Dumps the contents of indexs.

2000-10-24  Not Zed  <NotZed@HelixCode.com>

        * dumpindex.c: Dumps the contents of indexs.

        * hash.c (ibex_hash_dump_rec): Also print the word count.

        * wordindex.c (unindex_name): Cross-check the cache as well.

svn path=/trunk/; revision=6139

5 files changed, 91 insertions, 6 deletions

diff --git a/libibex/ChangeLog b/libibex/ChangeLog
index ec753aa483..212169fcfa 100644
--- a/libibex/ChangeLog
+++ b/libibex/ChangeLog
@@ -1,3 +1,11 @@
+2000-10-24  Not Zed  <NotZed@HelixCode.com>
+
+	* dumpindex.c: Dumps the contents of indexs.
+
+	* hash.c (ibex_hash_dump_rec): Also print the word count.
+
+	* wordindex.c (unindex_name): Cross-check the cache as well.
+
 2000-10-12  Not Zed  <NotZed@HelixCode.com>
 
 	* ibex_internal.h (struct ibex): Bumped ibex rev.
diff --git a/libibex/Makefile.am b/libibex/Makefile.am
index cfd8a1b4d5..fdde10aca3 100644
--- a/libibex/Makefile.am
+++ b/libibex/Makefile.am
@@ -21,10 +21,10 @@ INCLUDES = -I$(srcdir) $(GLIB_CFLAGS) $(UNICODE_CFLAGS) \
 		-DG_LOG_DOMAIN=\"libibex\"
 
 
-#noinst_PROGRAMS = hash
+noinst_PROGRAMS = dumpindex
 
-#hash_SOURCES = disktail.c
-#hash_LDADD = libibex.la $(GLIB_LIBS) $(UNICODE_LIBS)
+dumpindex_SOURCES = dumpindex.c
+dumpindex_LDADD = libibex.la $(GLIB_LIBS) $(UNICODE_LIBS)
 
 #noinst_PROGRAMS = mkindex lookup
 #
diff --git a/libibex/dumpindex.c b/libibex/dumpindex.c
new file mode 100644
index 0000000000..92f4b08845
--- /dev/null
+++ b/libibex/dumpindex.c
@@ -0,0 +1,32 @@
+/*
+  Dump the hash tables from an ibex file.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ibex_internal.h"
+
+extern void ibex_hash_dump(struct _IBEXIndex *index);
+
+int main(int argc, char **argv)
+{
+	ibex *ib;
+
+	if (argc != 2) {
+		printf("Usage: %s ibexfile\n", argv[0]);
+		return 1;
+	}
+	ib = ibex_open(argv[1], O_RDONLY, 0);
+	if (ib == NULL) {
+		perror("Opening ibex file\n");
+		return 1;
+	}
+
+	ibex_hash_dump(ib->words->wordindex);
+	ibex_hash_dump(ib->words->nameindex);
+
+	ibex_close(ib);
+
+	return 0;
+}
diff --git a/libibex/hash.c b/libibex/hash.c
index 25a6e92ef9..9395e1e00b 100644
--- a/libibex/hash.c
+++ b/libibex/hash.c
@@ -622,9 +622,20 @@ hash_insert(struct _IBEXIndex *index, const char *key, int keylen)
 
 /* debug */
 void ibex_hash_dump(struct _IBEXIndex *index);
+static void ibex_hash_dump_rec(struct _IBEXIndex *index, int *words, int *wordslen);
 
-void
-ibex_hash_dump(struct _IBEXIndex *index)
+void ibex_hash_dump(struct _IBEXIndex *index)
+{
+	int words = 0, wordslen=0;
+
+	ibex_hash_dump_rec(index, &words, &wordslen);
+
+	printf("Total words = %d, bytes = %d, ave length = %f\n", words, wordslen, (double)wordslen/(double)words);
+}
+
+
+static void
+ibex_hash_dump_rec(struct _IBEXIndex *index, int *words, int *wordslen)
 {
 	int i;
 	struct _hashtableblock *table;
@@ -643,6 +654,8 @@ ibex_hash_dump(struct _IBEXIndex *index)
 		while (hashbucket) {
 			int len;
 
+			*words = *words + 1;
+
 			bucket = (struct _hashblock *)ibex_block_read(index->blocks, HASH_BLOCK(hashbucket));
 			printf(" bucket %d: [used %d]", hashbucket, bucket->used);
 			if (HASH_INDEX(hashbucket) == 0) {
@@ -655,6 +668,9 @@ ibex_hash_dump(struct _IBEXIndex *index)
 			printf("'%.*s' = %d next=%d\n", len, &bucket->hb_keydata[bucket->hb_keys[HASH_INDEX(hashbucket)].keyoffset],
 			       bucket->hb_keys[HASH_INDEX(hashbucket)].root,
 			       bucket->hb_keys[HASH_INDEX(hashbucket)].next);
+
+			*wordslen = *wordslen + len;
+
 			ibex_diskarray_dump(index->blocks,
 					    bucket->hb_keys[HASH_INDEX(hashbucket)].root << BLOCK_BITS,
 					    bucket->hb_keys[HASH_INDEX(hashbucket)].tail);
diff --git a/libibex/wordindex.c b/libibex/wordindex.c
index da25389b27..62e8859e15 100644
--- a/libibex/wordindex.c
+++ b/libibex/wordindex.c
@@ -145,6 +145,8 @@ static void unindex_name(struct _IBEXWord *idx, const char *name)
 	int i;
 	nameid_t nameid, wordid;
 	blockid_t nameblock, wordblock, newblock, nametail, wordtail, newtail;
+	char *word;
+	struct _wordcache *cache;
 
 	d(printf("unindexing %s\n", name));
 
@@ -168,7 +170,34 @@ static void unindex_name(struct _IBEXWord *idx, const char *name)
 		if (newblock != wordblock || newtail != wordtail)
 			idx->wordindex->klass->set_data(idx->wordindex, wordid, newblock, newtail);
 
-		/* FIXME: check cache as well */
+		/* now check the cache as well */
+		word = idx->nameindex->klass->get_key(idx->wordindex, wordid, NULL);
+		if (word) {
+			cache = g_hash_table_lookup(idx->wordcache, word);
+			if (cache) {
+				/* its there, update our head/tail pointers */
+				cache->wordblock = newblock;
+				cache->wordtail = newtail;
+
+				/* now check that we have a data entry in it */
+				if (cache->filealloc == 0 && cache->filecount == 1) {
+					if (cache->file.file0 == nameid) {
+						cache->filecount = 0;
+					}
+				} else {
+					int j;
+
+					for (j=0;j<cache->filecount;j++) {
+						if (cache->file.files[j] == nameid) {
+							cache->file.files[j] = cache->file.files[cache->filecount-1];
+							cache->filecount--;
+							break;
+						}
+					}
+				}
+			}
+			g_free(word);
+		}
 	}
 	g_array_free(words, TRUE);