getIdleSeconds suggestion

Proposal: Add ability to identify how long it has been since the last input
was received from the user.

Justification: In each of the major commercial applications I have worked on
in the past decade there has eventually arisen a request to do some action
if the user has been idle for some period. Sometimes this is related to
security--log out of the database if the user has been away from the
keyboard for X minutes. Sometimes this is related to background
tasks--during an idle period do some file maintenance. In the case of
security, the approach of locking the entire workstation (with the OS screen
saver and password) is not attractive because the champion or our
application wants to have additional security and not be at the mercy of the
local workstation setup. Also, the OS approach would require a separate OS
account for each user and in a point-of-sale application they want to be
able to leave the OS login as a generic account and have each clerk provide
an ID to the application and have the application recognize if the user has
gone away. One could keep track of user events internally, but that is a lot
of extra overhead. Fortunately, Windows has an API for getting information
about the last user input.

Public API: UserLibrary>>getIdleSeconds obtains the millisecondClockValue
associated with the last user input and compares it to the current time,
returning a number of seconds since the last user event. The method handles
the clock wrap-around that occurs when the 32-bit counter overflows.

Implementation:

UserLibrary>>getIdleSeconds
"
(Delay forSeconds: 5) wait. UserLibrary default getIdleSeconds.
"
 | lastEvent current wrap |
 lastEvent := self getLastInputInfoTime.
 current := Time millisecondClockValue.
 lastEvent <= current ifTrue: [^current - lastEvent // 1000].
 wrap := 2 raisedTo: 32.
 ^current + wrap - lastEvent // 1000 \\ wrap.

UserLibrary>>getLastInputInfoTime

 | buffer flag |
 buffer := ByteArray new: 8.
 buffer
  dwordAtOffset: 0
  put: buffer size.
 flag := self getLastInputInfo: buffer.
 flag ifFalse: [self error: 'Unable to get last input time'].
 ^buffer dwordAtOffset: 4.

UserLibrary>>getLastInputInfo: anExternalBuffer
"
 BOOL GetLastInputInfo(
  PLASTINPUTINFO plii <>   // last input event
 );
"
 <stdcall: bool GetLastInputInfo lpvoid>
 ^self invalidCall

"James Foster" <[hidden email]> wrote in message
news:[hidden email]...
> Proposal: Add ability to identify how long it has been since the last
> input was received from the user.

James, I think this is quite a reasonable proposal, and I have no objection
to it going into the base system. I'd like to see the abstraction
implemented in the InputState class rather than in UserLibrary. In this case
a trivial thing to address.

It clearly must have taken you some effort to formulate your proposal, and
thanks for that but it really isn't necessary to be quite so formal :-). A
few words of why you think something ought to be in the base rather than in
your own extension package, the code, and (critically) some tests will do
the trick. Mostly in justification we would be looking for either (a) a
common need and/or (b) great potential for a clash with extensions built by
others.

Regards

Blair

Blair,
Thanks for your response, and for your patience with my suggestions. I
apologize for the "tone" of some of my comments. For what it's worth, I'm
really enjoying the new tools. The IdeaSpace is taking some retraining on
habits, but I can see it will be useful. And I really like being rid of the
RichTextEdit widget for code. I'm still getting used to the text completion,
and haven't (yet) discovered that I can't live without it. On the other
hand, it certainly does give you parity with the best of the competing
languages' code editors (for those who like to check-off a feature list).
Have a good weekend, and thanks again for a wonderful product.
James
"Blair McGlashan" <[hidden email]> wrote in message
news:[hidden email]...