Difference between revisions of "CacheClean"

From ArchWiki
Jump to: navigation, search
m (add original author)
(wikify some external links, use https for archlinux.org)
(17 intermediate revisions by 7 users not shown)
Line 1: Line 1:
[[Category:Scripts (English)]]
+
[[Category:Scripts]]
[[Category:Package management (English)]]
+
[[Category:Package management]]
 +
[[zh-CN:CacheClean]]
 +
Cacheclean is python script to clean {{ic|/var/cache/pacman/pkg}} allowing users to specify how many package versions should be retained.  In function, it is similar to {{ic|pacman -Sc}} except users select how many old versions to keep.  Another difference is unlike, {{ic|pacman -Sc}} cacheclean does not discriminate against packages that are not currently installed as it inspects {{ic|/var/cache/pacman/pkg}}.
  
This script has been included in the AUR at [[http://aur.archlinux.org/packages.php?ID=37572 this page]].
+
== Installation ==
 +
[[AUR]] package: {{AUR|cacheclean}}
  
It was originally written by "alterkacker" - see his forum thread: [[http://bbs.archlinux.org/viewtopic.php?id=9104||A utility for cleaning /var/cache/pacman/pkg]]
+
== Usage ==
  
 +
{{bc|cacheclean {-p} {-v} <# of copies to keep>
 +
# of copies to keep - (required) how many generations of each package to keep
 +
-p - (optional) preview what would be deleted; forces verbose (-v) mode.
 +
-v - (optional) show deleted packages.}}
  
 +
== Code ==
 +
For those wishing to contribute, the script's source code is hosted in [https://github.com/graysky2/cacheclean/tree/master/cacheclean this github repository].  To simplify updates, please fork and send a pull request with updates/changes.
  
<pre>
+
== See also ==
#!/usr/bin/env python
+
*Author's original forum thread: [https://bbs.archlinux.org/viewtopic.php?id=9104 A utility for cleaning /var/cache/pacman/pkg]
"""cache_clean - a simple python script to clean up the /var/cache/pacman/pkg directory.
 
More versatile than 'pacman -Sc' in that you can select how many old versions
 
to keep.
 
Usage: cache_clean {-p} {-v} <# of copies to keep>
 
  # of copies to keep - (required) how many generations of each package to keep
 
  -p - (optional) preview what would be deleted; forces verbose (-v) mode.
 
  -v - (optional) show deleted packages."""
 
# Note that the determination of package age is done by simply looking at the date-time
 
# modified stamp on the file. There is just enough variation in the way package version
 
# is done that I thought this would be simpler & just as good.
 
# Also note that you must be root to run this script.
 
 
 
import getopt
 
import os
 
import re
 
import sys
 
 
 
# helper function to get tuple of (file dtm, file name, file sz)
 
def fn_stats(fn):
 
  s = os.stat(fn)
 
  return (s.st_mtime, fn, s.st_size)
 
 
 
# cleanup does the actual pkg file deletion
 
def cleanup(run_list):
 
  # strictly speaking only the first two of these globals need to be listed.
 
  global n_deleted, bytes_deleted, opt_verbose, opt_preview, n_to_keep
 
  # return if the run_list is too short
 
  #print run_list
 
  #return
 
  if len(run_list) <= n_to_keep: return
 
  # Build list of tuples (date-time file modified, file name, file size)
 
  dtm_list = [fn_stats(tfn) for tfn in run_list]
 
  # Sort the list by date-time
 
  dtm_list.sort()
 
  # Build list of the filenames to delete (all but last n_to_keep).
 
  # <showing_off>
 
  #kill_list = [tfn[1] for tfn in dtm_list[:-n_to_keep]]
 
  #bytes_deleted = sum(x[2] for x in dtm_list[:-n_to_keep])
 
  # </showing_off>
 
  kill_list = []
 
  for x in dtm_list[:-n_to_keep]:
 
    kill_list.append(x[1])
 
    bytes_deleted += x[2]
 
  if opt_verbose: print kill_list
 
  n_deleted += len(kill_list)
 
  # and finally delete (if not in preview mode)
 
  if not opt_preview:
 
    for dfn in kill_list:
 
      os.unlink(dfn)
 
 
 
######################################################################
 
# mainline processing
 
 
 
# process command line options
 
try:
 
  opts, pargs = getopt.getopt(sys.argv[1:], 'vp')
 
  opt_dict = dict(opts)
 
  opt_preview = '-p' in opt_dict
 
  opt_verbose = '-v' in opt_dict
 
  if opt_preview: opt_verbose = True
 
  if len(pargs) == 1:
 
    n_to_keep = pargs[0]
 
  else:
 
    raise getopt.GetoptError("missing required argument.")
 
  try:
 
    n_to_keep = int(n_to_keep)
 
    if n_to_keep <= 0: raise ValueError
 
  except ValueError, e:
 
    raise getopt.GetoptError("# of copies to keep must be numeric > 0!")
 
except getopt.GetoptError, msg:
 
  print "Error:",msg,"\n",__doc__
 
  sys.exit(1)
 
 
 
# change to the pkg directory & get a sorted list of its contents
 
os.chdir('/var/cache/pacman/pkg')
 
pkg_fns = os.listdir('.')
 
pkg_fns.sort()
 
 
 
# Pattern to use to extract the package name from the tar file name:
 
# for pkg e.g. 'gnome-common-2.8.0-1-i686.pkg.tar.gz' group(1) is 'gnome-common'.
 
bpat = re.compile(r'^(.+?)(?:-\d.*?)(-i686|-x86_64|-any)?\.pkg\.tar\.(?:gz|xz)$')
 
 
 
n_deleted = 0
 
bytes_deleted = 0
 
pkg_base_nm = ''
 
# now look for "runs" of the same package name differing only in version info.
 
for run_end in range(len(pkg_fns)):
 
  fn = pkg_fns[run_end]
 
 
 
  # make sure we skip directories
 
  if not os.path.isdir(fn):
 
    mo = bpat.match(fn) # test for a match of the package name pattern
 
    if mo:
 
      tbase = mo.group(1) # gets the 'base' package name
 
      # include the architecture in the name if it's present
 
      if mo.lastindex > 1:
 
        tbase += mo.group(2)
 
      #print 'tbase: ',tbase,'  ',mo.lastindex
 
      # is it a new base name, i.e. the start of a new run?
 
      if tbase != pkg_base_nm: # if so then process the prior run
 
        if pkg_base_nm != '':
 
          cleanup(pkg_fns[run_start:run_end])
 
        pkg_base_nm = tbase # & setup for the new run
 
        run_start = run_end
 
    else:
 
      print >>sys.stderr, "File '"+fn+"' doesn't match package pattern!"
 
  else:
 
    print >>sys.stdout, "skipping directory '"+fn+"'!"
 
 
 
# catch the final run of the list
 
run_end += 1
 
cleanup(pkg_fns[run_start:run_end])
 
 
 
if opt_verbose:
 
  if opt_preview:
 
    print "Preview mode (no files deleted):",
 
  print n_deleted,"files deleted,",bytes_deleted/1000,"kbytes."
 
</pre>
 

Revision as of 15:52, 4 December 2012

zh-CN:CacheClean Cacheclean is python script to clean /var/cache/pacman/pkg allowing users to specify how many package versions should be retained. In function, it is similar to pacman -Sc except users select how many old versions to keep. Another difference is unlike, pacman -Sc cacheclean does not discriminate against packages that are not currently installed as it inspects /var/cache/pacman/pkg.

Installation

AUR package: cachecleanAUR

Usage

cacheclean {-p} {-v} <# of copies to keep>
	# of copies to keep - (required) how many generations of each package to keep
	-p - (optional) preview what would be deleted; forces verbose (-v) mode.
	-v - (optional) show deleted packages.

Code

For those wishing to contribute, the script's source code is hosted in this github repository. To simplify updates, please fork and send a pull request with updates/changes.

See also