patch set!
| Newsgroups | gmane.comp.gnu.gift.general |
|---|---|
| Message-ID | <[email protected]> |
here are the patches i've been talking about. changelogs for each, etc etc.. 29 patches. breifly, make gift build in a builddir (AKA, instead of doing "./configure; make" you do "mkdir build; cd build; ../configure ; make") fix two buffer overflows (one mine, one davids) move from uint32_t to uint_fast32_t, which is uint32_t on 32bit, and uint64_t on 64bit. cleanups.. thats about it. here goes! Julia Longtin <[email protected]> _______________________________________________ help-GIFT mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-gift
00-ChangeLog
(text/plain, 353 B)
--- ../../anoncvs/gift/ChangeLog 2006-10-22 09:59:12.000000000 -0500 +++ ChangeLog 2006-11-18 11:03:04.000000000 -0600 @@ -1,3 +1,9 @@ +2006-11-18 <[email protected]> + + * bootstrap-cvs.sh + error out when a step fails. + remove --force from autoconf and aclocal invocations. + 2006-10-10 <[email protected]> * FeatureExtraction/gabor.c
00-bootstrap_cvs_sh-remove_force_and_error_out.patch
(text/plain, 1 KB)
--- ../../anoncvs/gift/bootstrap-cvs.sh 2006-02-28 07:48:21.000000000 -0600
+++ bootstrap-cvs.sh 2006-11-18 10:39:19.000000000 -0600
@@ -33,12 +33,12 @@
echo "running libtoolize"
libtoolize --force --ltdl || { echo "Libtoolize failed. THIS MIGHT CAUSE PROBLEMS LATER, continuing " ; }
echo "running aclocal"
-aclocal --force -I . || { echo "bootstrap failed, because aclocal failed. exiting. " , exit 100; }
+aclocal -I . || { echo "bootstrap failed, because aclocal failed. exiting. " ; exit 100; }
echo "...done"
echo "running automake -a"
-automake -a --force || { echo "bootstrap failed, because automake failed. exiting." , exit 101; }
+automake -a || { echo "bootstrap failed, because automake failed. exiting." ; exit 101; }
echo "...done"
echo "running autoconf"
-autoconf || { echo "bootstrap failed, because autoconf failed. exiting." , exit 102; }
+autoconf || { echo "bootstrap failed, because autoconf failed. exiting." ; exit 102; }
echo "...done"
echo "Now you should be able to run ./configure"
01-ChangeLog
(text/plain, 398 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 11:26:47.000000000 -0600
+++ ChangeLog 2006-11-18 11:40:45.000000000 -0600
@@ -3,6 +3,8 @@
* bootstrap-cvs.sh
error out when a step fails.
remove --force from autoconf and aclocal invocations.
+ * configure.in
+ call perl test program with a path, so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
01-configure_in-make_build_in_subdir.patch
(text/plain, 604 B)
--- ../../dev3/gift/configure.in 2006-07-11 15:05:23.000000000 -0500
+++ configure.in 2006-11-18 11:10:17.000000000 -0600
@@ -66,7 +66,7 @@
AC_PATH_PROG(PERL,perl)
-if perl gift-check-perl-config.pl "XML::DOM" "XML::XQL" "XML::XQL::DOM" "Text::Iconv" "HTTP::Request" "XML::Parser";
+if perl ${srcdir}/gift-check-perl-config.pl "XML::DOM" "XML::XQL" "XML::XQL::DOM" "Text::Iconv" "HTTP::Request" "XML::Parser";
then echo "Your perl setup seems to be OK for running the GIFT" ; else
AC_MSG_ERROR([There are some modules missing from your perl configuration. Please see the output above.])
fi
02-ChangeLog
(text/plain, 524 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 11:43:33.000000000 -0600
+++ ChangeLog 2006-11-18 12:02:21.000000000 -0600
@@ -5,6 +5,9 @@
remove --force from autoconf and aclocal invocations.
* configure.in
call perl test program with a path, so that we can build gift in a directory other than ${srcdir}
+ * dtd/Makefile.am
+ make mv invocations use ${top_builddir}, so that we can build gift in a directory other than ${srcdir}
+ replace "../" with appropriate autoconf variables.
2006-10-10 <[email protected]>
02-dtd_Makefile_am-make_build_in_subdir.patch
(text/plain, 864 B)
--- ../../dev2/gift/dtd/Makefile.am 2001-12-11 18:26:21.000000000 +0100
+++ dtd/Makefile.am 2006-10-31 22:22:31.000000000 +0100
@@ -1,14 +1,14 @@
-noinst_DATA= ../libMRML/cc/mrml_const.cc
+noinst_DATA= ${top_builddir}/libMRML/cc/mrml_const.cc
#
# making a new mrml_const.cc if mrml.dtd has been changed
#
-../libMRML/cc/mrml_const.cc: mrml.dtd
- $(PERL) ../scripts/perl/gift-dtd-to-keywords.pl ../dtd/mrml.dtd mrml mrml_const charmer.mrml
- mv mrml_const.h ../libMRML/include
- mv mrml_const.cc ../libMRML/cc
- mv mrml_const.java ../
+${top_builddir}/libMRML/cc/mrml_const.cc: mrml.dtd
+ $(PERL) ${top_builddir}/scripts/perl/gift-dtd-to-keywords.pl ${srcdir}/mrml.dtd mrml mrml_const charmer.mrml
+ mv mrml_const.h ${top_builddir}/libMRML/include
+ mv mrml_const.cc ${top_builddir}/libMRML/cc
+ mv mrml_const.java ${top_builddir}/
EXTRA_DIST=mrml.dtd
03-ChangeLog
(text/plain, 542 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 12:03:45.000000000 -0600
+++ ChangeLog 2006-11-18 12:18:55.000000000 -0600
@@ -8,6 +8,9 @@
* dtd/Makefile.am
make mv invocations use ${top_builddir}, so that we can build gift in a directory other than ${srcdir}
replace "../" with appropriate autoconf variables.
+ * libMRML/cc/CAttributeList.cc
+ include wordsize.h, so we can tell if we're a 64 or 32bit binary during build.
+ when dumping pointer as int, check the size of the pointer before casting.
2006-10-10 <[email protected]>
03-libMRML_cc_CAttributeList_cc-check_size_of_pointer_before_dumping.patch
(text/plain, 1 KB)
--- ../../dev3/gift/libMRML/cc/CAttributeList.cc 2006-02-28 07:48:22.000000000 -0600
+++ libMRML/cc/CAttributeList.cc 2006-11-18 12:14:09.000000000 -0600
@@ -27,6 +27,8 @@
#include "libMRML/include/CMutex.h"
#include <iostream>
#include <algorithm>
+/* for __WORDSIZE */
+#include <bits/wordsize.h>
extern CMutex* gMutex;
CAttributeList::CAttributeList(const char * const * const inAttributeList){
@@ -94,7 +96,11 @@
if(strcmp(i->second.first,
i->second.second)){
- cout << "DELETING FAILED" << flush << int(i->second.second) << flush << i->second.first << i->second.second << endl;
+#if __WORDSIZE==64 /* this should tell us what size a float* is, so that we can dump it properly.. */
+ cout << "DELETING FAILED" << flush << int64_t(i->second.second) << flush << i->second.first << i->second.second << endl;
+#else
+ cout << "DELETING FAILED" << flush << int32_t(i->second.second) << flush << i->second.first << i->second.second << endl;
+#endif
assert(0);
}
delete[] i->second.first;
04-ChangeLog
(text/plain, 505 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 12:27:56.000000000 -0600
+++ ChangeLog 2006-11-18 12:30:56.000000000 -0600
@@ -11,6 +11,8 @@
* libMRML/cc/CAttributeList.cc
include wordsize.h, so we can tell if we're a 64 or 32bit binary during build.
when dumping pointer as int, check the size of the pointer before casting.
+ * FeatureExtraction/Makefile.am
+ change $(top_srcdir) to $(top_builddir), so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
04-FeatureExtraction_Makefile_am-make_build_in_subdir.patch
(text/plain, 638 B)
--- ../../dev/gift/FeatureExtraction/Makefile.am 2006-01-10 02:53:02.000000000 -0600 +++ FeatureExtraction/Makefile.am 2006-11-06 04:54:19.000000000 -0600 @@ -28,9 +28,9 @@ gift_write_feature_descs_SOURCES= $(david_sources) write_feature_descs.c -gift_extract_features_LDADD = -L$(top_srcdir)/libSquirePPM -lSquirePPM -lm +gift_extract_features_LDADD = -L$(top_builddir)/libSquirePPM -lSquirePPM -lm -gift_write_feature_descs_LDADD= -L$(top_srcdir)/libSquirePPM -lSquirePPM -lm +gift_write_feature_descs_LDADD= -L$(top_builddir)/libSquirePPM -lSquirePPM -lm bin_PROGRAMS = gift-extract-features gift-write-feature-descs
05-ChangeLog
(text/plain, 764 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 12:32:29.000000000 -0600
+++ ChangeLog 2006-11-18 12:53:08.000000000 -0600
@@ -13,6 +13,10 @@
when dumping pointer as int, check the size of the pointer before casting.
* FeatureExtraction/Makefile.am
change $(top_srcdir) to $(top_builddir), so that we can build gift in a directory other than ${srcdir}
+ * GIFTServer/Makefile.am
+ remove $(srcdir), not being used.
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ NOTE: in reality, before the "libraries" can be moved into being seperately installable packages, these paths need changed to <directory/target.h>
2006-10-10 <[email protected]>
05-GIFTServer_Makefile_am-make_build_in_subdir.patch
(text/plain, 572 B)
--- ../../dev3/gift/GIFTServer/Makefile.am 2003-04-22 03:03:46.000000000 -0500 +++ GIFTServer/Makefile.am 2006-11-18 12:49:25.000000000 -0600 @@ -16,7 +16,7 @@ # include directories for all programs in this dir -INCLUDES = -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(srcdir) -I. +INCLUDES = -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(top_srcdir) -I. #-------------------------------------------------
06-ChangeLog
(text/plain, 723 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:04:43.000000000 -0600
+++ ChangeLog 2006-11-18 13:09:45.000000000 -0600
@@ -17,6 +17,8 @@
remove $(srcdir), not being used.
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
NOTE: in reality, before the "libraries" can be moved into being seperately installable packages, these paths need changed to <directory/target.h>
+ * libGIFTAcDistanceMatrix/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
06-libGIFTAcDistanceMatrix_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 935 B)
--- ../../dev/gift/libGIFTAcDistanceMatrix/cc/Makefile.am 2006-02-28 07:48:22.000000000 -0600 +++ libGIFTAcDistanceMatrix/cc/Makefile.am 2006-11-06 04:12:22.000000000 -0600 @@ -25,7 +25,7 @@ libGIFTAcDistanceMatrix_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcInvertedFile/cc/.libs/libGIFTAcInvertedFile.a -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcDistanceMatrix/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(top_srcdir)/libGIFTQuBayesian/include -I$(HOME)/browsing/libGIFTQuBayesian/include +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcDistanceMatrix/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(top_srcdir)/libGIFTQuBayesian/include -I$(top_srcdir) -I$(HOME)/browsing/libGIFTQuBayesian/include libGIFTAcDistanceMatrix_la_SOURCES= \
07-ChangeLog
(text/plain, 725 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:11:06.000000000 -0600
+++ ChangeLog 2006-11-18 13:14:38.000000000 -0600
@@ -19,6 +19,8 @@
NOTE: in reality, before the "libraries" can be moved into being seperately installable packages, these paths need changed to <directory/target.h>
* libGIFTAcDistanceMatrix/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTAcHierarchy/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
07-libGIFTAcHierarchy_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 618 B)
--- ../../dev/gift/libGIFTAcHierarchy/cc/Makefile.am 2004-12-03 15:15:19.000000000 +0100 +++ libGIFTAcHierarchy/cc/Makefile.am 2006-11-01 04:24:01.000000000 +0100 @@ -30,7 +30,7 @@ libGIFTAcHierarchy_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcHierarchy/include -I$(top_srcdir)/expat/xmlparse -I$(top_srcdir)/expat/xmltok +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcHierarchy/include -I$(top_srcdir)/expat/xmlparse -I$(top_srcdir)/expat/xmltok -I$(top_srcdir)/ libGIFTAcHierarchy_la_SOURCES= \
08-ChangeLog
(text/plain, 734 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:16:38.000000000 -0600
+++ ChangeLog 2006-11-18 13:23:28.000000000 -0600
@@ -21,6 +21,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTAcHierarchy/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTAcInvertedFile/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
08-libGIFTAcInvertedFile_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 665 B)
--- ../../dev3/gift/libGIFTAcInvertedFile/cc/Makefile.am 2006-02-28 07:48:22.000000000 -0600 +++ libGIFTAcInvertedFile/cc/Makefile.am 2006-11-18 13:17:51.000000000 -0600 @@ -27,7 +27,7 @@ libGIFTAcInvertedFile_la_LDFLAGS= -module "@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcURL2FTS/cc/.libs/libGIFTAcURL2FTS.a -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcURL2FTS/include +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(top_srcdir) #CAcIFFileSystem.cc: CAcIFFileSystem.pre-cc
09-ChangeLog
(text/plain, 729 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:29:01.000000000 -0600
+++ ChangeLog 2006-11-18 13:28:54.000000000 -0600
@@ -23,6 +23,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTAcInvertedFile/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTAcPerl/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
09-libGIFTAcPerl_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 540 B)
--- /disk1/gift/src/merge/dev2/gift/libGIFTAcPerl/cc/Makefile.am 2002-02-26 14:20:18.000000000 +0100 +++ libGIFTAcPerl/cc/Makefile.am 2006-11-01 20:51:29.000000000 +0100 @@ -25,7 +25,7 @@ libGIFTAcPerl_la_LDFLAGS=@PRE_INSTALL_LIBDIRS@ --rpath../../libMRML/cc/.libs/ -lMRML -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcPerl/include +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcPerl/include -I$(top_srcdir) # libGIFTAcPerl_a_SOURCES= libGIFTAcPerl_la_SOURCES= \
10-ChangeLog
(text/plain, 724 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:30:38.000000000 -0600
+++ ChangeLog 2006-11-18 13:32:10.000000000 -0600
@@ -25,6 +25,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTAcPerl/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTAcURL2FTS/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
10-libGIFTAcURL2FTS_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 531 B)
--- ../../dev3/gift/libGIFTAcURL2FTS/cc/Makefile.am 2002-02-26 07:20:18.000000000 -0600 +++ libGIFTAcURL2FTS/cc/Makefile.am 2006-11-18 13:31:34.000000000 -0600 @@ -25,7 +25,7 @@ libGIFTAcURL2FTS_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcURL2FTS/include \ - -I$(top_srcdir)/expat/xmlparse -I$(top_srcdir)/expat/xmltok + -I$(top_srcdir)/expat/xmlparse -I$(top_srcdir)/expat/xmltok -I$(top_srcdir)/ libGIFTAcURL2FTS_la_SOURCES= \ CAcURL2FTS.cc\
11-ChangeLog
(text/plain, 729 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:45:38.000000000 -0600
+++ ChangeLog 2006-11-18 13:45:32.000000000 -0600
@@ -27,6 +27,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTAcURL2FTS/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTQuHierarchy/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
11-libGIFTQuHierarchy_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 624 B)
--- ../../dev3/gift/libGIFTQuHierarchy/cc/Makefile.am 2004-12-03 08:15:20.000000000 -0600 +++ libGIFTQuHierarchy/cc/Makefile.am 2006-11-18 13:44:52.000000000 -0600 @@ -28,6 +28,6 @@ libGIFTQuHierarchy_la_LDFLAGS=@PRE_INSTALL_LIBDIRS@ -lMRML -lGIFTAcHierarchy -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcHierarchy/include -I$(top_srcdir)/libGIFTQuHierarchy/include +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcHierarchy/include -I$(top_srcdir)/libGIFTQuHierarchy/include -I$(top_srcdir)/ libGIFTQuHierarchy_la_SOURCES= CQHierarchy.cc plug_in_fo.cc
12-ChangeLog
(text/plain, 734 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 13:56:43.000000000 -0600
+++ ChangeLog 2006-11-18 13:56:26.000000000 -0600
@@ -29,6 +29,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTQuHierarchy/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTQuInvertedFile/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
12-libGIFTQuInvertedFile_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 668 B)
--- /disk1/gift/src/merge/dev2/gift/libGIFTQuInvertedFile/cc/Makefile.am 2006-02-28 14:48:22.000000000 +0100 +++ libGIFTQuInvertedFile/cc/Makefile.am 2006-11-01 20:41:28.000000000 +0100 @@ -29,7 +29,7 @@ # # The trick here is: do not compile c code with the -D bool=char define. #-rdynamic -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTQuInvertedFile/include -I$(top_srcdir)/libGIFTAcInvertedFile/include +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTQuInvertedFile/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/ libGIFTQuInvertedFile_la_SOURCES= \ CWeightingFunction.cc \
13-ChangeLog
(text/plain, 729 B)
--- ../../dev3/gift/ChangeLog 2006-11-18 14:00:15.000000000 -0600
+++ ChangeLog 2006-11-18 14:00:07.000000000 -0600
@@ -31,6 +31,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTQuInvertedFile/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libGIFTQuPerl/cc/Makefile.am
+ HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
2006-10-10 <[email protected]>
13-libGIFTQuPerl_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 572 B)
--- ../../dev3/gift/libGIFTQuPerl/cc/Makefile.am 2002-02-26 07:20:19.000000000 -0600 +++ libGIFTQuPerl/cc/Makefile.am 2006-11-18 13:59:16.000000000 -0600 @@ -27,7 +27,7 @@ # # The trick here is: do not compile c code with the -D bool=char define. #-rdynamic -INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcPerl/include -I$(top_srcdir)/libGIFTQuPerl/include +INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcPerl/include -I$(top_srcdir)/libGIFTQuPerl/include -I$(top_srcdir)/ libGIFTQuPerl_la_SOURCES= \
14-ChangeLog
(text/plain, 603 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 10:37:15.000000000 -0600
+++ ChangeLog 2006-11-19 10:48:54.000000000 -0600
@@ -33,6 +33,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libGIFTQuPerl/cc/Makefile.am
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
+ * libMRML/cc/Makefile.am
+ use $(srcdir)/ to find libMRML-config.pre-cc.
2006-10-10 <[email protected]>
14-libMRML_cc_Makefile_am-make_build_in_subdir.patch
(text/plain, 508 B)
--- ../../dev3/gift/libMRML/cc/Makefile.am 2006-01-10 02:53:03.000000000 -0600
+++ libMRML/cc/Makefile.am 2006-11-19 10:46:41.000000000 -0600
@@ -114,7 +114,7 @@
s§%LDFLAGS%§$(LDFLAGS)§g;\
s§%LIBS%§$(LIBS)§ g;\
s§%GIFT_MYSQL_DIR%§@GIFT_MYSQL_DIR@§g;\
- s§%libdir%§$(libdir)§g;" < libMRML-config.pre-cc > libMRML-config.cc
+ s§%libdir%§$(libdir)§g;" < $(srcdir)/libMRML-config.pre-cc > libMRML-config.cc
bin_PROGRAMS= libMRML-config
15-ChangeLog
(text/plain, 522 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 11:06:05.000000000 -0600
+++ ChangeLog 2006-11-19 11:07:27.000000000 -0600
@@ -35,6 +35,8 @@
HACK: add $(top_srcdir) to the path, to fix #includes of the form "directory/include/target.h", so that we can build gift in a directory other than ${srcdir}
* libMRML/cc/Makefile.am
use $(srcdir)/ to find libMRML-config.pre-cc.
+ * libGIFTAcDistanceMatrix/cc/Makefile.am
+ link against .la file, instead of .a file. more portable this way.
2006-10-10 <[email protected]>
15-libGIFTAcDistanceMatrix_cc_Makefile_am-use_portable_library.patch
(text/plain, 796 B)
--- ../../dev/gift/libGIFTAcDistanceMatrix/cc/Makefile.am 2006-11-06 04:16:40.000000000 -0600 +++ libGIFTAcDistanceMatrix/cc/Makefile.am 2006-11-06 04:16:18.000000000 -0600 @@ -23,7 +23,7 @@ lib_LTLIBRARIES = libGIFTAcDistanceMatrix.la -libGIFTAcDistanceMatrix_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcInvertedFile/cc/.libs/libGIFTAcInvertedFile.a +libGIFTAcDistanceMatrix_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcInvertedFile/cc/.libs/libGIFTAcInvertedFile.la INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcDistanceMatrix/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(top_srcdir)/libGIFTQuBayesian/include -I$(top_srcdir) -I$(HOME)/browsing/libGIFTQuBayesian/include
16-ChangeLog
(text/plain, 445 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 11:08:34.000000000 -0600 +++ ChangeLog 2006-11-19 11:09:09.000000000 -0600 @@ -37,6 +37,8 @@ use $(srcdir)/ to find libMRML-config.pre-cc. * libGIFTAcDistanceMatrix/cc/Makefile.am link against .la file, instead of .a file. more portable this way. + * libGIFTAcInvertedFile/cc/Makefile.am + link against .la file, instead of .a file. more portable this way. 2006-10-10 <[email protected]>
16-libGIFTAcInvertedFile_cc_Makefile_am-use_portable_library.patch
(text/plain, 648 B)
--- ../../dev2/gift/libGIFTAcInvertedFile/cc/Makefile.am 2006-11-19 11:19:54.000000000 -0600 +++ libGIFTAcInvertedFile/cc/Makefile.am 2006-11-19 11:18:43.000000000 -0600 @@ -25,7 +25,7 @@ lib_LTLIBRARIES = libGIFTAcInvertedFile.la -libGIFTAcInvertedFile_la_LDFLAGS= -module "@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcURL2FTS/cc/.libs/libGIFTAcURL2FTS.a +libGIFTAcInvertedFile_la_LDFLAGS= -module "@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcURL2FTS/cc/.libs/libGIFTAcURL2FTS.la INCLUDES= -DHAS_BOOL -I$(top_srcdir)/libMRML/include -I$(top_srcdir)/libGIFTAcInvertedFile/include -I$(top_srcdir)/libGIFTAcURL2FTS/include -I$(top_srcdir)
17-ChangeLog
(text/plain, 464 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 11:22:04.000000000 -0600 +++ ChangeLog 2006-11-19 11:29:50.000000000 -0600 @@ -39,6 +39,8 @@ link against .la file, instead of .a file. more portable this way. * libGIFTAcInvertedFile/cc/Makefile.am link against .la file, instead of .a file. more portable this way. + * libGIFTQuInvertedFile/cc/Makefile.am + link against .la file, instead of .a file. more portable this way. 2006-10-10 <[email protected]>
17-libGIFTQuInvertedFile_cc_Makefile_am-use_portable_library.patch
(text/plain, 514 B)
--- /disk1/gift/src/merge/dev2/gift/libGIFTQuInvertedFile/cc/Makefile.am 2006-11-01 21:56:07.000000000 +0100 +++ libGIFTQuInvertedFile/cc/Makefile.am 2006-11-01 20:42:30.000000000 +0100 @@ -23,7 +23,7 @@ lib_LTLIBRARIES = libGIFTQuInvertedFile.la -libGIFTQuInvertedFile_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcInvertedFile/cc/.libs/libGIFTAcInvertedFile.a +libGIFTQuInvertedFile_la_LDFLAGS="@PRE_INSTALL_LIBDIRS@" -lMRML ../../libGIFTAcInvertedFile/cc/.libs/libGIFTAcInvertedFile.la #
18-ChangeLog
(text/plain, 595 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 11:38:48.000000000 -0600 +++ ChangeLog 2006-11-19 11:36:05.000000000 -0600 @@ -41,6 +41,10 @@ link against .la file, instead of .a file. more portable this way. * libGIFTQuInvertedFile/cc/Makefile.am link against .la file, instead of .a file. more portable this way. + * FeatureExtraction/gabor.c + fix an off-by-one error in the addressing of value_plane_double_reversed. my fault. + * FeatureExtraction/rgb2hsv_utils.c + fix an off-by-one error in the addressing of value_plane_double_reversed. my fault. 2006-10-10 <[email protected]>
18-FeatureExtraction_gabor_c-fix_pointer_error.patch
(text/plain, 1 KB)
--- ../../anoncvs/gift/FeatureExtraction/gabor.c 2006-10-22 09:59:12.000000000 -0500
+++ FeatureExtraction/gabor.c 2006-11-15 07:33:38.000000000 -0600
@@ -95,7 +95,7 @@
target_kernal=kernelsxy[filter_scale*num_gabors_per_scale+orientation];
for (x = 0; x < width; x++) {
for (y = 0; y < height; y++) {
- target_image=&image[(width*height)-(y*width+x+kernal_size[filter_scale]/2)];
+ target_image=&image[(width*height-1)-(y*width+x+kernal_size[filter_scale]/2)];
if ((x>=kernal_size[filter_scale]/2) && ((x+kernal_size[filter_scale]/2)<width))
{
for (k = 0; k < kernal_size[filter_scale]; k++)
@@ -143,7 +143,7 @@
target_kernal=&target_kernal[kernal_size[filter_scale]];
for (x = 0; x < width; x++) {
for (y = 0; y < height; y++) {
- target_image=&image[(width*height)-(y*width+x+kernal_size[filter_scale]/2)];
+ target_image=&image[(width*height-1)-(y*width+x+kernal_size[filter_scale]/2)];
if ((x>=kernal_size[filter_scale]/2) && ((x+kernal_size[filter_scale]/2)<width))
{
for (k = 0; k < kernal_size[filter_scale]; k++)
18-FeatureExtraction_rgb2hsv_utils_c-fix_pointer_error.patch
(text/plain, 591 B)
--- ../../anoncvs/gift/FeatureExtraction/rgb2hsv_utils.c 2006-07-20 12:09:15.000000000 -0500 +++ FeatureExtraction/rgb2hsv_utils.c 2006-11-08 16:09:08.000000000 -0600 @@ -202,7 +202,7 @@ (*im_hsv)->pixel[3*i + HUE] = (byte)rint((H/360.0)*max_col); (*im_hsv)->pixel[3*i + SATURATION] = (byte)rint(S*max_col); (*im_hsv)->pixel[3*i + VALUE] = (byte)rint(V*max_col); - (*im_hsv)->value_plane_double_reversed[num_pixels-i]=(V*max_col); + (*im_hsv)->value_plane_double_reversed[num_pixels-i-1]=(V*max_col); } add_comment((*im_hsv), "# Image converted from RGB to HSV format.\n");
19-ChangeLog
(text/plain, 552 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 11:48:50.000000000 -0600 +++ ChangeLog 2006-11-19 11:57:06.000000000 -0600 @@ -45,6 +45,8 @@ fix an off-by-one error in the addressing of value_plane_double_reversed. my fault. * FeatureExtraction/rgb2hsv_utils.c fix an off-by-one error in the addressing of value_plane_double_reversed. my fault. + * FeatureExtraction/gabor.c + move from using an array to initialize the size of another array to using a define'd value. one less GCC warning, and slightly faster. 2006-10-10 <[email protected]>
19-FeatureExtraction_gabor_c-remove_array_warning.patch
(text/plain, 682 B)
--- ../../dev3/gift/FeatureExtraction/gabor.c 2006-11-15 07:50:31.000000000 -0600
+++ FeatureExtraction/gabor.c 2006-11-15 07:47:15.000000000 -0600
@@ -17,6 +17,8 @@
static int kernal_size[num_gabor_scales] = {gabor_kernel_size_0, gabor_kernel_size_1, gabor_kernel_size_2};
+#define MAX_KERNAL_SIZE gabor_kernel_size_2
+
void save_norm_double_pgm(double *double_im, int w, int h, char *fname) {
PPM *im;
@@ -89,7 +91,7 @@
double * target_kernal;
double * target_conv;
double * target_image;
- double temparray[kernal_size[2]];
+ double temparray[MAX_KERNAL_SIZE];
/* first convolution */
target_kernal=kernelsxy[filter_scale*num_gabors_per_scale+orientation];
20-ChangeLog
(text/plain, 547 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:02:38.000000000 -0600 +++ ChangeLog 2006-11-19 12:02:31.000000000 -0600 @@ -47,6 +47,8 @@ fix an off-by-one error in the addressing of value_plane_double_reversed. my fault. * FeatureExtraction/gabor.c move from using an array to initialize the size of another array to using a define'd value. one less GCC warning, and slightly faster. + * libSquirePPM/ppm_comment.c + fix an off-by-one error in the allocation of ppm->comments. NOT my fault. finally! :) 2006-10-10 <[email protected]>
20-libSquirePPM_ppm_comment_c-fix_pointer_error.patch
(text/plain, 464 B)
--- ../../anoncvs/gift/libSquirePPM/ppm_comment.c 2001-02-01 12:22:41.000000000 +0100
+++ libSquirePPM/ppm_comment.c 2006-11-09 05:27:17.000000000 +0100
@@ -40,7 +40,7 @@
free(the_ppm->comments);
the_ppm->comment_length = i + strlen(new_comment);
- the_ppm->comments = malloc(the_ppm->comment_length*sizeof(char));
+ the_ppm->comments = malloc((the_ppm->comment_length+1)*sizeof(char));
i = 0;
if (original_comments != NULL) {
c = original_comments;
21-ChangeLog
(text/plain, 578 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:03:35.000000000 -0600 +++ ChangeLog 2006-11-19 12:07:39.000000000 -0600 @@ -49,6 +49,8 @@ move from using an array to initialize the size of another array to using a define'd value. one less GCC warning, and slightly faster. * libSquirePPM/ppm_comment.c fix an off-by-one error in the allocation of ppm->comments. NOT my fault. finally! :) + * FeatureExtraction/gabor.c + move from using uint32_t to using uint_fast32_t. this becomes 64bit on AMD64, solving the (wierd) crashing issue. 2006-10-10 <[email protected]>
21-FeatureExtraction_gabor_c-change_datatypes.patch
(text/plain, 870 B)
--- ../../dev3/gift/FeatureExtraction/gabor.c 2006-11-15 07:52:32.000000000 -0600
+++ FeatureExtraction/gabor.c 2006-11-15 07:55:50.000000000 -0600
@@ -54,8 +54,8 @@
void create_filter_kernels(double ** kernelsxy) {
- uint32_t i, j;
- int32_t x, x_c; /* ints to force type promotion later */
+ uint_fast32_t i, j;
+ int_fast32_t x, x_c; /* ints to force type promotion later */
double u0, u, v, sigma, theta;
/* set the values of the kernels */
@@ -86,8 +86,8 @@
/* conv and conv2 are just temporary space, for juggling image data between filters */
void gabor_filter(double *image, int width, int height, int filter_scale, int orientation, double **kernelsxy, double *conv, double *conv2, double *output) {
- uint32_t x, y;
- uint32_t k;
+ uint_fast32_t x, y;
+ uint_fast32_t k;
double * target_kernal;
double * target_conv;
double * target_image;
22-ChangeLog
(text/plain, 513 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:08:25.000000000 -0600 +++ ChangeLog 2006-11-19 12:11:25.000000000 -0600 @@ -51,6 +51,8 @@ fix an off-by-one error in the allocation of ppm->comments. NOT my fault. finally! :) * FeatureExtraction/gabor.c move from using uint32_t to using uint_fast32_t. this becomes 64bit on AMD64, solving the (wierd) crashing issue. + * FeatureExtraction/gabor.c + formatting and comment additions. no functional changes in this patch. 2006-10-10 <[email protected]>
22-FeatureExtraction_gabor_c-formatting_and_comments.patch
(text/plain, 4.3 KB)
--- ../../dev2/gift/FeatureExtraction/gabor.c 2006-11-15 07:55:50.000000000 -0600
+++ FeatureExtraction/gabor.c 2006-11-15 08:03:00.000000000 -0600
@@ -1,6 +1,7 @@
#include <stdio.h>
#include <stdlib.h>
#include <malloc.h>
+/* for M_PI */
#include <math.h>
/* for memset(), others */
#include <string.h>
@@ -9,7 +10,7 @@
#include "gabor.h"
-/* for uint32_t */
+/* for uint_fast32_t, int_fast32_t */
#include <stdint.h>
/* note that all this is taken from the paper JaH98 in ~/tex/bib/squizz.bib */
@@ -69,13 +70,13 @@
x_c = kernal_size[i]/2; /* since sizes are odd, this gives the
centre value */
- kernelsxy[(i*num_gabors_per_scale+j)] = (double *)malloc(kernal_size[i]*sizeof(double)*4);
+ kernelsxy[i*num_gabors_per_scale+j] = (double *)malloc(kernal_size[i]*sizeof(double)*4);
for (x = 0; x < kernal_size[i]; x++) {
/* note that x is "y" for kernels 12 and 22 */
- kernelsxy[(i*num_gabors_per_scale+j)][x] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*cos(2*M_PI*u*(x - x_c)));
- kernelsxy[(i*num_gabors_per_scale+j)][x+kernal_size[i]] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*cos(2*M_PI*v*(x - x_c)));
- kernelsxy[(i*num_gabors_per_scale+j)][x+kernal_size[i]*2] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*sin(2*M_PI*u*(x - x_c)));
- kernelsxy[(i*num_gabors_per_scale+j)][x+kernal_size[i]*3] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*sin(2*M_PI*v*(x - x_c)));
+ kernelsxy[i*num_gabors_per_scale+j][x] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*cos(2*M_PI*u*(x - x_c)));
+ kernelsxy[i*num_gabors_per_scale+j][x+kernal_size[i]] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*cos(2*M_PI*v*(x - x_c)));
+ kernelsxy[i*num_gabors_per_scale+j][x+kernal_size[i]*2] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*sin(2*M_PI*u*(x - x_c)));
+ kernelsxy[i*num_gabors_per_scale+j][x+kernal_size[i]*3] = (1/(sqrt(2*M_PI)*sigma)*exp(-sq((x - x_c))/(2*sq(sigma)))*sin(2*M_PI*v*(x - x_c)));
}
}
u0 = u0/2;
@@ -103,13 +104,13 @@
for (k = 0; k < kernal_size[filter_scale]; k++)
temparray[k]= target_kernal[k]*target_image[k];
for (k = 0; k < kernal_size[filter_scale]; k++)
- conv[((width*height)-1)-(x*height+y)] += temparray[k];
+ conv[(width*height-1)-(x*height+y)] += temparray[k];
}
else
{
for (k=0; k < kernal_size[filter_scale]; k++) {
if ((x+kernal_size[filter_scale]/2 >= k) && (x+kernal_size[filter_scale]/2 < width+k)) {
- conv[((width*height)-1)-(x*height+y)] +=
+ conv[(width*height-1)-(x*height+y)] +=
target_kernal[k]*target_image[k];
}
}
@@ -121,7 +122,7 @@
target_kernal=&target_kernal[kernal_size[filter_scale]];
for (x = 0; x < width; x++) {
for (y = 0; y < height; y++) {
- target_conv=&conv[((width*height)-1)-(x*height+y+(kernal_size[filter_scale]/2))];
+ target_conv=&conv[(width*height-1)-(x*height+y+(kernal_size[filter_scale]/2))];
if (((y>=kernal_size[filter_scale]/2)) && ((y+kernal_size[filter_scale]/2)<height))
{
/* first do the matrix multiply, then do the summation, so the matrix multiply can be vectored. */
@@ -151,13 +152,13 @@
for (k = 0; k < kernal_size[filter_scale]; k++)
temparray[k]= target_kernal[k]*target_image[k];
for (k = 0; k < kernal_size[filter_scale]; k++)
- conv2[((width*height)-1)-(x*height+y)] += temparray[k];
+ conv2[(width*height-1)-(x*height+y)] += temparray[k];
}
else
{
for (k=0; k < kernal_size[filter_scale]; k++) {
if ((x+kernal_size[filter_scale]/2 >= k) && (x+kernal_size[filter_scale]/2 < width+k)) {
- conv2[((width*height)-1)-(x*height+y)] +=
+ conv2[(width*height-1)-(x*height+y)] +=
target_kernal[k]*target_image[k];
}
}
@@ -169,7 +170,7 @@
target_kernal=&target_kernal[kernal_size[filter_scale]];
for (x = 0; x < width; x++) {
for (y = 0; y < height; y++) {
- target_conv=&conv2[((width*height)-1)-(x*height+y+(kernal_size[filter_scale]/2))];
+ target_conv=&conv2[(width*height-1)-(x*height+y+(kernal_size[filter_scale]/2))];
if (((y>=kernal_size[filter_scale]/2)) && ((y+kernal_size[filter_scale]/2)<height))
{
/* first do the matrix multiply, then do the summation, so the matrix multiply can be vectored. */
23-ChangeLog
(text/plain, 483 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:13:09.000000000 -0600 +++ ChangeLog 2006-11-19 12:14:27.000000000 -0600 @@ -53,6 +53,8 @@ move from using uint32_t to using uint_fast32_t. this becomes 64bit on AMD64, solving the (wierd) crashing issue. * FeatureExtraction/gabor.c formatting and comment additions. no functional changes in this patch. + * FeatureExtraction/extract_block_features.c + add missing headers, update a comment. 2006-10-10 <[email protected]>
23-FeatureExtraction_extract_block_features_c-header_cleanups.patch
(text/plain, 425 B)
--- ../../anoncvs/gift/FeatureExtraction/extract_block_features.c 2006-10-22 16:59:12.000000000 +0200 +++ FeatureExtraction/extract_block_features.c 2006-11-16 01:14:22.000000000 +0100 @@ -30,7 +30,13 @@ /* for M_PI */ #include <math.h> -/* for uint32_t */ +/* for memset() */ +#include <string.h> + +/* for exit() */ +#include <stdlib.h> + +/* for uint32_t, uint_fast32_t */ #include <stdint.h> #ifdef HAVE_CONFIG_H
24-ChangeLog
(text/plain, 501 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:16:44.000000000 -0600 +++ ChangeLog 2006-11-19 12:25:06.000000000 -0600 @@ -55,6 +55,8 @@ formatting and comment additions. no functional changes in this patch. * FeatureExtraction/extract_block_features.c add missing headers, update a comment. + * FeatureExtraction/extract_block_features.c + move from the (erronous) use of j in output, to using gabor_filter(num_filter_scales-1). happened to work on AMD64. 2006-10-10 <[email protected]>
24-FeatureExtraction_extract_block_features_c-switch_from_j_to_gabor_filter.patch
(text/plain, 1 KB)
--- ../../dev3/gift/FeatureExtraction/extract_block_features.c 2006-11-19 12:16:44.000000000 -0600
+++ FeatureExtraction/extract_block_features.c 2006-11-19 12:22:29.000000000 -0600
@@ -680,7 +680,13 @@
for (scale = 0; scale < num_gabor_scales; scale++) {
for (orientation = 0; orientation < num_gabors_per_scale; orientation++) {
for (j = 0; j < num_gabor_ranges; j++) {
- fprintf(out_file, "%d %d GABOR_HST SCALE, ORIENTATION, ENERGY UPPER BOUND = %d, %d, %f\n", feature_index, GABOR_HST, scale, orientation, j);
+#if num_gabor_ranges==16 /* again, we should really be checking the type... */
+ fprintf(out_file, "%d %d GABOR_HST SCALE, ORIENTATION, ENERGY UPPER BOUND = %d, %d, %f\n", feature_index, GABOR_HST, scale, orientation, gabor_ranges((num_gabor_ranges-1)));
+#else
+#if num_gabor_ranges==10
+ fprintf(out_file, "%d %d GABOR_HST SCALE, ORIENTATION, ENERGY UPPER BOUND = %d, %d, %d.000000\n", feature_index, GABOR_HST, scale, orientation, gabor_ranges((num_gabor_ranges-1)));
+#endif
+#endif
feature_index++;
}
}
25-ChangeLog
(text/plain, 564 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:25:44.000000000 -0600 +++ ChangeLog 2006-11-19 12:32:49.000000000 -0600 @@ -57,6 +57,10 @@ add missing headers, update a comment. * FeatureExtraction/extract_block_features.c move from the (erronous) use of j in output, to using gabor_filter(num_filter_scales-1). happened to work on AMD64. + * FeatureExtraction/extract_block_features.c + move from using uint32_t to using uint_fast32_t. + * FeatureExtraction/gabor.c + move from accepting int to accepting uint_fast32_t. 2006-10-10 <[email protected]>
25-FeatureExtraction_extract_block_features_c-convert_uint32_t_to_uint_fast32_t.patch
(text/plain, 3.4 KB)
--- ../../dev3/gift/FeatureExtraction/extract_block_features.c 2006-11-16 15:44:35.000000000 +0100
+++ FeatureExtraction/extract_block_features.c 2006-11-16 15:37:56.000000000 +0100
@@ -108,32 +108,32 @@
#endif
/* this version just does the work. no variable width/height information is taken into account. 256x256. */
-void init_feature_variables(uint32_t colmap_size, uint32_t ** col_counts, uint32_t *** block_gabor_class, uint32_t *** gabor_histogram) {
+void init_feature_variables(uint_fast32_t colmap_size, uint_fast32_t ** col_counts, uint_fast32_t *** block_gabor_class, uint_fast32_t *** gabor_histogram) {
- uint32_t i, j, k=0;
+ uint_fast32_t i, j, k=0;
/* colour features */
for (i = 0; i < num_colour_scales; i++) {
for (j = 0; j < num_blocks_at_scale(i); j++) {
- col_counts[k++] = (uint32_t *)malloc(colmap_size*sizeof(uint32_t));
+ col_counts[k++] = (uint_fast32_t *)malloc(colmap_size*sizeof(uint_fast32_t));
}
}
/* Gabor features */
for (i = 0; i < num_gabor_scales; i++) {
- block_gabor_class[i] = (uint32_t **)malloc(num_gabors_per_scale*sizeof(uint32_t *));
- gabor_histogram[i] = (uint32_t **)malloc(num_gabors_per_scale*sizeof(uint32_t *));
+ block_gabor_class[i] = (uint_fast32_t **)malloc(num_gabors_per_scale*sizeof(uint_fast32_t *));
+ gabor_histogram[i] = (uint_fast32_t **)malloc(num_gabors_per_scale*sizeof(uint_fast32_t *));
for (j = 0; j < num_gabors_per_scale; j++) {
- block_gabor_class[i][j] = (uint32_t *)malloc(square(image_size/gabor_block_size)*sizeof(uint32_t));
- gabor_histogram[i][j] = (uint32_t *)calloc(num_gabor_ranges, sizeof(uint32_t));
+ block_gabor_class[i][j] = (uint_fast32_t *)malloc(square(image_size/gabor_block_size)*sizeof(uint_fast32_t));
+ gabor_histogram[i][j] = (uint_fast32_t *)calloc(num_gabor_ranges, sizeof(uint_fast32_t));
}
}
}
-void extract_gabor_features(PPM *im_hsv, uint32_t *** block_gabor_class, uint32_t *** gabor_histogram) {
+void extract_gabor_features(PPM *im_hsv, uint_fast32_t *** block_gabor_class, uint_fast32_t *** gabor_histogram) {
- int i, j, x, y, k;
- int scale, orientation;
- int energy_class;
+ uint_fast32_t i, j, x, y, k;
+ uint_fast32_t scale, orientation;
+ uint_fast32_t energy_class;
PPM *value_plane;
double *value_image_dbl, *filtered_image;
double *conv;
@@ -199,7 +199,7 @@
}
}
-void extract_mode_features(PPM *im, uint32_t *colmap, uint32_t colmap_size, uint32_t ** col_counts, byte * block_mode, uint32_t * col_histogram) {
+void extract_mode_features(PPM *im, uint_fast32_t *colmap, uint_fast32_t colmap_size, uint_fast32_t ** col_counts, byte * block_mode, uint_fast32_t * col_histogram) {
int i, j, k, last_k, k1, m, n, r, x, y;
byte colour;
@@ -360,7 +360,7 @@
col_counts[k + 2][n] + col_counts[k + 3][n];
}
-enum ppm_error write_mode_features(char *out_fname, uint32_t colmap_size, uint32_t * col_histogram, uint32_t *** block_gabor_class, uint32_t *** gabor_histogram, byte * block_mode) {
+enum ppm_error write_mode_features(char *out_fname, uint_fast32_t colmap_size, uint_fast32_t * col_histogram, uint_fast32_t *** block_gabor_class, uint_fast32_t *** gabor_histogram, byte * block_mode) {
FILE *out_file = NULL;
int feature_index;
@@ -535,7 +535,7 @@
return(PPM_OK);
}
-void fts2blocks(char *fts_fname, byte * block_mode, uint32_t * col_histogram) {
+void fts2blocks(char *fts_fname, byte * block_mode, uint_fast32_t * col_histogram) {
FILE *fts_file, *outfile;
int num_features, feature_index;
25-FeatureExtraction_gabor_c-convert_int_to_uint_fast32_t.patch
(text/plain, 1.1 KB)
--- ../../dev2/gift/FeatureExtraction/gabor.c 2006-11-15 21:05:25.000000000 +0100
+++ FeatureExtraction/gabor.c 2006-11-16 16:11:53.000000000 +0100
@@ -16,7 +16,7 @@
/* note that all this is taken from the paper JaH98 in ~/tex/bib/squizz.bib */
/* (with my corrections!) */
-static int kernal_size[num_gabor_scales] = {gabor_kernel_size_0, gabor_kernel_size_1, gabor_kernel_size_2};
+static int_fast8_t kernal_size[num_gabor_scales] = {gabor_kernel_size_0, gabor_kernel_size_1, gabor_kernel_size_2};
#define MAX_KERNAL_SIZE gabor_kernel_size_2
@@ -85,7 +85,7 @@
/* conv, conv2, and output need to be cleared before feeding them to this function. */
/* conv and conv2 are just temporary space, for juggling image data between filters */
-void gabor_filter(double *image, int width, int height, int filter_scale, int orientation, double **kernelsxy, double *conv, double *conv2, double *output) {
+void gabor_filter(double *image, uint_fast32_t width, uint_fast32_t height, uint_fast32_t filter_scale, uint_fast32_t orientation, double **kernelsxy, double *conv, double *conv2, double *output) {
uint_fast32_t x, y;
uint_fast32_t k;
26-ChangeLog
(text/plain, 384 B)
--- ../../dev3/gift/ChangeLog 2006-11-19 12:34:49.000000000 -0600 +++ ChangeLog 2006-11-19 12:35:20.000000000 -0600 @@ -61,6 +61,8 @@ move from using uint32_t to using uint_fast32_t. * FeatureExtraction/gabor.c move from accepting int to accepting uint_fast32_t. + * FeatureExtraction/extract_block_features.c + remove unused variables. 2006-10-10 <[email protected]>
26-FeatureExtraction_extract_block_features_c-remove_unused_variables.patch
(text/plain, 901 B)
--- ../../dev2/gift/FeatureExtraction/extract_block_features.c 2006-11-16 15:37:56.000000000 +0100
+++ FeatureExtraction/extract_block_features.c 2006-11-16 15:39:50.000000000 +0100
@@ -134,8 +134,7 @@
uint_fast32_t i, j, x, y, k;
uint_fast32_t scale, orientation;
uint_fast32_t energy_class;
- PPM *value_plane;
- double *value_image_dbl, *filtered_image;
+ double *filtered_image;
double *conv;
double *conv2;
double gabor_mean;
@@ -201,9 +200,9 @@
void extract_mode_features(PPM *im, uint_fast32_t *colmap, uint_fast32_t colmap_size, uint_fast32_t ** col_counts, byte * block_mode, uint_fast32_t * col_histogram) {
- int i, j, k, last_k, k1, m, n, r, x, y;
+ int i, j, k, last_k, m, n, r, x, y;
byte colour;
- int scale, block_size, num_blocks, old_num_blocks;
+ int scale, block_size, num_blocks;
int max_count, mode_index;
int b1, b2, b3, b4;
#ifdef GENERATE_BLOCK_IMAGES