AspeQt: Yet another SIO2PC/APE clone

[+DrVenkman]

Since it also happens when AspeQt is more or less idling, I assume it's a polling or sleep state misconfiguration or something like that...

 

Actually it's something to do with opening the serial port. If you go into the AspeQt setting and pick a non-existent tty file, for instance, so that AspeQt cannot open the port for I/O, CPU usage drops down to nil. Once you specify a valid port and AspeQt opens, CPU goes through the roof. If that is NOT the behavior Win/Linux people are seeing, then AspeQt's OS X build has something specific wrong with it.

 

Anyway, I PM'd 'meliss' to see if he's interested in taking a look at the AspeQt Mac build; he's still active at AA so maybe he'll respond. Comparing how his code handles the I/O routines and opens the serial device with how AspeQt does it would be instructive.

Edited July 20, 2014 by DrVenkman

[Tempest]

I assume I just need to click on the AspeQt execuatable after downloading right? When I do that nothing seems to happen. I have an older version on my box (0.83) that appears to work, but it has different files and folders (no $boot folders and it has a bunch of other files that the new version doesn't).

[+DrVenkman]

Okay, another data point (and adding insult to injury in some ways!), the Win32 1.0.0. Preview 6 runs great from CrossOver in an XP bottle. I just had to manually set the Com port /dev entry. No pauses, no hitches, just runs. It peaks at about 10% on the CPU meter, if even that much.

 

This is some weird stuff that needs someone with serious coding-fu.

[atari8warez]

 

Anyway, I PM'd 'meliss' to see if he's interested in taking a look at the AspeQt Mac build; he's still active at AA so maybe he'll respond. Comparing how his code handles the I/O routines and opens the serial device with how AspeQt does it would be instructive.

 

Actually David Voswinkel - optixx is the person to ask because he helped me out with OS X compiles, modded the serial I/O routine to make it work with OS X and he actually uses (or used) AspeQt. He never told me there was a problem with CPU usage and/or with pauses and stutters, in any case If he comes up with a solution I will implement it in the mainline code so every OS X user will benefit from it.

 

I wasn't able to install OS X successfuly under VirtualBox so I have no means to debug the serial I/O under OS X.

Edited July 20, 2014 by atari8warez

[atari8warez]

I assume I just need to click on the AspeQt execuatable after downloading right? When I do that nothing seems to happen. I have an older version on my box (0.83) that appears to work, but it has different files and folders (no $boot folders and it has a bunch of other files that the new version doesn't).

 

Right, but make sure the OS is 32 bit. I have the same result if I try to run it on 64 bit Ubuntu. Also the binary was compiled under Ubuntu v14.04. 0.83 sounds to be awfully old for backward compatibility and in all likelihood would not have all required Qt libraries to start with.

 

You need $boot folders for folder image booting. The tar file in Sourceforge has all the required folders and files

Edited July 20, 2014 by atari8warez

[+DrVenkman]

I wasn't able to install OS X successfuly under VirtualBox so I have no means to debug the serial I/O under OS X.

 

That's the heck of it. Freetz has CPU usage but no data stutters. I am seeing both symptoms. But running your Win32 binary under CrossOver (set to a virtual XP environment) works great! Even through the overhead of OS translation and so on. I wonder if there's that much difference between Qt running on each platform?

[atari8warez]

 

That's the heck of it. Freetz has CPU usage but no data stutters. I am seeing both symptoms. But running your Win32 binary under CrossOver (set to a virtual XP environment) works great! Even through the overhead of OS translation and so on. I wonder if there's that much difference between Qt running on each platform?

 

I wish I could comment on that but my OS X knowledge is very limited to say the least, I 100% rely on optixx's mods for OS X functionality.

[atari8warez]

Here's David's AspeQt page, I think he runs it under Maverick.

 

EDIT: I emailed David and hoping to get some clues on how to solve these problems.

Edited July 20, 2014 by atari8warez

[+DrVenkman]

Thanks for your continued help, Ray. I appreciate it. Until I get a better handle on what's going on with my OS X build, I'm gonna use the Win32 version under CrossOver.

 

In the meantime, another quick question, this time regarding the 10502PC functionality. I'm working through your disk copying tutorial but what I'm finding with both of my 1050 drives is that whenever the device is connected to them, they will not function. They power on and the disk seeks Track 0. They also pass through the SIO signal chain as normal. So in your tutorial, my system sees D1: and D4: through AspeQt but my physical drive gives an ERROR 138 when I try to access it. COPY2000 doesn't see D2: at all.

I tried it first with one of my 1050's that is already configured as Drive 2 and it didn't work. I unplugged the device and booted from my normal physical D1:. After that, DOS let me see and access D2: as normal. I then pulled D1: out of the chain, plugged D2: straight into my 800XL and then daisy-chained the device off D2: again. I could boot from AspeQt and access the virtual drives fine but D2: gave me ERROR 138 again. I then swapped my physical drives. I set D1: to D2:; plugged it into my 800XL, connected the 10502PC device to that drive and booted from AspeQt again. Same results.

 

Putting my physical drives back into their original state and booting from D1: let me see both drives as normal again.

 

So for some reason, when the device is connected to my drives, it seems to disable the drives from responding. Is there anything I am doing wrong or something I can check further on my end?

[mellis]

Tried SIO Server the other day and while it ostensibly works fine with my FTDI interface, it has serious problems with double-density disks and keeps timing out. Single-density ATRs seem to boot without issue, which I find quite strange. I'd like to know what's going on with that. If SIO Server also did folder imaging (and I could iron out the transmission problems), it would be a really nice alternative.

 

Hi folks.

 

I really don't want to hijack this AspeQt thread, but I will post a few bits to contribute to the turn the conversation has taken.

 

I wrote SIO Server to scratch an itch I had a few years ago. It was designed to satisfy my own personal requirements, and then I released it to the public just in case others might derive some utility from it.

 

Until recently, SIO Server has had two active users that I know of. The double-density bug is definitely in there (and fixable), but I never did get around to correcting it because it wasn't a problem for my personal use-case and there were other peripheral emulator solutions that everyone seemed pretty happy with.

 

SIO Server is actually two programs: the GUI process and the communication daemon. When the GUI launches, it spawns a priority-elevated communication daemon as a child process. The daemon and the GUI communicate with one another using piped I/O. The approach minimizes CPU usage and helps to ensure that the priority-elevated daemon can continue to communicate even if the host system is delayed updating the GUI.

 

Although I am a long-time Mac OS X (née NEXTSTEP)/iOS developer, the GUI is written in Java in order to facilitate porting the app to other POSIX-compliant OSes. The core of the communication daemon was written in portable, POSIX-compliant C by Belboz, and I added many enhancements to it. The communication daemon has been tested on Mac OS X and Linux systems, and it can be launched from the command line.

 

In short, I never got around to fixing the double-density issue or to creating a Linux build because: I was happy with the software (for what I needed), nobody else was terribly interested, and I've had little time to spare anyway. That said, I am glad to see that the program has proven helpful in some way: as a troubleshooting tool.

[flashjazzcat]

Preview 6 runs great from CrossOver in an XP bottle. I just had to manually set the Com port /dev entry.

 

Hmmm... can't get that to work either. Tried Wine and CrossOver. Everything runs, the FTDI driver COM port is detected, connection made, but no communication takes place (mounted disks time out at the A8 end).

[Tempest]

Any answer to my question?

[flashjazzcat]

Yes: Aspeqt should simply run from the containing folder.

[atari8warez]

Any answer to my question?

 

See message #805 above.

[Tempest]

I must have missed the response. Sorry.

 

 

 

Right, but make sure the OS is 32 bit. I have the same result if I try to run it on 64 bit Ubuntu. Also the binary was compiled under Ubuntu v14.04. 0.83 sounds to be awfully old for backward compatibility and in all likelihood would not have all required Qt libraries to start with.

 

You need $boot folders for folder image booting. The tar file in Sourceforge has all the required folders and files

Ah, I have 64-bit Ubuntu (14.04). Is there a way to run it in 64-bit or a way to do 32-bit compatibility? I'll check the tar file again, but I swear it didn't install those folders.

[+DrVenkman]

 

Hmmm... can't get that to work either. Tried Wine and CrossOver. Everything runs, the FTDI driver COM port is detected, connection made, but no communication takes place (mounted disks time out at the A8 end).

 

I don't know what the problem would be then. So far as I can tell, once you specify the /dev/tty.usbserial-YOURDEVICESERIALNUMBER in AspeQt's Options menu, CrossOver should let it "reach through" OS X to grab the device and use it. In fact, given how squirrelly my Mac build of AspeQt is, using the Win32 version in CrossOver is the only way I can get things to work well.

 

As it happens, I'm spending a big chunk of tonight installing Win7 in a VirtualBox instance on my MacBook Pro [for reasons]. Interestingly, a full virtualized Win-instance seems to have no trouble seeing the SIO2PC device just fine. I hate to install Windows just to play with my Ataris, but fortunately (or not) I have several other reasons I need to use Windows lately anyway. At least this way I can keep it confined to a window (lowercase!) while I do my real stuff in OS X.

[atari8warez]

I must have missed the response. Sorry.

 

 

Ah, I have 64-bit Ubuntu (14.04). Is there a way to run it in 64-bit or a way to do 32-bit compatibility? I'll check the tar file again, but I swear it didn't install those folders.

 

I am not sure if there is a way to run it in 64 bit, but I will try to compile a 64 bit version once I install QT to my 64 bit version of Ubuntu.

[flashjazzcat]

I don't know what the problem would be then. So far as I can tell, once you specify the /dev/tty.usbserial-YOURDEVICESERIALNUMBER in AspeQt's Options menu, CrossOver should let it "reach through" OS X to grab the device and use it. In fact, given how squirrelly my Mac build of AspeQt is, using the Win32 version in CrossOver is the only way I can get things to work well.

Well, something's telling me I should stick to Win 7 for all my A8 development needs.

[mellis]

I have a question for the OS X users out there:

Assuming SIO Server's double-density bug was fixed, is there a single additional feature that would make it more usable for you? Folder imaging was mentioned earlier.

 

To me it seems running a Win32 compatibility layer is a bit heavyweight for a peripheral emulator.

[flashjazzcat]

Assuming SIO Server's double-density bug was fixed, is there a single additional feature that would make it more usable for you? Folder imaging was mentioned earlier.

Folder imaging alone would make it usable for me, but generally speaking I'd just be delighted if you continued work on SIO Server. Mac OS X builds of Aspeqt are very nice indeed, but there will always be a place for a "native" Mac peripheral server, IMO.

[+DrVenkman]

I have a question for the OS X users out there:

Assuming SIO Server's double-density bug was fixed, is there a single additional feature that would make it more usable for you? Folder imaging was mentioned earlier.

 

To me it seems running a Win32 compatibility layer is a bit heavyweight for a peripheral emulator.

 

Simple printer emulation would be the most-used feature for me - just ATASCII to ASCII translation of the character stream from the SIO interface and dump to a window, which could then be copied and pasted into TextEdit or whatever.

 

A lovely "nice-to-have" feature would be serial modem emulation for Telnet BBS access through the host's internet connection. SIO2OSX can do this but I would only use it occasionally myself.

Edited July 25, 2014 by DrVenkman

[freetz]

While I'd also like to see SIO Server continue to grow, maybe we should have a different thread for this?

If not, I'd like to see a feature to boot a xex file and refresh it on the click of a button (nice for debugging on real hardware). And close runner-up would be serial mode emulation. Of course, if any other software than APE would be able to run the 1050 standalone to transfer disks, this would get all my votes ...

[+DrVenkman]

Of course, if any other software than APE would be able to run the 1050 standalone to transfer disks, this would get all my votes ...

 

Amen.

Yeah, I'd change my vote for if this was a possibility.

[kogden]

I gave up trying to get Aspeqt going properly under OSX 10.9. SIO2OSX works well but "folder imaging" support seems to be broken on recent OSX versions. Other than that SIO2OSX has been flawless. I should probably register it one day, it's worth the $25.

 

SIOServer has a great looking interface. If the DD support was added and "folder imaging" it would be downright awesome. Peripheral emulation like printers, the 1020 plotter and "Internet Modem" emulation would be cool but not necessarily essential. Maybe slap together a plugin API and let the community handle addons if they want them bad enough?

 

Atari800MacX is a decent emulator and Eclipse/WUDSN runs great under OSX. The Mac is a decent A8 cross-development platform. I can live without Altirra and Aspeqt. The pain of dealing with Windows even in a VM or with a compatibility layer like WINE is simply not worth the effort for me. OSX and UNIX in general are a superior platform, why would I want to taint it with Winblows stuff just to play with my Atari?

[+DrVenkman]

Okay, after playing with APE/ProSystem on a virtual Win7 system on my Mac, let me just add this thought:

 

If someone could duplicate the functionality of ProSystem (connecting directly to a floppy drive) with all the myriad options, I'd happily drop the money to pay for it. I think the $50 AtariMax wants is a bit high, but $25 would be a no-brainer for a lot of people. I think he'd more than double his paying customer base if he made it cheaper.

 

Just my tuppence if anyone wants to tackle that kind of functionality natively in OS X.