RespeQt general discussion

[+DrVenkman]

I'll build new OS X and RPi2 binaries this weekend and give them a try. Looking forward to playing around with PCLINK functionality, too.

[+DrVenkman]

Just an FYI to Joey Z. as the project maintainer, and to any other other OS X users besides me - my ancient MacBook Pro has bitten the dust so I will be unable to compile binaries for future releases, at least until I get this one repaired or acquire another machine. I don't know when that might be.

[TheMontezuma]

I thought this never happens to Apple devices...

[+DrVenkman]

I thought this never happens to Apple devices...

Hah, hah, hah ... Very funny.

 

When the machine in question is 5-1/2 years old and has been used daily for probably an average of 6 - 12 hours a day, every day, for that entire time, and the machine has a defective piece of shit AMD Radeon video card that has suffered an industry-leading early failure rate way out of proportion to any other card of the era - so high that Apple has extended out-of-warranty repairs to nearly SIX YEARS after purchase - then yeah, it happens. I already had the card replace 13 months ago (after 4-1/2 years of daily use) but I don't know if Apple will replace it a second time. We'll see.

 

In the meantime, feel free to compile the OS X binaries for everyone too while you're busy making asinine jokes.

[TheMontezuma]

It was not a joke, I really did.

Hopefully your MacBook will be repaired again.

We all count on Apple in this case

[kogden]

AMD laptop GPU's from that era are famous for overheating and popping solder joints under the bga package. A lot of manufacturers (HPaq) used crappy soft heat transfer shims to the heatsink which dry out after a while resulting in an even quicker death. G3 iBooks were famous for this crap too. I've had to repair several.

 

Hopefully Apple will replace your board. If they won't, iFixit has good teardown guides for most Mac laptops if you go the DIY route. Apple usually makes good gear, my family still has plenty of old Mac hardware still running.... even PowerPC machines.

 

If I was more familiar with C++ and QT I'd try to dive in and contribute. I can barely deal with ANSI C and that was years ago.

[+DrVenkman]

Interesting development here. Mark, developer of SIO2OSX, has made the program freeware and will be open-sourcing the project "as time allows." As a paid user of the product, I can say it is an excellent and well-done product. While I'm sad that he isn't able to continue development, this is an excellent gesture on his part.

 

http://atariage.com/forums/topic/255349-atari800macx-and-sio2osx/?p=3561620

[kogden]

Interesting development here. Mark, developer of SIO2OSX, has made the program freeware and will be open-sourcing the project "as time allows." As a paid user of the product, I can say it is an excellent and well-done product. While I'm sad that he isn't able to continue development, this is an excellent gesture on his part.

 

http://atariage.com/forums/topic/255349-atari800macx-and-sio2osx/?p=3561620

 

Yeah, it's quite stable and works very well. Certainly the best OSX-based SIO2PC app at this point even if it's getting dated. I paid as well, certainly cheaper than APE. Definitely interested to see how some things work under the hood. Might take some work to get it building on newer versions of XCode though.

 

While a native Mac app and not based on QT, it'd be interesting to see some cross-pollination of ideas between RespeQT and SIO2OSX.

[KLund1]

I must be a dope, but I can't find a link here to download this software for my win10 x64 machine.

Could a smarter person than me post a link.

Thanks

[+DrVenkman]

I must be a dope, but I can't find a link here to download this software for my win10 x64 machine.

Could a smarter person than me post a link.

Thanks

 

Sorry for the delay in responding but here you go:

 

https://github.com/jzatarski/RespeQt/releases/download/r3/RespeQt-r3-win32.zip

[KLund1]

No trouble.

I got loaded it and it runs good. The application dose not scale well on a gForce driven, 4K monitor, fyi in Win 10x64, build170112-1758. Buttons are about 1/4" square. Pulling on a corner just increases the space between small buttons. Mini mode is not usable.

But, again, the program runs GREAT.

many thanks for all your hard work on this!

[+DrVenkman]

No trouble.

I got loaded it and it runs good. The application dose not scale well on a gForce driven, 4K monitor, fyi in Win 10x64, build170112-1758. Buttons are about 1/4" square. Pulling on a corner just increases the space between small buttons. Mini mode is not usable.

But, again, the program runs GREAT.

many thanks for all your hard work on this!

 

Right-click on the executable icon, select Properties, and then click the Compatibility tab. Look down to the Settings field and check the box that says "Override high DPI scaling behavior ..." In the little dropdown that becomes accessible, I use "System (enhanced)" - should take care of the issue. (My Windows machine has a 4K display and GTX1070 video card).

[KLund1]

Many thanks again.

Nothing become available with that check mark.

I do not see "System(enhanced)" (can get 8-bit and 16-bit color)

 

[+DrVenkman]

Many thanks again.

Nothing become available with that check mark.

I do not see "System(enhanced)" (can get 8-bit and 16-bit color)

 

 

Hmm, must be some variation due to graphic drivers or Windows sub-version. This is what I see:

 

 

 

And with that selected, I see a nice, reasonably sized and legible application.

 

[KLund1]

Yes most likely. I'll check it after new builds load to see if that option appears.

I'm on the Insider fast loop MS program, so I'm always running beta's. Who knows what I get.

Also your video card is a bit better then mine. I'm only a GTX1060 OC. But I'm running Samsung's 2x sm961 m.2 in raid 0! Seems like stuff loads before I click it!

But thanks once again for build a kill 8-bit app for us all here!

[+DrVenkman]

I saw a number of changes boucing through the Github emails today so I pulled the latest changes to my local copy and just built a new executable on my RPi ZeroW. Seems to be working just fine.

[TheMontezuma]

I have just built the binaries for Windows, OSX and Raspberry Pi:

https://github.com/TheMontezuma/RespeQt/releases/tag/r4rc2

 

On top of the recent changes I added a new backend based on the Qt Serial Port (it is not better than the standard one, it is just another option worth to try if there are problems with standard backend).

Please download, test and report any bugs you find.

[+DrVenkman]

I have just built the binaries for Windows, OSX and Raspberry Pi:

https://github.com/TheMontezuma/RespeQt/releases/tag/r4rc2

 

On top of the recent changes I added a new backend based on the Qt Serial Port (it is not better than the standard one, it is just another option worth to try if there are problems with standard backend).

Please download, test and report any bugs you find.

 

So you're calling this RC4 ... did you fork from Joey Z's repository? Just curious about the etiquette involved here.

 

I downloaded your RPi binary and gave it a run on my RPi Zero W. The binary you built is a bit different than the one I built from Joey's Master branch a few days back; it seems to require launching by double-clicking or invoking the shell script.

 

Now having said all that, I gave the Qt backend a quick runthrough and it seems to work fine all the way up to POKEY divisor 0 with handshaking set to NONE.

[TheMontezuma]

So you're calling this RC4 ... did you fork from Joey Z's repository? Just curious about the etiquette involved here.

 

Sorry for confusion.

Yes, I created a fork from Joey's repository (like all contributors, this is how github works) and I keep it in sync with the official one.

My fork is not supposed to be another version of AspeQt/RespeQt.

I only wanted to build up-to-date binaries, so everyone can give it a try (I'm pretty sure the most users to do have Qt environment for compiling RespeQt).

The binaries could be uploaded here at AtariAge, but I thought Github is a better place for uploading huge binaries.

And I incremented the version number (called it r4 RC2) just to have a way to distinguish the binaries.

 

To summarize: Joey is the maintainer of RespeQt and he creates official releases (I already created a PULL request with my Qt backend changes)

 

I downloaded your RPi binary and gave it a run on my RPi Zero W. The binary you built is a bit different than the one I built from Joey's Master branch a few days back; it seems to require launching by double-clicking or invoking the shell script.

 

There are two ways of getting RespeQt to run on Raspberry Pi:

 

1) You can install Qt development libraries and compile RespeQt. In this case you can start the RespeQt binary directly, because all needed shared libraries will be found automatically.

 

Note: QSerialPort is not included in qt-default / qtbase, so you need to install libqt5serialport5-dev in order to compile RespeQt:

sudo apt-get update
sudo apt-get install build-essential git qt5-default qtbase5-dev libqt5serialport5-dev

2) You can download the RPI RespeQt binaries from Github and unpack them in the pi's home directory (without installing Qt).

This time the required shared libraries are not included in the LD_LIBRARY_PATH, so launching the RespeQt binary will fail with a message, that one of the shared libraries is missing. That's why I created a scrpit (respeqt.sh), which is setting up the environment and launching RespeQt:

#!/bin/sh
#sudo rfcomm bind /dev/rfcomm0 98:D3:31:B0:95:A7
export LD_LIBRARY_PATH=./lib:$LD_LIBRARY_PATH
export QT_QPA_PLATFORM_PLUGIN_PATH=./lib
./RespeQt

If you use Bluetooth, you can uncomment the second line and replace the BT address with the address of your SIO2BT Bluetooth module. It will create /dev/rfcomm0 device, which can be used in RespeQt.

 

I also created a file called RespeQt.desktop:

[Desktop Entry]
Name=RespeQt
Exec=lxterminal --working-directory=/home/pi/RespeQt/ -e ./respeqt.sh
Icon=/home/pi/RespeQt/RespeQt.ico
Type=Application

Please copy it to your desktop and you will get a RespeQt icon on the desktop and will be able to launch RespeQt by double clicking the icon.

 

Now having said all that, I gave the Qt backend a quick runthrough and it seems to work fine all the way up to POKEY divisor 0 with handshaking set to NONE.

 

Great news.

 

The Qt serial backend is based on the Qt serial port abstraction.

This means the same RespeQt code is used for handling low level communication under Windows, Linux or Mac. At the end, the QSerialPort implementation uses native API of the corresponding OS, like it is done in the Standard Backend right now. From the first point of view, you may see no benefit of introducing new backend, but the benefit is that if the user encounters any problem with the native implementation of the Standard Backend, he can switch the backend and perhaps have better luck with the Qt code. At the time of writing of the AspeQt, there was no serial port support in the Qt, so there was no other choice, as to write it on your own. Today we can benefit of having two approaches

 

Setting the DTR/RTS line occurred to be necessary when using Arduino Micro Pro (as a SIO2PC).

There is a new project called SIO CART which is using this hardware and they wanted to make the USB port usefull for users.

Edited March 23, 2017 by TheMontezuma

[TheMontezuma]

I have just tried to run the RespeQt binaries on my old Windows XP laptop and was not able to start it (got an error message).

 

Google told me, that Qt 5.8 does not support Windows XP anymore:

https://bugreports.qt.io/browse/QTBUG-58407

 

That's a pity. I guess many old machines are used as file servers for ATARI, so do not let them die

I downgraded the Qt software to 5.7.1 and recompiled RespeQt. The version uploaded to Github (link above) works well under Windows XP.

Edited March 23, 2017 by TheMontezuma

[JoSch]

I saw a number of changes boucing through the Github emails today so I pulled the latest changes to my local copy and just built a new executable on my RPi ZeroW. Seems to be working just fine.

Two of them was me. I fixed the application exit crash on OSX and made some minor changes to the code.

 

 

 

Sorry for confusion.

Yes, I created a fork from Joey's repository (like all contributors, this is how github works) and I keep it in sync with the official one.

My fork is not supposed to be another version of AspeQt/RespeQt.

I only wanted to build up-to-date binaries, so everyone can give it a try (I'm pretty sure the most users to do have Qt environment for compiling RespeQt).

The binaries could be uploaded here at AtariAge, but I thought Github is a better place for uploading huge binaries.

And I incremented the version number (called it r4 RC2) just to have a way to distinguish the binaries.

 

To summarize: Joey is the maintainer of RespeQt and he creates official releases (I already created a PULL request with my Qt backend changes)

I'm working on a branch for printer support. I have the basic structure down, but need to test the code, before I can release it.

There are also some structural C++ questions, I still have to answer.

So please see this announcement as a teaser.

Edited March 23, 2017 by JoSch

[+DrVenkman]

Thanks for the clarifications, gentlemen. Much appreciated.

​Anyone making any further effort towards Atari 850 interface emulations and/or some kind of R: terminal emulation so we can connect to telnet BBS's? There was some discussion about 18 months ago when Joey got started on it.

[TheMontezuma]

Thanks for the clarifications, gentlemen. Much appreciated.

​Anyone making any further effort towards Atari 850 interface emulations and/or some kind of R: terminal emulation so we can connect to telnet BBS's? There was some discussion about 18 months ago when Joey got started on it.

 

I don't know if anybody commited to do it so far. Since it is somehow related to networking functionality from my SIO2BT project I can take a look at it.

At least I have everything (books, specs, tools, hardware and software) required to work on R: device emulation

[+DrVenkman]

 

I don't know if anybody commited to do it so far. Since it is somehow related to networking functionality from my SIO2BT project I can take a look at it.

At least I have everything (books, specs, tools, hardware and software) required to work on R: device emulation

 

If anyone could pull it off, I'm confident you could.

​

​As an aside, it's nice to see RespeQt slowly gaining functionality, cleaner and more stable code, etc. It seems like forever (less than two years) since I downloaded the OS X code and tried to kludge up some kind of fix for the timeouts and CPU usage spikes. I'm so glad people who know what they're doing have taken it upon themselves to fix things properly!

[JoSch]

So, to give a milestone on the printer support: I could print out a text to a simulated 1027 printer.

So the question now is, how to go about configuring the native printer: Should it be the same for all four P: devices, or should it be configurable for each P:?

What do you think?