[1][1] # Long name if no short name.
def GetShortPropertyValueName(prop, vname):
@ -396,10 +399,15 @@ def AddBinaryProperty(short_name, long_name):
_properties[NormPropName(long_name)] = prop
def AddPOSIXBinaryProperty(short_name, long_name):
AddBinaryProperty(short_name, long_name)
def AddPOSIXBinaryProperty(name):
# We only define a long name for ICU-specific (non-UCD) POSIX properties.
_null_values[name] = False
bin_prop = _properties["Math"]
prop = ("Binary", ["", name], bin_prop[2], bin_prop[3])
_properties[name] = prop
_properties[NormPropName(name)] = prop
# This is to match UProperty UCHAR_POSIX_ALNUM etc.
_properties["posix" + NormPropName(short_name)] = _properties[short_name]
_properties["posix" + NormPropName(name)] = prop
# Match a comment line like
@ -509,11 +517,11 @@ def ParsePropertyAliases(in_file):
AddBinaryProperty("segstart", "Segment_Starter")
# C/POSIX character classes that do not have Unicode property [value] aliases.
# See uchar.h.
AddPOSIXBinaryProperty("alnum", "alnum")
AddPOSIXBinaryProperty("blank", "blank")
AddPOSIXBinaryProperty("graph", "graph")
AddPOSIXBinaryProperty("print", "print")
AddPOSIXBinaryProperty("xdigit", "xdigit")
AddPOSIXBinaryProperty("alnum")
AddPOSIXBinaryProperty("blank")
AddPOSIXBinaryProperty("graph")
AddPOSIXBinaryProperty("print")
AddPOSIXBinaryProperty("xdigit")
def ParsePropertyValueAliases(in_file):
@ -1624,12 +1632,14 @@ def ParseUCharHeader(icu_src_root):
def WritePNamesDataHeader(out_path):
# Build a sorted list of (key0, enum) tuples
# Build a sorted list of (key0, enum, aliases) tuples
# to emulate the output order of the old genpname/preparse.pl.
# key0 is either a preparse.pl property type string (for property names)
# or a Unicode short property name (for property value names).
# enum is the ICU4C enum constant name.
# TODO: rename prop to not collide with usual properties[x]
# aliases is the tuple of the property's or value's names and aliases.
# (We use a tuple, not the original list,
# so that we can use it as a dict key.)
# TODO: once we are sure this works, simplify the order;
# for example, change all "_bp" etc. to just ""
# (outputs property names first in enum order),
@ -1646,9 +1656,12 @@ def WritePNamesDataHeader(out_path):
"Numeric": "_dp",
"String": "_sp"
}
pnames_data = [("binprop", "0"), ("binprop", "1")]
# Only properties that have ICU API.
pnames_data = [
("binprop", "0", tuple(_binary_values["N"])),
("binprop", "1", tuple(_binary_values["Y"]))
]
missing_enums = []
# Only properties that have ICU API.
for (pname, prop_enum) in _property_name_to_enum.iteritems():
prop = _properties[pname]
# Sometimes the uchar.h UProperty type differs
@ -1659,7 +1672,7 @@ def WritePNamesDataHeader(out_path):
type = "_op"
else:
type = prop_type_to_old_type[prop[0]]
pnames_data.append((type, prop_enum))
pnames_data.append((type, prop_enum, tuple(prop[1])))
if type != "_bp" and pname != "age":
short_name_to_enum = prop[2]
if pname.endswith("ccc"):
@ -1667,15 +1680,16 @@ def WritePNamesDataHeader(out_path):
# as "enum" values.
# In the UCD data, these numeric strings are the first value names,
# followed by the short & long value names.
# Omit the numeric strings from the aliases as well.
for name in short_name_to_enum:
pnames_data.append((pname, name))
pnames_data.append((pname, name, tuple(prop[3][name][1:])))
else:
if pname == "gc":
# See comment about _gc_vname_to_enum in ParseUCharHeader().
short_name_to_enum = _gc_vname_to_enum
for (name, enum) in short_name_to_enum.iteritems():
if enum:
pnames_data.append((pname, enum))
pnames_data.append((pname, enum, tuple(prop[3][name])))
else:
missing_enums.append((pname, name))
if missing_enums:
@ -1683,13 +1697,144 @@ def WritePNamesDataHeader(out_path):
"missing uchar.h enum constants for some property values: %s" %
missing_enums)
pnames_data.sort()
for item in pnames_data:
print item
short_script_name_to_enum = _properties["sc"][2]
# print short_script_name_to_enum
# print _property_name_to_enum
# print _properties["ea"][2]
# print _properties["gcm"][2]
# Write pnames_data.h.
year = datetime.date.today().strftime("%Y")
with open(out_path, "w") as out_file:
out_file.write("""/**
* Copyright (C) 2002-""" + year +
""", International Business Machines Corporation and
* others. All Rights Reserved.
*
* machine-generated by: icu/tools/unicode/py/preparseucd.py
*/
""")
out_file.write("/* Unicode version %s */\n" % _ucd_version)
v = _ucd_version.split(".")
while len(v) < 4: v.append("0")
for i in xrange(4):
out_file.write("const uint8_t VERSION_%d = %s;\n" % (i, v[i]))
out_file.write("\n")
# Write the string table with all of the names and aliases
# of all of the properties and their values.
# Unique strings in ASCII order.
# The first string must be the empty string.
strings_set = set([""])
for (key0, enum, aliases) in pnames_data:
for s in aliases: strings_set.add(s)
strings = sorted(strings_set)
out_file.write("const int32_t STRING_COUNT = %d;\n\n" % len(strings))
# While printing, create a mapping from string table entry to index.
string_to_id = {}
out_file.write("const AliasName STRING_TABLE[] = {\n")
for s in strings:
i = len(string_to_id)
out_file.write(' AliasName("%s", %d),\n' % (s, i))
string_to_id[s] = i
out_file.write("};\n\n")
# Emit the name group table.
# [A table of name groups. A name group is the list of names and aliases
# for a property or property value.
# The name group table looks like this:
#
# 114, -115, 116, -117, 0, -118, 65, -64, ...
# [0] [2] [4] [6]
#
# The entry at [0] consists of 2 strings, 114 and 115.
# The entry at [2] consists of 116 and 117, etc.
# The last entry is negative.
# Build the name group list with nameGroup indices.
name_groups = []
# Count the total number of values, not just the groups.
name_groups_total_length = 0
# Check for duplicate name groups, and reuse the first of a kind.
group_to_int = {}
for (key0, enum, aliases) in pnames_data:
index = group_to_int.get(aliases)
if index == None:
index = name_groups_total_length
group_to_int[aliases] = index
name_groups.append([string_to_id[s] for s in aliases])
name_groups_total_length += len(aliases)
out_file.write("const int32_t NAME_GROUP_COUNT = %d;\n\n" %
name_groups_total_length)
out_file.write("int32_t NAME_GROUP[] = {\n")
# Emit one group per line, with annotations.
max_names = 0 # Maximum number of names & aliases per item.
start = 0
for group in name_groups:
line = " "
aliases = [] # For comments.
for i in group[:-1]:
line += "%d, " % i
aliases.append('"%s"' % strings[i])
# Negative entry terminates the group.
i = group[-1]
line += "%d, " % -i
aliases.append('"%s"' % strings[i])
out_file.write(
line + " " * (24 - len(line)) +
"/* %3d: " % start + ", ".join(aliases) + " */\n")
length = len(group)
if length > max_names: max_names = length
start += length
out_file.write("};\n\n")
out_file.write("#define MAX_NAMES_PER_GROUP %d\n\n" % max_names)
# Emit the enumerated property values.
i = 0
while i < len(pnames_data):
pname = pnames_data[i][0]
if pname.startswith("_"):
i += 1
continue
if pname == "binprop":
count = 2
elif pname == "gc":
count = len(_gc_vname_to_enum)
else:
count = len(_properties[pname][2])
out_file.write("const int32_t VALUES_%s_COUNT = %d;\n\n" %
(pname, count))
out_file.write("const Alias VALUES_%s[] = {\n" % pname)
limit = i + count
while i < limit:
(pname, enum, aliases) = pnames_data[i]
out_file.write(" Alias((int32_t) %s, %d),\n" %
(enum, group_to_int[aliases]))
i += 1
out_file.write("};\n\n")
# Emit the top-level properties (binary, enumerated, etc.).
out_file.write("const int32_t PROPERTY_COUNT = %d;\n\n" %
len(_property_name_to_enum))
out_file.write("const Property PROPERTY[] = {\n")
for (pname, enum, aliases) in pnames_data:
if not pname.startswith("_"): continue
pname = aliases[0]
if not pname: pname = aliases[1] # Long name if no short name.
prop = _properties[pname]
group_index = group_to_int[aliases]
if prop[2] and pname != "age": # Property with named values.
if prop[0] == "Binary": pname = "binprop"
out_file.write(
" Property((int32_t) %s, %d, VALUES_%s_COUNT, VALUES_%s),\n" %
(enum, group_index, pname, pname))
else:
out_file.write(" Property((int32_t) %s, %d, 0, NULL),\n" %
(enum, group_index))
out_file.write("};\n")
# main() ------------------------------------------------------------------- ***