Merge branch 'master' of ssh://git.indexdata.com/home/git/pub/idzebra

diff --git a/Doxyfile.in b/Doxyfile.in
index 2f64226..c8502a5 100644 (file)
--- a/Doxyfile.in
+++ b/Doxyfile.in
@@ -154,13 +154,6 @@ QT_AUTOBRIEF           = NO
 
 MULTILINE_CPP_IS_BRIEF = NO
 
-# If the DETAILS_AT_TOP tag is set to YES then Doxygen 
-# will output the detailed description near the top, like JavaDoc.
-# If set to NO, the detailed description appears after the member 
-# documentation.
-
-DETAILS_AT_TOP         = NO
-
 # If the INHERIT_DOCS tag is set to YES (the default) then an undocumented 
 # member inherits the documentation from any documented member that it 
 # re-implements.

diff --git a/NEWS b/NEWS
index ffb4d50..8483ed0 100644 (file)
--- a/NEWS
+++ b/NEWS
@@ -1,3 +1,13 @@
+--- 2.0.39 2009/07/09
+
+Improved speed of record fetching by using a different block sizes
+for recd{A,B}.
+
+Added zlib-compression of recd-storage (record data). This is enabled
+by using recordCompression: zlib in zebra.cfg.
+
+Dictionary delete more agressive (dict_del_string).
+
 --- 2.0.38 2009/06/19
 
 Improved estimated hits for AND operation (bug #2907)

diff --git a/configure.ac b/configure.ac
index 936f76a..a362772 100644 (file)
--- a/configure.ac
+++ b/configure.ac
@@ -2,7 +2,7 @@ dnl This file is part of the Zebra server.
 dnl   Copyright (C) 1994-2009 Index Data
 dnl
 AC_PREREQ(2.60)
-AC_INIT([idzebra],[2.0.38],[zebra-help@indexdata.dk])
+AC_INIT([idzebra],[2.0.39],[zebra-help@indexdata.dk])
 AC_CONFIG_SRCDIR(configure.ac)
 AC_CONFIG_AUX_DIR(config)
 AM_INIT_AUTOMAKE([1.9])
@@ -174,6 +174,11 @@ else
     fi
 fi
 dnl
+dnl ------- zlib
+AC_CHECK_LIB([z],[compress2])
+if test "$ac_cv_lib_z_compress2" = "yes"; then
+    AC_CHECK_HEADERS([zlib.h])
+fi
 dnl ------ -lm
 AC_CHECK_LIB(m,sqrt)
 dnl ------ -ldl

diff --git a/debian/changelog b/debian/changelog
index 48a92df..a4a7c94 100644 (file)
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,9 @@
+idzebra (2.0.39-1indexdata) unstable; urgency=low
+
+  * Upstream.
+
+ -- Adam Dickmeiss <adam@indexdata.dk>  Thu, 09 Jul 2009 10:32:04 +0200
+
 idzebra (2.0.38-1indexdata) unstable; urgency=low
 
   * Upstream.

diff --git a/debian/control b/debian/control
index 5e98673..61afbb6 100644 (file)
--- a/debian/control
+++ b/debian/control
@@ -2,7 +2,7 @@ Source: idzebra
 Priority: optional
 Maintainer: Adam Dickmeiss <adam@indexdata.dk>
 Standards-Version: 3.6.0
-Build-Depends: debhelper (>= 4.0.0), libyaz3-dev (>= 3.0.30), libexpat1-dev, tcl8.4-dev, libbz2-dev, libxslt1-dev
+Build-Depends: debhelper (>= 4.0.0), libyaz3-dev (>= 3.0.30), libexpat1-dev, tcl8.4-dev, libbz2-dev, libxslt1-dev, zlib1g-dev
 
 Package: idzebra-2.0
 Section: text

diff --git a/examples/gils/zebra.cfg b/examples/gils/zebra.cfg
index 1fe604b..72b13ea 100644 (file)
--- a/examples/gils/zebra.cfg
+++ b/examples/gils/zebra.cfg
@@ -14,6 +14,9 @@ isam: b
 # Flags for the non-authenticated user. w=write (allows ES Updates)
 perm.anonymous: rw
 
+# Record compression (none, bzip2 or zlib)
+recordcompression: none
+
 #rank: zvrank
 recordId: (bib-1,title)
 

diff --git a/idzebra.spec.in b/idzebra.spec.in
index 1355f00..46cf0d2 100644 (file)
--- a/idzebra.spec.in
+++ b/idzebra.spec.in
@@ -9,7 +9,7 @@ Source: idzebra-%{version}.tar.gz
 BuildRoot: %{_tmppath}/idzebra-%{version}-root
 Packager: Adam Dickmeiss <adam@indexdata.dk>
 URL: http://www.indexdata.dk/zebra/
-BuildRequires: libyaz3-devel expat-devel bzip2-devel tcl
+BuildRequires: libyaz3-devel expat-devel bzip2-devel tcl zlib-devel
 Summary: High-performance, structured text indexing and retrival engine.
 
 %description

diff --git a/index/recindex.h b/index/recindex.h
index a3078c1..8e854ce 100644 (file)
--- a/index/recindex.h
+++ b/index/recindex.h
@@ -85,14 +85,26 @@ ZEBRA_RES rec_close (Records *p);
 */
 Records rec_open(BFiles bfs, int rw, int compression_method);
 
+/** \brief check whether a compression method is supported
+    \param compression_method (REC_COMPRESS_..)
+    \retval 0 if method is unsupported
+    \retval 1 if method is supported
+*/
+int rec_check_compression_method(int compression_method);
+
 char *rec_strdup(const char *s, size_t *len);
 void rec_prstat(Records p, int verbose);
 
 zint rec_sysno_to_int(zint sysno);
 
-/** \brief compression types */
+
+/** \brief No compression ("none") */
 #define REC_COMPRESS_NONE   0
+/** \brief BZIP2 compression (slow and requires big chunks) */
 #define REC_COMPRESS_BZIP2  1
+/** \brief zlib compression (faster and works off small chunks) */
+#define REC_COMPRESS_ZLIB   2
+
 
 enum { 
     recInfo_fileType, 

diff --git a/index/records.c b/index/records.c
index ac05fa8..797ffa5 100644 (file)
--- a/index/records.c
+++ b/index/records.c
@@ -44,6 +44,9 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #if HAVE_BZLIB_H
 #include <bzlib.h>
 #endif
+#if HAVE_ZLIB_H
+#include <zlib.h>
+#endif
 
 #define REC_BLOCK_TYPES 2
 #define REC_HEAD_MAGIC "recindex"
@@ -277,6 +280,28 @@ static ZEBRA_RES rec_write_tmp_buf(Records p, int size, zint *sysnos)
     return ZEBRA_OK;
 }
 
+int rec_check_compression_method(int compression_method)
+{
+    switch(compression_method)
+    {
+    case REC_COMPRESS_ZLIB:
+#if HAVE_ZLIB_H
+        return 1;
+#else
+        return 0;
+#endif
+    case REC_COMPRESS_BZIP2:
+#if HAVE_BZLIB_H
+        return 1;
+#else
+        return 0;
+#endif
+    case REC_COMPRESS_NONE:
+        return 1;
+    }
+    return 0;
+}
+
 Records rec_open(BFiles bfs, int rw, int compression_method)
 {
     Records p;
@@ -492,6 +517,32 @@ static ZEBRA_RES rec_flush_shared(Records p, short ref_count, zint *sysnos,
        char compression_method = p->compression_method;
        switch (compression_method)
        {
+        case REC_COMPRESS_ZLIB:
+#if HAVE_ZLIB_H
+           csize = out_offset + (out_offset >> 6) + 620;
+            while (1)
+            {
+                int r;
+                uLongf destLen = csize;
+                rec_tmp_expand(p, csize);
+                r = compress((Bytef *) p->tmp_buf+sizeof(zint)+sizeof(short)+
+                             sizeof(char),
+                             &destLen, (const Bytef *) out_buf, out_offset);
+                csize = destLen;
+                if (r == Z_OK)
+                {
+                    break;
+                }
+                if (r != Z_MEM_ERROR)
+                {
+                    yaz_log(YLOG_WARN, "compress error: %d", r);
+                    csize = 0;
+                    break;
+                }
+                csize = csize * 2;
+            }
+#endif
+            break;
        case REC_COMPRESS_BZIP2:
 #if HAVE_BZLIB_H       
            csize = out_offset + (out_offset >> 6) + 620;
@@ -509,8 +560,6 @@ static ZEBRA_RES rec_flush_shared(Records p, short ref_count, zint *sysnos,
                yaz_log(YLOG_WARN, "bzBuffToBuffCompress error code=%d", i);
                csize = 0;
            }
-           yaz_log(YLOG_LOG, "compress %4d %5d %5d", ref_count, out_offset,
-                 csize);
 #endif
            break;
        case REC_COMPRESS_NONE:
@@ -708,9 +757,6 @@ static Record rec_get_int(Records p, zint sysno)
     char *nptr, *cptr;
     char *in_buf = 0;
     char *bz_buf = 0;
-#if HAVE_BZLIB_H
-    unsigned int bz_size;
-#endif
     char compression_method;
 
     assert(sysno > 0);
@@ -761,27 +807,57 @@ static Record rec_get_int(Records p, zint sysno)
     in_size = entry.size - sizeof(short) - sizeof(char);
     switch (compression_method)
     {
+    case REC_COMPRESS_ZLIB:
+#if HAVE_ZLIB_H
+        if (1)
+        {
+            unsigned int bz_size = entry.size * 20 + 100;
+            while (1)
+            {
+                uLongf destLen = bz_size;
+                bz_buf = (char *) xmalloc(bz_size);
+                i = uncompress((Bytef *) bz_buf, &destLen,
+                               (const Bytef *) in_buf, in_size);
+                if (i == Z_OK)
+                {
+                    bz_size = destLen; 
+                    break;
+                }
+                yaz_log(YLOG_LOG, "failed");
+                xfree(bz_buf);
+                bz_size *= 2;
+            }
+            in_buf = bz_buf;
+            in_size = bz_size;
+        }
+#else
+        yaz_log(YLOG_FATAL, "cannot decompress record(s) in ZLIB format");
+        return 0;
+#endif
+        break;
     case REC_COMPRESS_BZIP2:
 #if HAVE_BZLIB_H
-       bz_size = entry.size * 20 + 100;
-       while (1)
-       {
-           bz_buf = (char *) xmalloc(bz_size);
+        if (1)
+        {
+            unsigned int bz_size = entry.size * 20 + 100;
+            while (1)
+            {
+                bz_buf = (char *) xmalloc(bz_size);
 #ifdef BZ_CONFIG_ERROR
-           i = BZ2_bzBuffToBuffDecompress
+                i = BZ2_bzBuffToBuffDecompress
 #else
-           i = bzBuffToBuffDecompress
+                    i = bzBuffToBuffDecompress
 #endif
-                (bz_buf, &bz_size, in_buf, in_size, 0, 0);
-           yaz_log(YLOG_LOG, "decompress %5d %5d", in_size, bz_size);
-           if (i == BZ_OK)
-               break;
-           yaz_log(YLOG_LOG, "failed");
-           xfree(bz_buf);
-            bz_size *= 2;
-       }
-       in_buf = bz_buf;
-       in_size = bz_size;
+                    (bz_buf, &bz_size, in_buf, in_size, 0, 0);
+                if (i == BZ_OK)
+                    break;
+                yaz_log(YLOG_LOG, "failed");
+                xfree(bz_buf);
+                bz_size *= 2;
+            }
+            in_buf = bz_buf;
+            in_size = bz_size;
+        }
 #else
        yaz_log(YLOG_FATAL, "cannot decompress record(s) in BZIP2 format");
        return 0;

diff --git a/index/zebraapi.c b/index/zebraapi.c
index 2b69a2d..d23f8ca 100644 (file)
--- a/index/zebraapi.c
+++ b/index/zebraapi.c
@@ -334,7 +334,7 @@ struct zebra_register *zebra_register_open(ZebraService zs, const char *name,
 {
     struct zebra_register *reg;
     int record_compression = REC_COMPRESS_NONE;
-    const char *recordCompression = 0;
+    const char *compression_str = 0;
     const char *profilePath;
     int sort_type = ZEBRA_SORT_TYPE_FLAT;
     ZEBRA_RES ret = ZEBRA_OK;
@@ -420,11 +420,25 @@ struct zebra_register *zebra_register_open(ZebraService zs, const char *name,
     zebraRankInstall(reg, rank_similarity_class);
     zebraRankInstall(reg, rank_static_class);
 
-    recordCompression = res_get_def(res, "recordCompression", "none");
-    if (!strcmp(recordCompression, "none"))
+    compression_str = res_get_def(res, "recordCompression", "none");
+    if (!strcmp(compression_str, "none"))
        record_compression = REC_COMPRESS_NONE;
-    if (!strcmp(recordCompression, "bzip2"))
+    else if (!strcmp(compression_str, "bzip2"))
        record_compression = REC_COMPRESS_BZIP2;
+    else if (!strcmp(compression_str, "zlib"))
+       record_compression = REC_COMPRESS_ZLIB;
+    else
+    {
+        yaz_log(YLOG_FATAL, "invalid recordCompression: %s", compression_str);
+        ret = ZEBRA_FAIL;
+    }
+
+    if (!rec_check_compression_method(record_compression))
+    {
+        yaz_log(YLOG_FATAL, "unsupported recordCompression: %s",
+                compression_str);
+        ret = ZEBRA_FAIL;
+    }
 
     {
        const char *index_fname = res_get_def(res, "index", "default.idx");