Add pkgconfig file

[rossburton]

Now that snowball builds a shared library a pkgconfig file to find the library would be much appreciated.

[ojwb]

Snowball doesn't yet build a shared library (there are patches floating around for it, but nothing in a suitable state for merging yet).

[eli-schwartz]

Snowball also still doesn't install anything... people who are distributing a shared library have to manually cp the library into /usr/lib64 or wherever.

[ojwb]

Snowball also still doesn't install anything

Yes, but that's literally a separate issue (#189) - please don't mix discussions.

[eli-schwartz]

It really is not a separate thing, though. It's a prerequisite. A pkg-config file has to include the path that the library and header file is installed to, it cannot do that unless it is in fact installed.

[sidrat]

Hi all first time contributor

Here is a pkgconfig file i created for libstemmer

cat > libstemmer.pc << "EOF"

prefix=/usr
exec_prefix=${prefix}
libdir=${exec_prefix}/lib
sharedlibdir=${libdir}
includedir=${prefix}/include

Name: libstemmer
Description: Snowball compiler and stemming algorithms 
Version: 2.2.0

Requires:
Libs: -L${libdir} -L${sharedlibdir} -lstemmer
Cflags: -I${includedir} 

EOF

To conferm libstemmer.so is picked up from pkgconfig I use libxmlb-0.3.21 build from Beyond Linux From Scratch (BLFS) book

wget -c https://github.com/hughsie/libxmlb/releases/download/0.3.21/libxmlb-0.3.21.tar.xz
tar -xf libxmlb-0.3.21.tar.xz
cd libxmlb-0.3.21

mkdir build &&
cd    build &&

meson setup --prefix=/usr --buildtype=release -D gtkdoc=false -D stemmer=true .. &&
ninja
ninja test

ninja install

I use ldd to check that libstemmer is included in libxmlb.so
ldd /usr/lib/libxmlb.so

result:
	linux-vdso.so.1 (0x00007c29bcf88000)
	libgio-2.0.so.0 => /usr/lib/libgio-2.0.so.0 (0x00007c29bcc00000)
	libgobject-2.0.so.0 => /usr/lib/libgobject-2.0.so.0 (0x00007c29bcedf000)
	libglib-2.0.so.0 => /usr/lib/libglib-2.0.so.0 (0x00007c29bcaa1000)
	liblzma.so.5 => /usr/lib/liblzma.so.5 (0x00007c29bceaf000)
	libzstd.so.1 => /usr/lib/libzstd.so.1 (0x00007c29bc9a8000)  

libstemmer.so.0d => /usr/lib/libstemmer.so.0d (0x00007c29bc8df000)

	libc.so.6 => /usr/lib/libc.so.6 (0x00007c29bc6f9000)
	libgmodule-2.0.so.0 => /usr/lib/libgmodule-2.0.so.0 (0x00007c29bcea6000)
	libz.so.1 => /usr/lib/libz.so.1 (0x00007c29bce87000)
	libmount.so.1 => /usr/lib/libmount.so.1 (0x00007c29bce18000)
	libffi.so.8 => /usr/lib/libffi.so.8 (0x00007c29bce07000)
	libpcre2-8.so.0 => /usr/lib/libpcre2-8.so.0 (0x00007c29bc658000)
	/lib64/ld-linux-x86-64.so.2 (0x00007c29bcf8a000)
	libblkid.so.1 => /usr/lib/libblkid.so.1 (0x00007c29bc5fa000)

---

I do not know how to test the functions of libstemmer but i hope this points people in the right direction or gives someone an idea.

This is a install script I used after build libstemmer.

cat > install.sh << "EOF"
cp libstemmer.pc /usr/lib/pkgconfig
cp stemwords  /usr/bin
cp snowball  /usr/bin
cp include/*  /usr/include
cp libstemmer.a /usr/lib
cp libstemmer.so.0d.0.0 /usr/lib
ln -s /usr/lib/libstemmer.so.0d /usr/lib/libstemmer.so
ln -s /usr/lib/libstemmer.so.0d.0.0 /usr/lib/libstemmer.so.0d
ldconfig

EOF

This file is need is for the modified GNUmakefile
create a debian dir
mkdir debian

cat > debian/libstemmer.ver<< "EOF"
{
        global:
                sb_*;
        local:
                *;
};
EOF

---

I created a GNUmakefile.patch file from an early version of 2.2.0 diff file , to manually modify the GNUmakefile from the git folder . This was the closest matching diff to patch the GNUmakefile i could find.

The result is :

        libstemmer.so 
        libstemmer.so.0d 
        libstemmer.so.0d.0.0
        libstemmer.a

cat > GNUmakefile.patch<< "EOF"
--- GNUmakefile	2025-01-25 22:35:08.351722964 +1300
+++ GNUmakefile.new	2025-01-25 22:32:49.754808290 +1300
@@ -61,6 +61,7 @@
 # a sibling to this one.
 STEMMING_DATA ?= ../snowball-data
 STEMMING_DATA_ABS := $(abspath $(STEMMING_DATA))
+STEMWORDS=LD_LIBRARY_PATH=.:$(LD_LIBRARY_PATH) ./stemwords$(EXEEXT)
 
 # Keep one in $(THIN_FACTOR) entries from gzipped vocabularies.
 THIN_FACTOR ?= 3
@@ -187,12 +188,12 @@
 JAVA_CLASSES = $(JAVA_SOURCES:.java=.class)
 JAVA_RUNTIME_CLASSES=$(JAVARUNTIME_SOURCES:.java=.class)
 
-CFLAGS=-g -O2 -W -Wall -Wmissing-prototypes -Wmissing-declarations
-CPPFLAGS=
+override CFLAGS += -O2 -W -Wall -Wmissing-prototypes -Wmissing-declarations
+CPPFLAGS?=
 
 INCLUDES=-Iinclude
 
-all: snowball$(EXEEXT) libstemmer.a stemwords$(EXEEXT) $(C_OTHER_SOURCES) $(C_OTHER_HEADERS) $(C_OTHER_OBJECTS)
+all: snowball$(EXEEXT) libstemmer.a libstemmer.so stemwords$(EXEEXT) $(C_OTHER_SOURCES) $(C_OTHER_HEADERS) $(C_OTHER_OBJECTS)
 
 algorithms.mk: libstemmer/mkalgorithms.pl libstemmer/modules.txt
 	libstemmer/mkalgorithms.pl algorithms.mk libstemmer/modules.txt
@@ -200,8 +201,8 @@
 clean:
 	rm -f $(COMPILER_OBJECTS) $(RUNTIME_OBJECTS) \
 	      $(LIBSTEMMER_OBJECTS) $(LIBSTEMMER_UTF8_OBJECTS) $(STEMWORDS_OBJECTS) snowball$(EXEEXT) \
-	      libstemmer.a stemwords$(EXEEXT) \
-              libstemmer/modules.h \
+	      libstemmer.a libstemmer.so stemwords$(EXEEXT) \
+	      libstemmer/modules.h \
               libstemmer/modules_utf8.h \
 	      $(C_LIB_SOURCES) $(C_LIB_HEADERS) $(C_LIB_OBJECTS) \
 	      $(C_OTHER_SOURCES) $(C_OTHER_HEADERS) $(C_OTHER_OBJECTS) \
@@ -215,12 +216,11 @@
 	      stemtest$(EXEEXT) $(STEMTEST_OBJECTS) \
               libstemmer/mkinc.mak libstemmer/mkinc_utf8.mak \
               libstemmer/libstemmer.c libstemmer/libstemmer_utf8.c \
-	      algorithms.mk
+	      algorithms.mk \
+	      stemtest$(EXEEXT) $(STEMTEST_OBJECTS)
 	rm -rf ada/obj dist
-	-rmdir $(c_src_dir)
-	-rmdir $(python_output_dir)
-	-rmdir $(js_output_dir)
-
+	rm -rf dist .shared
+	
 update_version:
 	perl -pi -e 's/(SNOWBALL_VERSION.*?)\d+\.\d+\.\d+/$${1}$(SNOWBALL_VERSION)/' \
 		compiler/header.h \
@@ -254,17 +254,25 @@
 libstemmer.a: libstemmer/libstemmer.o $(RUNTIME_OBJECTS) $(C_LIB_OBJECTS)
 	$(AR) -cru $@ $^
 
+libstemmer.so: libstemmer/libstemmer.o $(RUNTIME_OBJECTS) $(C_LIB_OBJECTS)
+	$(CC) $(CFLAGS) -shared $(LDFLAGS) \
+	      -Wl,--version-script=debian/libstemmer.ver,-soname,libstemmer.so.0d \
+	      -o [email protected] ${^:%=shared/%}
+	ln -s [email protected] [email protected]
+	ln -s [email protected] $@
+	$(AR) -crs ${@:.so=.a} $^
+
 examples/%.o: examples/%.c
 	$(CC) $(CFLAGS) $(INCLUDES) $(CPPFLAGS) -c -o $@ $<
 
-stemwords$(EXEEXT): $(STEMWORDS_OBJECTS) libstemmer.a
-	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $^
+stemwords$(EXEEXT): $(STEMWORDS_OBJECTS) libstemmer.so
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(STEMWORDS_OBJECTS) -L. -lstemmer
 
 tests/%.o: tests/%.c
 	$(CC) $(CFLAGS) $(INCLUDES) $(CPPFLAGS) -c -o $@ $<
 
-stemtest$(EXEEXT): $(STEMTEST_OBJECTS) libstemmer.a
-	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $^
+stemtest$(EXEEXT): $(STEMTEST_OBJECTS) libstemmer.so
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(STEMTEST_OBJECTS) -L. -lstemmer
 
 csharp_stemwords$(EXEEXT): $(CSHARP_STEMWORDS_SOURCES) $(CSHARP_RUNTIME_SOURCES) $(CSHARP_SOURCES)
 	$(MCS) -unsafe -target:exe -out:$@ $(CSHARP_STEMWORDS_SOURCES) $(CSHARP_RUNTIME_SOURCES) $(CSHARP_SOURCES)
@@ -478,7 +486,7 @@
 	cp -a $(PYTHON_RUNTIME_SOURCES) $${dest}/src/$(python_runtime_dir) && \
 	cp -a $(COMMON_FILES) $(PYTHON_PACKAGE_FILES) $${dest} && \
 	(cd $${dest} && $(python) setup.py sdist bdist_wheel && cp dist/*.tar.gz dist/*.whl ..) && \
-	rm -rf $${dest}
+	cp -a $(COMMON_FILES) $(PYTHON_PACKAGE_FILES) $${dest}
 
 dist_libstemmer_js: $(JS_SOURCES) $(COMMON_FILES)
 	destname=jsstemmer-$(SNOWBALL_VERSION); \
@@ -507,7 +515,7 @@
 check: check_stemtest check_utf8 check_iso_8859_1 check_iso_8859_2 check_koi8r
 
 check_stemtest: stemtest$(EXEEXT)
-	./stemtest
+	LD_LIBRARY_PATH=.:$(LD_LIBRARY_PATH) ./stemtest
 
 check_utf8: $(libstemmer_algorithms:%=check_utf8_%)
 
@@ -520,9 +528,9 @@
 check_utf8_%: $(STEMMING_DATA)/% stemwords$(EXEEXT)
 	@echo "Checking output of $* stemmer with UTF-8"
 	@if test -f '$</voc.txt.gz' ; then \
-	  gzip -dc '$</voc.txt.gz'|./stemwords$(EXEEXT) -c UTF_8 -l $* -o tmp.txt; \
+	  gzip -dc '$</voc.txt.gz'|$(STEMWORDS) -c UTF_8 -l `echo $<|sed 's!.*/!!'` -o tmp.txt; \
 	else \
-	  ./stemwords$(EXEEXT) -c UTF_8 -l $* -i $</voc.txt -o tmp.txt; \
+	  $(STEMWORDS) -c UTF_8 -l `echo $<|sed 's!.*/!!'` -i $</voc.txt -o tmp.txt; \
 	fi
 	@if test -f '$</output.txt.gz' ; then \
 	  gzip -dc '$</output.txt.gz'|$(DIFF) -u - tmp.txt; \
@@ -534,7 +542,7 @@
 check_iso_8859_1_%: $(STEMMING_DATA)/% stemwords$(EXEEXT)
 	@echo "Checking output of $* stemmer with ISO_8859_1"
 	@$(ICONV) -f UTF-8 -t ISO-8859-1 '$</voc.txt' |\
-	    ./stemwords -c ISO_8859_1 -l $* -o tmp.txt
+	    $(STEMWORDS) -c ISO_8859_1 -l `echo $<|sed 's!.*/!!'` -o tmp.txt
 	@$(ICONV) -f UTF-8 -t ISO-8859-1 '$</output.txt' |\
 	    $(DIFF) -u - tmp.txt
 	@rm tmp.txt
@@ -542,7 +550,7 @@
 check_iso_8859_2_%: $(STEMMING_DATA)/% stemwords$(EXEEXT)
 	@echo "Checking output of $* stemmer with ISO_8859_2"
 	@$(ICONV) -f UTF-8 -t ISO-8859-2 '$</voc.txt' |\
-	    ./stemwords -c ISO_8859_2 -l $* -o tmp.txt
+	    $(STEMWORDS) -c ISO_8859_2 -l `echo $<|sed 's!.*/!!'` -o tmp.txt
 	@$(ICONV) -f UTF-8 -t ISO-8859-2 '$</output.txt' |\
 	    $(DIFF) -u - tmp.txt
 	@rm tmp.txt
@@ -550,7 +558,7 @@
 check_koi8r_%: $(STEMMING_DATA)/% stemwords$(EXEEXT)
 	@echo "Checking output of $* stemmer with KOI8R"
 	@$(ICONV) -f UTF-8 -t KOI8-R '$</voc.txt' |\
-	    ./stemwords -c KOI8_R -l $* -o tmp.txt
+	    $(STEMWORDS) -c KOI8_R -l `echo $<|sed 's!.*/!!'` -o tmp.txt
 	@$(ICONV) -f UTF-8 -t KOI8-R '$</output.txt' |\
 	    $(DIFF) -u - tmp.txt
 	@rm tmp.txt
@@ -795,4 +803,10 @@
 ada/bin/stemwords: $(ADA_SOURCES)
 	cd ada && $(gprbuild) -Pstemwords -p
 
+%.o: %.c
+	@mkdir -p $(shell dirname ${@:%=shared/%})
+	$(CC) $(CFLAGS) $(CPPFLAGS) -c -fPIC -o ${@:%=shared/%} $<
+	$(CC) $(CFLAGS) $(CPPFLAGS) -c -o $@ $<
+
+
 .SUFFIXES: .class .java

EOF

[sidrat]

Sorry the GNUmakefile.patch i posted earlier did not format well so i have placed it in my old git project file with a link for those who want the patch

https://github.com/sidrat/mylfsos/blob/master/patches/GNUmakefile.patch

[eli-schwartz]

The text box for GitHub comments says "use markdown to format your comment" as a hint that you should use markdown. Don't paste raw text, paste triple-backticks enclosed code blocks.

You can edit your previous comment to fix that mistake (or delete it entirely).

[sidrat]

thanks for the hint , i don't use markdown much.

[ojwb]

In general, if you want to submit a patch then open a PR instead of pasting patch text into comments or putting a patch file in some random other git repo.

However for this patch that wouldn't actually help as I can't merge a patch which adds shared library support but with a hardcoded assumptions that are wrong for some platforms (not all platforms name shared libraries with a.so extension and LD_LIBRARY_PATH isn't universal either). This is why none of the previous PRs for shared library support have been merged.

[eli-schwartz]

What constraints are there on how to add shared library support? Does it have to use a GNUmakefile or is using another build system (automake, meson) on the table?