On Fri, Jul 07, 2006 at 10:26:13AM -0700, Scott Wang wrote: > Hi Paul, > > Even, currently, there is no any zero size data file > in structure folder in my regression code coverage run > (lots of suites), I still got lots of messages like > below: > ----------- > Corrupted storable file (binary v2.7) at > ../../lib/Storable.pm (autosplit into > ../../lib/auto/Storable/_retrieve.al) line 331, at > /xxxxxxxx/Devel/Cover/DB/Structure.pm line 269 > END failed--call queue aborted. > ------------- > Seems some data files got corrupted somehow. > > The process killing (kill parent process before > killing child process and make chid process be > terminated abnormally) is one possibility, I also > noticed that useing system or other folk method to > execute command could also cause this problem. > > Any clue about above issue "Corrupted storable file > (binary v2.7) at ..." and does anybody also experience > this issue and know how to deal with it? > > Also, does anybody have similar experience on killing > process and folking process cause "Magic number ..." > issue and "orrupted storable file ..." issue?
In lieu of discussing either licences or alligators, let me try to answer this. Killing the process sounds like the most likely cause of this problem. If you manage to kill the process while it is writing out a storable file the file will very probably be corrupted. How are you killing the process? Are you sending it SIGKILL (9) for example? Maybe you could send it something a little nicer which might allow to process to clean up. Devel::Cover does its work in the very last END block. You really need to let it run to completion. -- Paul Johnson - [EMAIL PROTECTED] http://www.pjcj.net
