OnTriggerAvail not firing...

[Paul Breneman [TPX]]

Mike,

> Mark tried it with APRO 4 and it's now firing the OnTriggerAvail.

It is good that he found something that works, but I'm still wondering
what the problem is using 3.06. As far as I know that version should be
very reliable. I'm still wondering if this computer has two processors?
If not how can 3.06 be a problem with USB ports or XP Pro?

--
Regards,
Paul Breneman [TPX]
http://www.BrenemanLabs.com

Those of us with [TPX] after our names are not TurboPower employees but
a group of folks that have volunteered to assist TurboPower and help
fellow users who have questions about the operation of the products.

[Mike Welch [TurboPower Software]]

Mark,
There have been a few changes to our dispatcher that might affect this.
Can I email some test projects to you? (I assume that the pdq.bounce.net
address is invalid)

--
Mike Welch (TurboPower Software)
God Bless America

<za...@pdq.bounce.net> wrote in message
news:MPG.1879523e4...@news.turbopower.com...
> Hi,
>
> > Is this a single processor machine?
>
> > Can you make a simple test application that demonstrates the problem?
>
> No. ;) It is only ports that have the Belkin to USB converter. Regular
> serial ports work just fine. Belkin installs virtual serial ports.
>
> I can do any kind of testing to trace down exactly what is happening.
> From the log Apro is doing a ReadCom and getting the bytes. It just is
> not firing the OnTriggerAvail. At least it is not always firing the
> trigger.
>
> > I've got lots of programs in 24x7 operation using Delphi 5.01 and Apro
> > 3.05 so that should be similar to your situation with 3.06.
>
> On XP pro?
>
>
> Cheers,
>
>
> Mark
>
>
> In article <WuHHaX6r...@tpsmail01.turbopower.net>,
> Paul.Brenema...@TPX.TurboPower.com says...
> > Mark,
> >
> > > It works most of the time. Sometimes the trigger does not fire as the
> > > log shows. When I time out I can check the inBuffUsed and it will show
> > > the bytes for which I have not received a trigger.
> > >
> > > Other data. Delphi 5.01 and XP.
> >
> > I don't remember seeing problems with the OnTriggerAvail not firing
> > except for one instance where someone had a data trigger and in the
> > handler for that they were reading data from the buffer (which caused
> > the OnTriggerAvail to not fire sometimes).
> >
> > I've got lots of programs in 24x7 operation using Delphi 5.01 and Apro
> > 3.05 so that should be similar to your situation with 3.06.
> >
> > I don't have too many ideas at this point. Is this a single processor
> > machine? Can you make a simple test application that demonstrates the
> > problem?
> >
> >

[Mike Welch [TurboPower Software]]

Mark tried it with APRO 4 and it's now firing the OnTriggerAvail.

--

Mike Welch (TurboPower Software)
God Bless America

"Paul Breneman [TPX]" <Paul.Brenema...@TPX.TurboPower.com> wrote
in message news:oFHlNzJs...@tpsmail01.turbopower.net...
> Mark,

>
> >>Is this a single processor machine?
>

> I didn't read any response to that question. This PC doesn't have dual
> processors does it?

>
> >>Can you make a simple test application that demonstrates the problem?
> >
> >
> > No. ;) It is only ports that have the Belkin to USB converter. Regular
> > serial ports work just fine. Belkin installs virtual serial ports.
> >
> > I can do any kind of testing to trace down exactly what is happening.
> > From the log Apro is doing a ReadCom and getting the bytes. It just is
> > not firing the OnTriggerAvail. At least it is not always firing the
> > trigger.
>

> Since the data shows up in a ReadCom I don't see how the Belkin USB
> converter would make any difference. Maybe Mike will have some further
> ideas.

>
> >>I've got lots of programs in 24x7 operation using Delphi 5.01 and Apro
> >>3.05 so that should be similar to your situation with 3.06.
> >
> >
> > On XP pro?
>

> No, I recommend Win2K right now to my clients. Again, since the data is
> getting into AsyncPro OK I don't understand what difference the
> operating system would make (except for a dual processor machine).

[Vic Feria]

No problem Paul, I used both methods. When I expect what type or size of
input I get with all the conditions (startstring/endstring) then I use it
with the apdDataPacket. I used to just get the incoming chars directly from
apdComport when I wrote a paging program prior to apdTapPager.

I just want to know if one method is better than the other.

Thanks,
Victor

Paul Breneman [TPX]" <Paul.Brenema...@TPX.TurboPower.com> wrote in

message news:V#k#SX6rCH...@tpsmail01.turbopower.net...
> Victor,
>
> > Would apdDataPacket in combination with apdComport work more reliably?
>
> I have many programs in 24x7 operation using Delphi 5.01 and Apro 3.05
> and OnTriggerAvail and the data packets are both extremely reliable.
> What problems are you experiencing?

[Mike Welch [TurboPower Software]]

Mark,
It's on the way. In general, the USB converter should be completely
seamless/transparent. The only problem that I know of with USB serial
converters is with how it's powered. A self-powered device may not provide
all of the line signals (RTS, etc). That may be a contributing factor,
especially if we're using HW flow control. Try powering it externally.
A lot of the newer USB virtual ports are also not sending the
notification to the application that chars are ready, instead they are
relying on constant polling of the port. The drivers for the Motorola SM56
chipset were the first to do this that we found, but we changed to a polling
design for APRO 3.00.

--
Mike Welch (TurboPower Software)
God Bless America

<za...@pdq.bounce.net> wrote in message
news:MPG.187a0dd3...@news.turbopower.com...
> Hi Mike,

>
> > Can I email some test projects to you?
>

> Yes.

>
> > I assume that the pdq.bounce.net address is invalid
>

> Good eye.
>
> Remove the bounce.
>
> zayin at pdq.net
>
> Cheers,
>
> Mark

[Paul Breneman [TPX]]

Mark,

> I have a Belkin F5U103 USB to serial converter.
>
> Portmon shows that the bytes are arriving.
>
> The APRO log shows that the bytes are arriving. Sometimes the
> OnTriggerAvail does not fire.

If you have an OnTriggerAvail event defined then you should not use
GetChar, GetBlock, FlushInBuffer, or any similar functions anywhere
else. Are you "peeking" at the incoming data in any other place?

You need to read "Count" number of characters in your OnTriggerAvail
using GetChar or GetBlock. You might want to verify that.

Those are the only two things off the top of my head...

[Paul Breneman [TPX]]

Mark,

> It works most of the time. Sometimes the trigger does not fire as the
> log shows. When I time out I can check the inBuffUsed and it will show
> the bytes for which I have not received a trigger.
>
> Other data. Delphi 5.01 and XP.

I don't remember seeing problems with the OnTriggerAvail not firing
except for one instance where someone had a data trigger and in the
handler for that they were reading data from the buffer (which caused

the OnTriggerAvail to not fire sometimes).

I've got lots of programs in 24x7 operation using Delphi 5.01 and Apro

3.05 so that should be similar to your situation with 3.06.

I don't have too many ideas at this point. Is this a single processor

machine? Can you make a simple test application that demonstrates the
problem?

--

[Paul Breneman [TPX]]

Victor,

> Would apdDataPacket in combination with apdComport work more reliably?

I have many programs in 24x7 operation using Delphi 5.01 and Apro 3.05

and OnTriggerAvail and the data packets are both extremely reliable.
What problems are you experiencing?

--

[Vic Feria]

Paul,

Would apdDataPacket in combination with apdComport work more reliably?

--Victor

<za...@pdq.bounce.net> wrote in message
news:MPG.1878c6a45...@news.turbopower.com...
> Hi Paul,

>
> > Are you "peeking" at the incoming data in any other place?
>

> No.

>
> > You need to read "Count" number of characters in your OnTriggerAvail
> > using GetChar or GetBlock. You might want to verify that.
>

> That is exactly what I am doing via GetBlock.

>
> It works most of the time. Sometimes the trigger does not fire as the
> log shows. When I time out I can check the inBuffUsed and it will show
> the bytes for which I have not received a trigger.
>
> Other data. Delphi 5.01 and XP.
>

> Cheers,
>
> Mark
>
>
>
> In article <usJIK9vr...@tpsmail01.turbopower.net>,
> Paul.Brenema...@TPX.TurboPower.com says...

[P McNab]

Hi Paul, Mike

I have used A USB adaptor (Belden I think) successfully with Win2K and
V3.05/6 at 115k baud and rely totally on the OnTriggerAval for
reception. It has never given me any problems. The port was opended from
the VCL thread. Having accumulated a complete message I then pass it to
a thread driven queue for dispatch within the application, so there is
very little time spent in the OnTrigger event handler.

However, when we need to toggle RTS or DTR there appears to be a
significantly slower response than running via a 16C550 comm port.

Regards
Peter McNab

"Paul Breneman [TPX]" wrote:
>
> Mike,
>
> > I'm not sure which one of the changes (or combinations of changes) are the
> > critical ones. We've played a bit with the thread synchronization, the
> > polling of the port, and a few other things. Without being able to explain
> > it, I'm just glad it's working<g>
>
> Yes, I'm glad its working too. It is just that I have a lot of clients
> using AysncPro 3.05 programs. Some of those folks have used USB a
> little, and although I've been recommending Win2K it is only a matter of
> time till WinXP gets used on more of these systems. Call me paranoid
> but if there is a problem with 3.05/USB/XP I'd like to know it before
> one of my customers calls me. If you or Mark come up with an
> explanation later please remember to write this paranoid person...

[Mike Welch [TurboPower Software]]

I'm not sure which one of the changes (or combinations of changes) are the
critical ones. We've played a bit with the thread synchronization, the
polling of the port, and a few other things. Without being able to explain
it, I'm just glad it's working<g>

--

Mike Welch (TurboPower Software)
God Bless America

<za...@pdq.bounce.net> wrote in message
news:MPG.187b9d185...@news.turbopower.com...
> Hi,

>
> > I'm still wondering if this computer has two processors?
>

> Only one processor.

>
> > If not how can 3.06 be a problem with USB ports or XP Pro?
>

> Good question. ;)
>
> Cheers,
>
>
> Mark
>
>
> In article <UDcK0UQs...@tpsmail01.turbopower.net>,
> Paul.Brenema...@TPX.TurboPower.com says...

[Paul Breneman [TPX]]

> APRO v3.05
> Compiler : Delphi 5
> Operating System : Windows NT 4.0 Service Pack 6
> Device: Sportster 28800-33600 External #2

I would just like to point out to anyone looking at my log file header
that I believe good stable tools have their place... <G>

Of course I may eventually get run over by the C# and .NET wagon, but
right now development life is good...

[Mike Welch [TurboPower Software]]

Mark,
Not yet, but I'll log into that account shortly.

--
Mike Welch (TurboPower Software)
God Bless America

<za...@pdq.bounce.net> wrote in message
news:MPG.187dff1cd...@news.turbopower.com...
> Hi Mike,
>
> Have you seen the emails I sent to support?
>
>
> Cheers,
>
>
> Mark
>
> In article <iXo6FPns...@tpsmail01.turbopower.net>,
> please@keep_it_in_the_newsgroup.thanks says...

[Paul Breneman [TPX]]

Mike,

I just did some tests and found out something else. This problem only
happens when I have TAPI involved, not when I have a direct serial
connection to the device. Now I'm real confused as TAPI should not
matter as the connection was made long before this particular command is
done. But it is completely repeatable (which is good).

[Paul Breneman [TPX]]

Hi Peter,

> I have used A USB adaptor (Belden I think) successfully with Win2K and
> V3.05/6 at 115k baud and rely totally on the OnTriggerAval for
> reception. It has never given me any problems. The port was opended from
> the VCL thread. Having accumulated a complete message I then pass it to
> a thread driven queue for dispatch within the application, so there is
> very little time spent in the OnTrigger event handler.
>
> However, when we need to toggle RTS or DTR there appears to be a
> significantly slower response than running via a 16C550 comm port.

Thanks for jumping in with your experience. Maybe the significant thing
here is that you are using Win2K and Mark is using WinXP? I don't see
how the operating system would cause the OnTriggerAvail to be unreliable
so that is why I'm still searching for answers on this. But it is good
to hear that things are reliable for you too. That is my experience and
I hope that WinXP doesn't change that.

It is also good to know to expect larger delays in toggling the hardware
lines. Thanks for mentioning that.

[Paul Breneman [TPX]]

Mike,

> I'm not sure which one of the changes (or combinations of changes) are the
> critical ones. We've played a bit with the thread synchronization, the
> polling of the port, and a few other things. Without being able to explain
> it, I'm just glad it's working<g>

Yes, I'm glad its working too. It is just that I have a lot of clients

using AysncPro 3.05 programs. Some of those folks have used USB a
little, and although I've been recommending Win2K it is only a matter of
time till WinXP gets used on more of these systems. Call me paranoid
but if there is a problem with 3.05/USB/XP I'd like to know it before
one of my customers calls me. If you or Mark come up with an
explanation later please remember to write this paranoid person...

--

[Mike Welch [TurboPower Software]]

I did a quick scan of a few USB adapter drivers, and they're using the same
ones for W2K and XP. For all we know, the original problem could be
completely unrelated to the OS. Perhaps the afflicted system was running a
high-priority service or something like that, or they updated the USB
drivers, or any number of off-the-wall possibilites. Since it was happening
on a customer's machine, remote from the developer, who knows what they did
on their end to try to fix the problem (if they did anything). Reminds me
of something that happened a few years ago. They couldn't send a fax, so we
went through all kinds of debugging but couldn't find the problem. Out of
the blue, it started working, but we didn't find out until a few months
later that the customer got a new computer (the old computer had a bad PCI
bus)

--
Mike Welch (TurboPower Software)
God Bless America

"Paul Breneman [TPX]" <Paul.Brenema...@TPX.TurboPower.com> wrote
in message news:Ye6CtGss...@tpsmail01.turbopower.net...