r38779 MediaWiki - Code Review archive

Repository:	MediaWiki
Revision:	< r38778‎ \| r38779 \| r38780 >
Date:	13:57, 7 August 2008
Author:	river
Status:	old
Tags:
Comment:	makefile: cleanups affunctions: change glib to ICU (untested) eval.cpp: remote obsolete libxml include
Modified paths:	/trunk/extensions/AbuseFilter/parser_native/af_parser (deleted) (history) /trunk/extensions/AbuseFilter/parser_native/affunctions.cpp (modified) (history) /trunk/extensions/AbuseFilter/parser_native/affunctions.h (modified) (history) /trunk/extensions/AbuseFilter/parser_native/check (deleted) (history) /trunk/extensions/AbuseFilter/parser_native/eval.cpp (modified) (history) /trunk/extensions/AbuseFilter/parser_native/makefile (modified) (history) /trunk/extensions/AbuseFilter/parser_native/syntax_check (deleted) (history)

Diff [purge]

Index: trunk/extensions/AbuseFilter/parser_native/syntax_check
Cannot display: file marked as a binary type.
svn:mime-type = application/octet-stream
Index: trunk/extensions/AbuseFilter/parser_native/af_parser
Cannot display: file marked as a binary type.
svn:mime-type = application/octet-stream
Index: trunk/extensions/AbuseFilter/parser_native/check
Cannot display: file marked as a binary type.
svn:mime-type = application/octet-stream
Index: trunk/extensions/AbuseFilter/parser_native/affunctions.cpp
—	—	@@ -5,8 +5,10 @@
6	6	#include <ios>
7	7	#include <iostream>
8	8	#include <ctype.h>
9		~~-#include <glibmm/ustring.h>~~
10	9
	10	+#include <unicode/utf8.h>
	11	+#include <unicode/ustring.h>
	12	+
11	13	#define EQUIVSET_LOC "equivset.txt"
12	14
13	15	map<string,AFPFunction> af_functions;
—	—	@@ -179,7 +181,7 @@
180	182	}
181	183	}
182	184
183		~~- double ratio = (float)(specialcount) / (float)(((Glib::ustring)orig).size());~~
	185	+ double ratio = (float)(specialcount) / (float)(utf8_strlen(orig));
184	186
185	187	return AFPData(ratio);
186	188	}
—	—	@@ -234,7 +236,7 @@
235	237	throw AFPException( "Not enough arguments to lcase" );
236	238	}
237	239
238		~~- return AFPData( (long int)((Glib::ustring)args[0].toString()).size() );~~
	240	+ return AFPData( (long int)utf8_strlen(args[0].toString()) );
239	241	}
240	242
241	243	AFPData af_lcase( vector<AFPData> args ) {
—	—	@@ -242,10 +244,7 @@
243	245	throw AFPException( "Not enough arguments to lcase" );
244	246	}
245	247
246		~~- Glib::ustring s = args[0].toString();~~
247		~~- string s2 = s.lowercase();~~
248		-
249		~~- return AFPData(s);~~
	248	+ return AFPData(utf8_tolower(args[0].toString()));
250	249	}
251	250
252	251	string confusable_character_normalise( string orig ) {
—	—	@@ -354,6 +353,66 @@
355	354	return c;
356	355	}
357	356
	357	+std::size_t
	358	+utf8_strlen(std::string const &s)
	359	+{
	360	+std::size_t ret = 0;
	361	+ for (std::string::const_iterator it = s.begin(), end = s.end();
	362	+ it < end; ++it)
	363	+ {
	364	+ int skip = 1;
	365	+
	366	+ skip = U8_LENGTH(*it);
	367	+#if 0
	368	+ if (*it >= 0xc0) {
	369	+ if (*it < 0xe0)
	370	+ skip = 1;
	371	+ else if (*it < 0xf0)
	372	+ skip = 2;
	373	+ else
	374	+ skip = 3;
	375	+ } else
	376	+ skip = 1;
	377	+#endif
	378	+
	379	+ if (it + skip >= end)
	380	+ return ret; /* end of string */
	381	+
	382	+ it += skip;
	383	+ }
	384	+
	385	+ return ret;
	386	+}
	387	+
	388	+/*
	389	+ * This could almost certainly be done in a nicer way.
	390	+ */
	391	+std::string utf8_tolower(std::string const &s)
	392	+{
	393	+ std::vector<UChar> ustring;
	394	+ UErrorCode error = U_ZERO_ERROR;
	395	+
	396	+ for (int i = 0; i < s.size(); ) {
	397	+ UChar32 c;
	398	+ U8_NEXT(s.data(), i, s.size(), c);
	399	+ ustring.push_back(c);
	400	+ }
	401	+
	402	+ std::vector<UChar> dest;
	403	+ u_strToLower(&dest[0], dest.size(), &ustring[0], ustring.size(),
	404	+ NULL, &error);
	405	+
	406	+ if (U_FAILURE(error))
	407	+ return s;
	408	+
	409	+ std::vector<unsigned char> u8string;
	410	+ int i, j;
	411	+ for (i = 0, j = 0; i < dest.size(); j++) {
	412	+ U8_APPEND_UNSAFE(&u8string[0], i, dest[j]);
	413	+ }
	414	+ return std::string(u8string.begin(), u8string.begin() + i);
	415	+}
	416	+
358	417	// Ported from MediaWiki core function in PHP.
359	418	string codepointToUtf8( int codepoint ) {
360	419	string ret;
Index: trunk/extensions/AbuseFilter/parser_native/affunctions.h
—	—	@@ -1,3 +1,6 @@
	2	+#ifndef AFFUNCTIONS_H
	3	+#define AFFUNCTIONS_H
	4	+
2	5	#include "aftypes.h"
3	6	#include <map>
4	7	#include <vector>
—	—	@@ -18,3 +21,7 @@
19	22	AFPData callFunction( string name, AFPData arg );
20	23	string rmdoubles( string orig );
21	24	string rmspecials( string orig );
	25	+std::size_t utf8_strlen(std::string const &s);
	26	+std::string utf8_tolower(std::string const &s);
	27	+
	28	+#endif /* !AFFUNCTIONS_H */
Index: trunk/extensions/AbuseFilter/parser_native/eval.cpp
—	—	@@ -1,6 +1,5 @@
2	2	#include "afeval.h"
3	3	#include "affunctions.h"
4		~~-#include <libxml++/libxml++.h>~~
5	4	#include <iostream>
6	5	#include <string>
7	6	#include <sstream>
Index: trunk/extensions/AbuseFilter/parser_native/makefile
—	—	@@ -1,24 +1,41 @@
2		~~-all: check af_parser syntax_check af_expr~~
	2	+CXX = g++
	3	+CXXFLAGS = -O3
	4	+BOOST_INCLUDES = /opt/boost/include/boost-1_35
	5	+BOOST_LIBS = /opt/boost/lib
	6	+BOOST_TAG = -gcc34-mt
	7	+ICU_INCLUDES = /opt/icu/include
	8	+ICU_LIBS = /opt/icu/lib
3	9
4		~~-af_expr: afeval.o affunctions.o afparser.o aftypes.o afutils.o eval.o~~
5		~~- g++ -O3 -o af_expr afeval.o affunctions.o afparser.o aftypes.o afutils.o eval.o -lboost_regex -lxml++-2.6 -lxml2 -lglibmm-2.4 -lgobject-2.0 -lsigc-2.0 -lglib-2.0~~
	10	+CPPFLAGS = -I$(BOOST_INCLUDES) -I$(ICU_INCLUDES)
	11	+LDFLAGS = -L$(BOOST_LIBS) -L$(ICU_LIBS) -R$(ICU_LIBS) -R$(BOOST_LIBS)
6	12
7		~~-af_parser: afeval.o affunctions.o afparser.o aftypes.o afutils.o main.o~~
8		~~- g++ -O3 -o af_parser afeval.o affunctions.o afparser.o aftypes.o afutils.o main.o -lboost_regex -lglibmm-2.4~~
	13	+LIBS = -lboost_regex$(BOOST_TAG) -licuuc -licui18n -licudata -licui18n
9	14
10		~~-check: afeval.o affunctions.o afparser.o aftypes.o afutils.o check.o~~
11		~~- g++ -O3 -o check -lboost_regex afeval.o affunctions.o afparser.o aftypes.o afutils.o check.o -lglibmm-2.4~~
12		-
13		~~-syntax_check: afeval.o affunctions.o afparser.o aftypes.o afutils.o syntax_check.o~~
14		~~- g++ -O3 -o syntax_check afeval.o affunctions.o afparser.o aftypes.o afutils.o syntax_check.o -lboost_regex -lglibmm-2.4~~
	15	+af_expr_objs = afeval.o affunctions.o afparser.o aftypes.o afutils.o eval.o
	16	+af_parser_objs = afeval.o affunctions.o afparser.o aftypes.o afutils.o main.o
	17	+check_objs = afeval.o affunctions.o afparser.o aftypes.o afutils.o check.o
	18	+syntax_check_objs = afeval.o affunctions.o afparser.o aftypes.o afutils.o syntax_check.o
15	19
	20	+progs = check af_parser syntax_check af_expr
	21	+
	22	+all: $(progs)
	23	+
	24	+af_expr: $(af_expr_objs)
	25	+ $(CXX) $(CXXFLAGS) -o $@ $(af_expr_objs) $(LDFLAGS) $(LIBS)
	26	+af_parser: $(af_parser_objs)
	27	+ $(CXX) $(CXXFLAGS) -o $@ $(af_parser_objs) $(LDFLAGS) $(LIBS)
	28	+check: $(check_objs)
	29	+ $(CXX) $(CXXFLAGS) -o $@ $(check_objs) $(LDFLAGS) $(LIBS)
	30	+syntax_check: $(syntax_check_objs)
	31	+ $(CXX) $(CXXFLAGS) -o $@ $(syntax_check_objs) $(LDFLAGS) $(LIBS)
	32	+
16	33	.cpp.o:
17		- g++ -O3 -c $< -I/usr/include/libxml++-2.6 -I/usr/lib/libxml++-2.6/include -I/usr/include/libxml2 -I/usr/include/glibmm-2.4 -I/usr/lib/glibmm-2.4/include -I/usr/include/sigc++-2.0 -I/usr/lib/sigc++-2.0/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include
	34	+ $(CXX) $(CPPFLAGS) $(CXXFLAGS) -c $<
18	35
19	36	clean:
20		~~- rm -f *.o check af_parser syntax_check~~
	37	+ rm -f *.o $(progs)
21	38
22	39	clean-final:
23		~~- rm -f check af_parser syntax_check~~
	40	+ rm -f $(progs)
24	41
25	42	.SUFFIXES: .cpp .o

Status & tagging log

15:29, 12 September 2011 Meno25 (talk | contribs) changed the status of r38779 [removed: ok added: old]