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
+ 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
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.