From ac157895f10abf96ad7d1d8170e57a3cb4ee3df8 Mon Sep 17 00:00:00 2001
From: Dan McGee <dan@archlinux.org>
Date: Tue, 17 Dec 2013 08:26:47 -0600
Subject: reporead: implement delayed parsing of files data

This gives us some large memory savings in python due to the internal
storage of Unicode strings vs. byte strings, as well as saving us
processing time up front for filelist data we are never going to have to
actually use.

Signed-off-by: Dan McGee <dan@archlinux.org>
---
 devel/management/commands/reporead.py | 58 +++++++++++++++++++++--------------
 1 file changed, 35 insertions(+), 23 deletions(-)

(limited to 'devel')

diff --git a/devel/management/commands/reporead.py b/devel/management/commands/reporead.py
index 248eb62c..3244d593 100644
--- a/devel/management/commands/reporead.py
+++ b/devel/management/commands/reporead.py
@@ -82,8 +82,7 @@ class RepoPackage(object):
             'md5sum', 'sha256sum', 'url', 'packager' )
     number = ( 'csize', 'isize' )
     collections = ( 'depends', 'optdepends', 'makedepends', 'checkdepends',
-            'conflicts', 'provides', 'replaces', 'groups', 'license',
-            'files' )
+            'conflicts', 'provides', 'replaces', 'groups', 'license')
 
     def __init__(self, repo):
         self.repo = repo
@@ -98,7 +97,6 @@ def __init__(self, repo):
             setattr(self, k, ())
         self.builddate = None
         self.files = None
-        self.has_files = False
 
     def populate(self, values):
         for k, v in values.iteritems():
@@ -120,13 +118,21 @@ def populate(self, values):
                     logger.warning(
                             'Package %s had unparsable build date %s',
                             self.name, v[0])
-            elif k == 'files':
-                self.files = tuple(v)
-                self.has_files = True
             else:
                 # anything left in collections
                 setattr(self, k, tuple(v))
 
+    @property
+    def files_list(self):
+        data_file = io.TextIOWrapper(io.BytesIO(self.files), encoding='UTF-8')
+        try:
+            info = parse_info(data_file)
+        except UnicodeDecodeError:
+            logger.warn("Could not correctly decode files list for %s",
+                    self.name)
+            return None
+        return info['files']
+
     @property
     def full_version(self):
         '''Very similar to the main.models.Package method.'''
@@ -291,15 +297,18 @@ def populate_files(dbpkg, repopkg, force=False):
             return
 
     # only delete files if we are reading a DB that contains them
-    if repopkg.has_files:
+    if repopkg.files:
+        files = repopkg.files_list
+        # we had files data, but it couldn't be parsed, so skip
+        if not files:
+            return
         delete_pkg_files(dbpkg)
         logger.info("adding %d files for package %s",
-                len(repopkg.files), dbpkg.pkgname)
+                len(files), dbpkg.pkgname)
         pkg_files = []
         # sort in normal alpha-order that pacman uses, rather than makepkg's
         # default breadth-first, directory-first ordering
-        files = sorted(repopkg.files)
-        for f in files:
+        for f in sorted(files):
             if '/' in f:
                 dirname, filename = f.rsplit('/', 1)
                 dirname += '/'
@@ -507,24 +516,27 @@ def parse_repo(repopath):
 
     repodb = tarfile.open(repopath, "r")
     logger.debug("Starting package parsing")
-    dbfiles = ('desc', 'depends', 'files')
     newpkg = lambda: RepoPackage(reponame)
     pkgs = defaultdict(newpkg)
     for tarinfo in repodb.getmembers():
         if tarinfo.isreg():
             pkgid, fname = os.path.split(tarinfo.name)
-            if fname not in dbfiles:
-                continue
-            data_file = repodb.extractfile(tarinfo)
-            data_file = io.TextIOWrapper(io.BytesIO(data_file.read()),
-                    encoding='UTF-8')
-            try:
-                pkgs[pkgid].populate(parse_info(data_file))
-            except UnicodeDecodeError:
-                logger.warn("Could not correctly decode %s, skipping file",
-                        tarinfo.name)
-            data_file.close()
-            del data_file
+            if fname == 'files':
+                # don't parse yet for speed and memory consumption reasons
+                files_data = repodb.extractfile(tarinfo)
+                pkgs[pkgid].files = files_data.read()
+                del files_data
+            elif fname in ('desc', 'depends'):
+                data_file = repodb.extractfile(tarinfo)
+                data_file = io.TextIOWrapper(io.BytesIO(data_file.read()),
+                        encoding='UTF-8')
+                try:
+                    pkgs[pkgid].populate(parse_info(data_file))
+                except UnicodeDecodeError:
+                    logger.warn("Could not correctly decode %s, skipping file",
+                            tarinfo.name)
+                data_file.close()
+                del data_file
 
             logger.debug("Done parsing file %s/%s", pkgid, fname)
 
-- 
cgit v1.2.3-54-g00ecf