Created attachment 357628 [details]
Proposed patch

+++ This bug was initially created as a clone of Bug #469702 +++

Escalated to Bugzilla from IssueTracker

--- Additional comment from mpoole on 2008-11-03 11:13:31 EDT ---

There is a bug with the print code in that is simply reads the whole file sequentially to meet the man page of printing in UID order. Since the output does not actually include the UID it is a little redundant stating or assuming UID order which therefore gives an optimal way of reworking the code.

* How reproducible:

Always.

* Steps to Reproduce:

1) Make sure nfsnobody has a very large uid:

  # grep nfsnobody /etc/passwd
  nfsnobody:x:4294967294:4294967294:Anonymous

2) Initialize the faillog file

  # faillog -m 3

3) Check the size of the faillog file (huge sparse file):

  # ls -l /var/log/faillog
  -rw-r--r-- 1 root root 137438953440 Aug 17 11:43 /var/log/faillog

4) Check for failures

  # faillog -a

Actual result

  faillog reads sequentially the huge sparse file and takes a very long time to complete.

Expected results

  faillog picks the right records in the sparse file and completes instantly.

Additional information

Attaching the patch from Martin Poole in BZ#469702 ported to shadow 4.0.17 for el5.

Note: the problem is also present in F11.