[pacman-dev] [PATCH 1/2] Make sync DB reading a bit more flexible
Dan McGee
dan at archlinux.org
Fri Jun 24 01:18:46 EDT 2011
We can reorganize things a bit to not require reading a directory-only entry
first (or at all). This was noticed while working on some pactest
improvements, but should be a good step forward anyway.
Also make _alpm_splitname() a bit more generic in where it stores the data it
parses.
Signed-off-by: Dan McGee <dan at archlinux.org>
---
lib/libalpm/be_local.c | 3 +-
lib/libalpm/be_sync.c | 130 ++++++++++++++++++++++++-----------------------
lib/libalpm/util.c | 46 ++++++++++-------
lib/libalpm/util.h | 3 +-
4 files changed, 97 insertions(+), 85 deletions(-)
diff --git a/lib/libalpm/be_local.c b/lib/libalpm/be_local.c
index 4b2a301..9e4de54 100644
--- a/lib/libalpm/be_local.c
+++ b/lib/libalpm/be_local.c
@@ -382,7 +382,8 @@ static int local_db_populate(pmdb_t *db)
RET_ERR(db->handle, PM_ERR_MEMORY, -1);
}
/* split the db entry name */
- if(_alpm_splitname(name, pkg) != 0) {
+ if(_alpm_splitname(name, &(pkg->name), &(pkg->version),
+ &(pkg->name_hash)) != 0) {
_alpm_log(db->handle, PM_LOG_ERROR, _("invalid name for database entry '%s'\n"),
name);
_alpm_pkg_free(pkg);
diff --git a/lib/libalpm/be_sync.c b/lib/libalpm/be_sync.c
index f0b1736..6d1ab39 100644
--- a/lib/libalpm/be_sync.c
+++ b/lib/libalpm/be_sync.c
@@ -182,7 +182,61 @@ cleanup:
/* Forward decl so I don't reorganize the whole file right now */
static int sync_db_read(pmdb_t *db, struct archive *archive,
- struct archive_entry *entry, pmpkg_t *likely_pkg);
+ struct archive_entry *entry, pmpkg_t **likely_pkg);
+
+static pmpkg_t *load_pkg_for_entry(pmdb_t *db, const char *entryname,
+ const char **entry_filename, pmpkg_t *likely_pkg)
+{
+ char *pkgname = NULL, *pkgver = NULL;
+ unsigned long pkgname_hash;
+ pmpkg_t *pkg;
+
+ /* get package and db file names */
+ if(entry_filename) {
+ char *fname = strrchr(entryname, '/');
+ if(fname) {
+ *entry_filename = fname + 1;
+ } else {
+ *entry_filename = NULL;
+ }
+ }
+ if(_alpm_splitname(entryname, &pkgname, &pkgver, &pkgname_hash) != 0) {
+ _alpm_log(db->handle, PM_LOG_ERROR,
+ _("invalid name for database entry '%s'\n"), entryname);
+ return NULL;
+ }
+
+ if(likely_pkg && strcmp(likely_pkg->name, pkgname) == 0) {
+ pkg = likely_pkg;
+ } else {
+ pkg = _alpm_pkghash_find(db->pkgcache, pkgname);
+ }
+ if(pkg == NULL) {
+ pkg = _alpm_pkg_new();
+ if(pkg == NULL) {
+ RET_ERR(db->handle, PM_ERR_MEMORY, NULL);
+ }
+
+ pkg->name = pkgname;
+ pkg->version = pkgver;
+ pkg->name_hash = pkgname_hash;
+
+ pkg->origin = PKG_FROM_SYNCDB;
+ pkg->origin_data.db = db;
+ pkg->ops = &default_pkg_ops;
+ pkg->handle = db->handle;
+
+ /* add to the collection */
+ _alpm_log(db->handle, PM_LOG_FUNCTION, "adding '%s' to package cache for db '%s'\n",
+ pkg->name, db->treename);
+ db->pkgcache = _alpm_pkghash_add(db->pkgcache, pkg);
+ } else {
+ free(pkgname);
+ free(pkgver);
+ }
+
+ return pkg;
+}
/*
* This is the data table used to generate the estimating function below.
@@ -292,55 +346,20 @@ static int sync_db_populate(pmdb_t *db)
st = archive_entry_stat(entry);
if(S_ISDIR(st->st_mode)) {
- const char *name;
-
- pkg = _alpm_pkg_new();
- if(pkg == NULL) {
- archive_read_finish(archive);
- RET_ERR(db->handle, PM_ERR_MEMORY, -1);
- }
-
- name = archive_entry_pathname(entry);
-
- if(_alpm_splitname(name, pkg) != 0) {
- _alpm_log(db->handle, PM_LOG_ERROR, _("invalid name for database entry '%s'\n"),
- name);
- _alpm_pkg_free(pkg);
- pkg = NULL;
- continue;
- }
-
- /* duplicated database entries are not allowed */
- if(_alpm_pkghash_find(db->pkgcache, pkg->name)) {
- _alpm_log(db->handle, PM_LOG_ERROR, _("duplicated database entry '%s'\n"), pkg->name);
- _alpm_pkg_free(pkg);
- pkg = NULL;
- continue;
- }
-
- pkg->origin = PKG_FROM_SYNCDB;
- pkg->origin_data.db = db;
- pkg->ops = &default_pkg_ops;
- pkg->handle = db->handle;
-
- /* add to the collection */
- _alpm_log(db->handle, PM_LOG_FUNCTION, "adding '%s' to package cache for db '%s'\n",
- pkg->name, db->treename);
- db->pkgcache = _alpm_pkghash_add(db->pkgcache, pkg);
- count++;
+ continue;
} else {
/* we have desc, depends or deltas - parse it */
- if(sync_db_read(db, archive, entry, pkg) != 0) {
+ if(sync_db_read(db, archive, entry, &pkg) != 0) {
_alpm_log(db->handle, PM_LOG_ERROR,
_("could not parse package '%s' description file from db '%s'\n"),
pkg->name, db->treename);
- _alpm_pkg_free(pkg);
- pkg = NULL;
continue;
}
}
}
+ count = alpm_list_count(db->pkgcache->list);
+
if(count > 0) {
db->pkgcache->list = alpm_list_msort(db->pkgcache->list, (size_t)count, _alpm_pkg_cmp);
}
@@ -370,10 +389,9 @@ static int sync_db_populate(pmdb_t *db)
} while(1) /* note the while(1) and not (0) */
static int sync_db_read(pmdb_t *db, struct archive *archive,
- struct archive_entry *entry, pmpkg_t *likely_pkg)
+ struct archive_entry *entry, pmpkg_t **likely_pkg)
{
const char *entryname, *filename;
- char *pkgname, *p, *q;
pmpkg_t *pkg;
struct archive_read_buffer buf;
@@ -391,27 +409,12 @@ static int sync_db_read(pmdb_t *db, struct archive *archive,
/* 512K for a line length seems reasonable */
buf.max_line_size = 512 * 1024;
- /* get package and db file names */
- STRDUP(pkgname, entryname, RET_ERR(db->handle, PM_ERR_MEMORY, -1));
- p = pkgname + strlen(pkgname);
- for(q = --p; *q && *q != '/'; q--);
- filename = q + 1;
- for(p = --q; *p && *p != '-'; p--);
- for(q = --p; *q && *q != '-'; q--);
- *q = '\0';
-
- /* package is already in db due to parsing of directory name */
- if(likely_pkg && strcmp(likely_pkg->name, pkgname) == 0) {
- pkg = likely_pkg;
- } else {
- if(db->pkgcache == NULL) {
- RET_ERR(db->handle, PM_ERR_MEMORY, -1);
- }
- pkg = _alpm_pkghash_find(db->pkgcache, pkgname);
- }
+ pkg = load_pkg_for_entry(db, entryname, &filename, *likely_pkg);
+
if(pkg == NULL) {
- _alpm_log(db->handle, PM_LOG_DEBUG, "package %s not found in %s sync database",
- pkgname, db->treename);
+ _alpm_log(db->handle, PM_LOG_DEBUG,
+ "entry %s could not be loaded into %s sync database",
+ entryname, db->treename);
return -1;
}
@@ -498,6 +501,7 @@ static int sync_db_read(pmdb_t *db, struct archive *archive,
if(ret != ARCHIVE_EOF) {
goto error;
}
+ *likely_pkg = pkg;
} else if(strcmp(filename, "files") == 0) {
/* currently do nothing with this file */
} else {
@@ -505,12 +509,10 @@ static int sync_db_read(pmdb_t *db, struct archive *archive,
_alpm_log(db->handle, PM_LOG_DEBUG, "unknown database file: %s\n", filename);
}
- FREE(pkgname);
return 0;
error:
_alpm_log(db->handle, PM_LOG_DEBUG, "error parsing database file: %s\n", filename);
- FREE(pkgname);
return -1;
}
diff --git a/lib/libalpm/util.c b/lib/libalpm/util.c
index 4976703..e56c9f2 100644
--- a/lib/libalpm/util.c
+++ b/lib/libalpm/util.c
@@ -825,46 +825,54 @@ cleanup:
}
}
-int _alpm_splitname(const char *target, pmpkg_t *pkg)
+int _alpm_splitname(const char *target, char **name, char **version,
+ unsigned long *name_hash)
{
/* the format of a db entry is as follows:
* package-version-rel/
+ * package-version-rel/desc (we ignore the filename portion)
* package name can contain hyphens, so parse from the back- go back
* two hyphens and we have split the version from the name.
*/
- const char *version, *end;
+ const char *pkgver, *end;
- if(target == NULL || pkg == NULL) {
+ if(target == NULL) {
return -1;
}
- end = target + strlen(target);
- /* remove any trailing '/' */
- while(*(end - 1) == '/') {
- --end;
+ /* remove anything trailing a '/' */
+ end = strchr(target, '/');
+ if(!end) {
+ end = target + strlen(target);
}
/* do the magic parsing- find the beginning of the version string
* by doing two iterations of same loop to lop off two hyphens */
- for(version = end - 1; *version && *version != '-'; version--);
- for(version = version - 1; *version && *version != '-'; version--);
- if(*version != '-' || version == target) {
+ for(pkgver = end - 1; *pkgver && *pkgver != '-'; pkgver--);
+ for(pkgver = pkgver - 1; *pkgver && *pkgver != '-'; pkgver--);
+ if(*pkgver != '-' || pkgver == target) {
return -1;
}
/* copy into fields and return */
- if(pkg->version) {
- FREE(pkg->version);
+ if(version) {
+ if(*version) {
+ FREE(*version);
+ }
+ /* version actually points to the dash, so need to increment 1 and account
+ * for potential end character */
+ STRNDUP(*version, pkgver + 1, end - pkgver - 1, return -1);
}
- /* version actually points to the dash, so need to increment 1 and account
- * for potential end character */
- STRNDUP(pkg->version, version + 1, end - version - 1, return -1);
- if(pkg->name) {
- FREE(pkg->name);
+ if(name) {
+ if(*name) {
+ FREE(*name);
+ }
+ STRNDUP(*name, target, pkgver - target, return -1);
+ if(name_hash) {
+ *name_hash = _alpm_hash_sdbm(*name);
+ }
}
- STRNDUP(pkg->name, target, version - target, return -1);
- pkg->name_hash = _alpm_hash_sdbm(pkg->name);
return 0;
}
diff --git a/lib/libalpm/util.h b/lib/libalpm/util.h
index 778e20f..c68b07b 100644
--- a/lib/libalpm/util.h
+++ b/lib/libalpm/util.h
@@ -109,7 +109,8 @@ const char *_alpm_filecache_setup(pmhandle_t *handle);
int _alpm_lstat(const char *path, struct stat *buf);
int _alpm_test_md5sum(const char *filepath, const char *md5sum);
int _alpm_archive_fgets(struct archive *a, struct archive_read_buffer *b);
-int _alpm_splitname(const char *target, pmpkg_t *pkg);
+int _alpm_splitname(const char *target, char **name, char **version,
+ unsigned long *name_hash);
unsigned long _alpm_hash_sdbm(const char *str);
long _alpm_parsedate(const char *line);
--
1.7.5.4
More information about the pacman-dev
mailing list