Get estimated package count when populating databases

This works for both local and sync databases in slightly different ways. For
the local database, we can use the directory hard link count on the local/
folder. For sync databases, we use the archive size coupled with some
computed average per-package sizes to determine an estimate.

This is currently a dead assignment once calculated, but could be used to
set the initial size of a hash table.

Signed-off-by: Dan McGee <dan@archlinux.org>
Signed-off-by: Allan McRae <allan@archlinux.org>

1 files changed, 63 insertions, 1 deletions

diff --git a/lib/libalpm/be_sync.c b/lib/libalpm/be_sync.c
index f7101f54..0c968b45 100644
--- a/lib/libalpm/be_sync.c
+++ b/lib/libalpm/be_sync.c
@@ -145,9 +145,67 @@ int SYMEXPORT alpm_db_update(int force, pmdb_t *db)
 static int sync_db_read(pmdb_t *db, struct archive *archive,
 		struct archive_entry *entry, pmpkg_t *likely_pkg);
 
+/*
+ * This is the data table used to generate the estimating function below.
+ * "Weighted Avg" means averaging the bottom table values; thus each repo, big
+ * or small, will have equal influence.  "Unweighted Avg" means averaging the
+ * sums of the top table columns, thus each package has equal influence.  The
+ * final values are calculated by (surprise) averaging the averages, because
+ * why the hell not.
+ *
+ * Database   Pkgs  tar      bz2     gz      xz
+ * community  2096  5294080  256391  421227  301296
+ * core        180   460800   25257   36850   29356
+ * extra      2606  6635520  294647  470818  339392
+ * multilib    126   327680   16120   23261   18732
+ * testing      76   204800   10902   14348   12100
+ *
+ * Bytes Per Package
+ * community  2096  2525.80  122.32  200.97  143.75
+ * core        180  2560.00  140.32  204.72  163.09
+ * extra      2606  2546.25  113.06  180.67  130.23
+ * multilib    126  2600.63  127.94  184.61  148.67
+ * testing      76  2694.74  143.45  188.79  159.21
+
+ * Weighted Avg     2585.48  129.42  191.95  148.99
+ * Unweighted Avg   2543.39  118.74  190.16  137.93
+ * Average of Avgs  2564.44  124.08  191.06  143.46
+ */
+static int estimate_package_count(struct stat *st, struct archive *archive)
+{
+	unsigned int per_package;
+
+	switch(archive_compression(archive)) {
+		case ARCHIVE_COMPRESSION_NONE:
+			per_package = 2564;
+			break;
+		case ARCHIVE_COMPRESSION_GZIP:
+			per_package = 191;
+			break;
+		case ARCHIVE_COMPRESSION_BZIP2:
+			per_package = 124;
+			break;
+		case ARCHIVE_COMPRESSION_COMPRESS:
+			per_package = 193;
+			break;
+		case ARCHIVE_COMPRESSION_LZMA:
+		case ARCHIVE_COMPRESSION_XZ:
+			per_package = 143;
+			break;
+		case ARCHIVE_COMPRESSION_UU:
+			per_package = 3543;
+			break;
+		default:
+			/* assume it is at least somewhat compressed */
+			per_package = 200;
+	}
+	return((int)(st->st_size / per_package) + 1);
+}
+
 static int sync_db_populate(pmdb_t *db)
 {
-	int count = 0;
+	int est_count, count = 0;
+	struct stat buf;
 	struct archive *archive;
 	struct archive_entry *entry;
 	pmpkg_t *pkg = NULL;
@@ -169,6 +227,10 @@ static int sync_db_populate(pmdb_t *db)
 		archive_read_finish(archive);
 		RET_ERR(PM_ERR_DB_OPEN, 1);
 	}
+	if(lstat(_alpm_db_path(db), &buf) != 0) {
+		RET_ERR(PM_ERR_DB_OPEN, 1);
+	}
+	est_count = estimate_package_count(&buf, archive);
 
 	while(archive_read_next_header(archive, &entry) == ARCHIVE_OK) {
 		const struct stat *st;