>Hoping someone on this list knows the answer to this...
>
>We have a software ap running on a notebook computer with Windows2000
>system. There is a pcmcia card with a serial port stuck in to interface to
>external hardware. The customer has it setup to startup automatically into
>the application program.
>
>The problem is that about 10% of the time, the program starts up and either
>hasn't found the plugin card yet or it somehow gets configured wrong. Then
>the program runs but won't allow comms. The user must exit the program and
>restart it.
>
>It appears to be simply due to how windows starts up and how/when/where it
>recognizes the plugin card. If this is so, does anyone know how to control
>this behavior so the card is found 100% of the time?
>
>Thanks,
>Tom M.
>
>--
>http://www.piclist.com hint: To leave the PICList
>.....piclist-unsubscribe-requestKILLspam@spam@mitvma.mit.edu
>
>
>
>
>
>
>

>Tom Messenger <KILLspamkristKILLspamTHEGRID.NET> wrote:
>
>
>>We have a software ap running on a notebook computer with Windows2000
>>system. There is a pcmcia card with a serial port stuck in to interface to
>>external hardware. The customer has it setup to startup automatically into
>>the application program.
>>
>>The problem is that about 10% of the time, the program starts up and either
>>hasn't found the plugin card yet or it somehow gets configured wrong. Then
>>the program runs but won't allow comms. The user must exit the program and
>>restart it.
>>
>>It appears to be simply due to how windows starts up and how/when/where it
>>recognizes the plugin card. If this is so, does anyone know how to control
>>this behavior so the card is found 100% of the time?
>>
>>
>
>Is there another program that runs to configure the PCMCIA serial card?
>If so, then there's probably a simple fix, if you're comfortable with
>editing the registry.
>
>Look in the registry for the key
>  HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Run
>
>See if there's an entry for the PCMCIA serial card. (I assume you'll also
>find an entry for the application program here.)
>
>If so, move the entry for the PCMCIA card to the following key:
>  HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\RunOnce
>
>The difference is that the programs in ...\RunOnce are guaranteed to run
>to completion before all the programs in ...\Run start. The latter programs
>multitask normally and are not guaranteed to run in any particular sequence.
>
>This may cure your problem.
>
>-- Dave Tweed
>
>--
>http://www.piclist.com hint: To leave the PICList
>RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu
>
>
>
>
>
>
>

>If it is being started as a service, I think there is a way to determine the
>order of starting, but have never gone digging into this. It may be
>necessary to have a time delay or retry built into the app as others
>suggest, as it sounds like you have a race condition occurring between the
>app and the drivers it relies on.
>
>Another solution may be to have the app start at a real low priority level,
>and retry on the dependencies, and then once the link is made push its own
>priority level up to something suitable - just a thought, again not
>something I have tried.

>
> > >If it is being started as a service, I think there is a way to determine
> the
> > >order of starting, but have never gone digging into this. It may be
> > >necessary to have a time delay or retry built into the app as others
> > >suggest, as it sounds like you have a race condition occurring between
> the
> > >app and the drivers it relies on.
>
> The service manager allows you to mark a service as dependent on another
> service IIRC.

>
> "M. Adam Davis" <EraseMEadampicUBASICS.COM> wrote:
> > According to
> > support.microsoft.com/default.aspx?scid=kb;EN-US;q137367
> > runonce entries are run once then deleted from the registry.  They are
> > used for setup programs, etc.
>
> Yeah, that occurred to me right after I sent the message.
>
> OK, forget what I said. I don't see a simple way to force to commands to
> run sequentially on every boot, without writing a wrapper as someone else
> suggested.
>
> -- Dave Tweed
>
>