VMS.notes

This  file  contains   special   notes   concerning   VMS  platforms.  All
information on how to fix the issues described below  is welcome, provided
it is complete,  clean, and  (most of all)  working.  We have no access to
VMS, hence we cannot test anything ourselves!

1. The  VMS  port  of xautolock is not complete.  Some known functionality
   problems are:

    + Multi-headed displays may not be supported under VMS.

    + The -notifier option may either misbehave or even not work at
      all on VMS.

    + The -unlocknow option only works on some VMS versions 
      (if it compiles, it should also work on the same machine).
    
    + The -killpid option might fail to work.

2. Due to some  VMS quirks,  the inner workings  of which  are  a complete
   mystery to us,  there seems  to be something  not quite right  with the
   code  used to detect  the presence  of the locker  during the first few 
   seconds  of its existence.   This sometimes manifests itself as follows
   (quoting from the original problem report):

     "One  person reported that  using the corners options,  he  had a
     problem.  Apparently, if we guessed right,  if it takes more than
     5 seconds for the locker to start, the timer will time out again,
     and fire off a second locker screen."

   While this cannot ever occur on UNIX,  even if the locker takes several
   hours to start, it does seem to be possible under VMS. As a workaround,
   the code contains  a conditional sleep() call  right after starting the
   locker,  which can be compiled in be #define-ing SLOW_VMS.  By default,
   this is not used.

   Obviously,  defining  SLOW_VMS  will influence  the timing behaviour of
   xautolock in ways which are incompatible with what is stated in the man
   page.

Note  for users of previous versions:  xautolock is no longer sensitive to
the   sometimes  serious  mistimings   caused  by  the  braindead  sleep()
implementation of VMS.  It may therefore  be needed  to modify your  -time
and -cornerdelay settings.