Jump to content

Photo

RespeQt r3 released

r3

56 replies to this topic

#1 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • 876 posts
  • Location:Hoffman Estates, IL

Posted Wed Mar 16, 2016 4:09 PM

RespeQt release 3
    * removed settings migration feature due to foreseeable
    incompatibilities between AspeQt settings and RespeQt settings
    * code cleanup by TheMontezuma to reorganize various SIO constants into
    a pair of enumerations
    * updated contact info in readme.txt (thanks to Kyle22 IIRC)
    * fix for hardware handshaking issue in Windows which could cause data
    which are not commands to be interpreted as commands. (issue #21)
    * fix for performance issue with Software Handshaking using wired
    SIO2PC (issue #21)
    * Software handshaking for bluetooth (TheMontezuma)
    * fix for serial port name in log windows (TheMontezuma)
    * format command timeout value fixed (TheMontezuma)
    * bug fix for COM ports higher than COM9 in windows (TheMontezuma)
    * Spanish language update, thanks to ascrnet
    * reorganization of messy HTML in about.html, thanks to ascrnet
    * Spanish version is included in the about.html, thanks to ascrnet
    * fixed broken language option in about dialog (#14)
    * reorganization of messy HTML in RespeQt User Manual-English.html,
    thanks to ascrnet
 
NOTE: software and 'none' handshaking have small risks of data loss when used alongside real atari devices. The risk arises from the slight possibility that some consecutive 5 bytes will form a valid format command and cause a drive in RespeQt to format. The risk is miniscule, but existent, so it's worth mentioning. Hardware handshaking does not carry this risk.
 
https://github.com/j...espeQt/releases
 
There is source, as well as a win32 build, and I suspect DrVenkman will upload an OSX build when he gets a chance.
 
I am officially declaring the win32 build as stable now, since nobody has reported any issues with the win32 releases that have not been fixed.
 
the current ever-growing todo list:
  put NONE and software handshake data loss warnings in documentation (there's always something I forget to do when a release comes around)
  need to deal with AspeCL at some point
  translations
  sioworker.cpp: have devices provide their own string for debug output? - SioWorker::deviceName
  Qt4 version?
  R: support
  folder imaging write support
  SDFS folder imaging
  better printer support - MX-80 and 1020, possibly others. direct to file, full 8 bit support, etc.
  VAPI/ATX
  conditional compilation for nativeMenuBar stuff in OSX
  PCLINK.SYS support?
  XF551 compatibility issue: http://atariage.com/...ased/?p=3427201

#2 DrVenkman OFFLINE  

DrVenkman

    River Patroller

  • 3,953 posts
  • Back off, man! I'm a scientist.
  • Location:KMBT

Posted Wed Mar 16, 2016 4:39 PM

Updating  Xcode now. A new OS X build should be up tonight. 

 

EDITED: Okay, Github has an OS X build available now, and I'll attach it directly here as well. 

 

Attached File  RespeQt-OSX.zip   9.77MB   294 downloads


Edited by DrVenkman, Wed Mar 16, 2016 5:18 PM.


#3 TheEditor OFFLINE  

TheEditor

    Chopper Commander

  • 162 posts

Posted Wed Mar 16, 2016 8:44 PM

Has git been updated so I can compile a new Pi version?

Sent from my LG-V496 using Tapatalk

#4 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Wed Mar 16, 2016 9:00 PM

Has git been updated so I can compile a new Pi version?

Sent from my LG-V496 using Tapatalk

Yes, git stays up to date. Still, I recommend you download source from the releases page. The 'master' source may have unstable changes. Currently, it doesn't, but at any given time it could.



#5 GoodByteXL OFFLINE  

GoodByteXL

    Moonsweeper

  • 339 posts
  • Location:XL heaven

Posted Thu Mar 17, 2016 3:55 AM

RespeQt release 3
...

Great, thanks for your ongoing efforts, will try it later today.

@AtariXLE:
RespeQt version 1 & 2 for linux (debian 64) work fine (except some hassles with SIO highspeed).
will you catch up soon?

#6 TheMontezuma OFFLINE  

TheMontezuma

    Dragonstomper

  • 674 posts
  • Location:Hildesheim, D / Kraków, PL

Posted Thu Mar 17, 2016 11:20 AM

I hope we can count on Hias regarding SIO highspeed improvements for linux ;)



#7 GoodByteXL OFFLINE  

GoodByteXL

    Moonsweeper

  • 339 posts
  • Location:XL heaven

Posted Thu Mar 17, 2016 2:12 PM

I hope we can count on Hias regarding SIO highspeed improvements for linux icon_wink.gif

I guess Hias is the only A8 user with appropriate insights :-).

RespeQt R3 so far works fine on Win 7-64 & 32. Not tested yet on Win 10.

While reading the user manual I spotted an editing failure: the first mentioning of SD 3.2f is an incomplete paragraph.
The second mentioning is the same para but complete, as it seems. Nevertheless it needs correction:

At first it should read SD 3.2G as version F is a buggy version superseeded by G as declared by FTE when it was released as shareware back in 1994.

Then

You can not load drivers or run an autorun.bat file during the boot process

was fixed already with version AspeQt 0.6 iirc,
 

Once booted from, a Folder Image won't be bootable again until ...

same as above. Enter RUN E477 on the command line and watch it reboot and load e.g. a rtc driver.

SpartaDOS and SpartaDOS X work pretty good with AspeQt/RespeQt on Win (Speed PD 0) and Linux (Speed PD 4), cannot say for MacOX.

The speed issue under Linux seems to be related to a failure in timing adjustments between the A8 and the pc side as they do not exist under Win. One can help this by using a FIFO on the A8 side, which accelerates up to PD 0 (e.g. SD 3.2G) depending on the capabilities of the software.

What puzzles me most is that the UltraSpeed driver of the IDE+ 2.0 (Rev C here) speeds up every software always under Win to PD 0 whereas the same under Linux leads to instant timing errors and stops transfering.

The pc hardware for both Win 7-64 and Linux 64 is the same, just two different hard drives.

#8 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Thu Mar 17, 2016 2:35 PM

RespeQt R3 so far works fine on Win 7-64 & 32. Not tested yet on Win 10.


good to hear.
 

While reading the user manual I spotted an editing failure: the first mentioning of SD 3.2f is an incomplete paragraph.
The second mentioning is the same para but complete, as it seems. Nevertheless it needs correction:

At first it should read SD 3.2G as version F is a buggy version superseeded by G as declared by FTE when it was released as shareware back in 1994.


That's not incorrect. RespeQt comes with a copy of 3.2F, that's why the docs are that way. I haven't changed the boot files since I forked it. Would someone get mad if I stripped the boot sectors and files off a 3.2G disk?
 

Then
was fixed already with version AspeQt 0.6 iirc,
 
same as above. Enter RUN E477 on the command line and watch it reboot and load e.g. a rtc driver.

SpartaDOS and SpartaDOS X work pretty good with AspeQt/RespeQt on Win (Speed PD 0) and Linux (Speed PD 4), cannot say for MacOX.


Have you tested these recently? I'd like to be sure before I update the manual.
 

The speed issue under Linux seems to be related to a failure in timing adjustments between the A8 and the pc side as they do not exist under Win. One can help this by using a FIFO on the A8 side, which accelerates up to PD 0 (e.g. SD 3.2G) depending on the capabilities of the software.

What puzzles me most is that the UltraSpeed driver of the IDE+ 2.0 (Rev C here) speeds up every software always under Win to PD 0 whereas the same under Linux leads to instant timing errors and stops transfering.

The pc hardware for both Win 7-64 and Linux 64 is the same, just two different hard drives.


The speed issue is a result of all USB serial adapters being horrible in general. You have one timing setting, and you get issues at 19200. Another timing value, you've got issues at divisor 0. USB-serial adapters just don't have accurate enough timing to work for everything. As I said before somewhere else, it's a real crapshoot.

 

EDIT: source for the windows and linux serial backends are separate, which explains the difference between windows and linux performance, BTW.


Edited by Joey Z, Thu Mar 17, 2016 2:37 PM.


#9 DrVenkman OFFLINE  

DrVenkman

    River Patroller

  • 3,953 posts
  • Back off, man! I'm a scientist.
  • Location:KMBT

Posted Thu Mar 17, 2016 4:06 PM

SpartaDOS and SpartaDOS X work pretty good with AspeQt/RespeQt on Win (Speed PD 0) and Linux (Speed PD 4), cannot say for MacOX.
 

 

 

I don't bother messing around with manually setting a POKEY divisor but I can tell you that through r2, RespeQt works great at 38,400 on OS X and through 57,600 on my RPi2. I haven't actually done any speed tests with r3 yet, nor compiled it on the RPi2.



#10 TheMontezuma OFFLINE  

TheMontezuma

    Dragonstomper

  • 674 posts
  • Location:Hildesheim, D / Kraków, PL

Posted Fri Mar 18, 2016 2:32 AM

I attached the R3 Binaries for Raspberry Pi:

Attached File  RespeQt-r3.tar.gz   7.82MB   208 downloads

 

@Joe

Could you please upload it to Github?

 

How to get a desktop icon:

1) Extract "RespeQt-r3" to your HOME folder (/home/pi/)

2) Open a terminal and type in:
 

cd Desktop
ln -s ../RespeQt-r3/RespeQt.desktop .

Now you can just doubleclick the icon on the desktop.

 

The tool is started via respeqt.sh script.

You may edit the script and uncomment the line for automatic binding of the Bluetooth device (just remove the '#' character at the beginning of the line and edit the BT address).

 

BTW:

I tested the Windows binaries on the WIN10 (64bit) and on WIN7 (64bit).


Edited by TheMontezuma, Fri Mar 18, 2016 2:36 AM.


#11 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Fri Mar 18, 2016 2:37 AM

@Joe
Could you please upload it to Github?


done

#12 GoodByteXL OFFLINE  

GoodByteXL

    Moonsweeper

  • 339 posts
  • Location:XL heaven

Posted Fri Mar 18, 2016 11:22 AM

 

GoodByteXL, on 17 Mar 2016 - 9:12 PM, said:
Nevertheless it needs correction:
At first it should read SD 3.2G as version F is a buggy version superseeded by G as
declared by FTE when it was released as shareware back in 1994.

That's not incorrect. RespeQt comes with a copy of 3.2F, that's why the docs are
that way. I haven't changed the boot files since I forked it. Would someone get mad
if I stripped the boot sectors and files off a 3.2G disk?

 

Though having no programming capabilities I would go for the most accurate version.

Does it have any outcome which SD version is used?
 

 

GoodByteXL, on 17 Mar 2016 - 9:12 PM, said:
SpartaDOS and SpartaDOS X work pretty good with AspeQt/RespeQt on Win (Speed PD 0)

and Linux (Speed PD 4), cannot say for MacOX.

Have you tested these recently? I'd like to be sure before I update the manual.

 

This is daily business as I use it that way. Currently testing SDX 4.48 for the
updates to the user guide and frequently using SD 3.2G and also Steven's RealDOS, I
haven't seen it fail in more than a decade.  
 

 

GoodByteXL, on 17 Mar 2016 - 9:12 PM, said:
The speed issue under Linux seems to be related to a failure in timing adjustments
between the A8 and the pc side as they do not exist under Win. One can help this by
using a FIFO on the A8 side, which accelerates up to PD 0 (e.g. SD 3.2G) depending
on the capabilities of the software.

What puzzles me most is that the UltraSpeed driver of the IDE+ 2.0 (Rev C here)
speeds up every software always under Win to PD 0 whereas the same under Linux
leads to instant timing errors and stops transfering.

The pc hardware for both Win 7-64 and Linux 64 is the same, just two different hard
drives.

The speed issue is a result of all USB serial adapters being horrible in general.
You have one timing setting, and you get issues at 19200. Another timing value,
you've got issues at divisor 0. USB-serial adapters just don't have accurate enough
timing to work for everything. As I said before somewhere else, it's a real
crapshoot.

 

I wouldn't sign that as my experiences are different. It seems to depend very much on the hard-/software combo on the pc side and how it manages the USB ports. Even when using the first USBtoSerial breakout boards from prolific were working well as soon as the right conditions for the pc combo was found.

Same applies to the story about the 4 capacitators in XL/XE which are recommended to be de-soldered. I never experienced any influence of these concerning SIO2PC with RS232 or USB2Serial. Nevertheless, following the proposal of the makers of the IDE+ I desoldered last week those caps in a 600XL to see if this causes the speed problems under Linux when using the Ultra Speed driver of the IDE+. No difference.
 

EDIT: source for the windows and linux serial backends are separate, which

explains the difference between windows and linux performance, BTW.

I guess that's the real reason why the performance is so different. With Win7 I even had "waving" transfer rates with a Asus M3N mainboard and a AMD processor up to PD 0. It sounded like a record player that cannot keep its rpm. To "fix" this I started a tool to get the processor to 100%. With Linux on the same system the max. was 3xSIO.

 

 

I don't bother messing around with manually setting a POKEY divisor...

 

Hey, that's what it's all about, having a real fast SIO connection to the pc/mac. My preset PD is always 0 until something else proofs a need to change it.

 

P.S.: Sorry for the bad quoting, but the forum software under FF on Win7 and Linux doesn't allow any proper writing here. I have to do copy & paste from/to a text editor.



#13 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Fri Mar 18, 2016 5:40 PM

This is daily business as I use it that way. Currently testing SDX 4.48 for the
updates to the user guide and frequently using SD 3.2G and also Steven's RealDOS, I
haven't seen it fail in more than a decade.


OK, just to be clear, you tested these issues specifically when booting sparta from a bootable folder, right? That's when the issues listed in the docs occur.
 

I wouldn't sign that as my experiences are different. It seems to depend very much on the hard-/software combo on the pc side and how it manages the USB ports. Even when using the first USBtoSerial breakout boards from prolific were working well as soon as the right conditions for the pc combo was found.


Well yeah, that's what I'm saying. There's no way to be sure you're going to get good results. USB to serial converters just aren't going to be consistent from machine to machine, and aren't even all that consistent byte to byte. For that matter, real serial ports might not be either, although that's been less of an issue in my experience. But if you look at serial data transfer with an oscilloscope, you can easily see the timing jitter on the USB-serial converters.



#14 GoodByteXL OFFLINE  

GoodByteXL

    Moonsweeper

  • 339 posts
  • Location:XL heaven

Posted Sat Mar 19, 2016 6:09 AM

OK, just to be clear, you tested these issues specifically when booting sparta from a bootable folder, right?

Ups, you're true, I was riding the wrong horse, sorry. The "bootable folder" A:/D1: isn't really usable with SpartaDOS. The folder option gets just used here to copy files to hard disk images.

Nevertheless, if SD 3.2G could be made available, it would omit the errors built into SD 3.2F (hopefully).

#15 atarixle OFFLINE  

atarixle

    Moonsweeper

  • 316 posts
  • Location:Germany

Posted Sun Mar 20, 2016 3:13 AM

Get my Linux builds of R3 here:

 

http://www.abbuc.de/...wnload/#respeqt



#16 atarixle OFFLINE  

atarixle

    Moonsweeper

  • 316 posts
  • Location:Germany

Posted Sun Mar 20, 2016 3:34 PM

Thanks to the power of the VirtualBox, you get five builds of RespeQt R3 here: http://www.abbuc.de/...wnload/#respeqt

 

it is built for

debian 8.2 x86

debian 8.2 x86_64

debian 7.9 x86_64 QT5static

Ubuntu 12.04 LTS x86_64 QT5static

Ubuntu 14.04 LTS x86_64 QT5static


Edited by atarixle, Sun Mar 20, 2016 3:34 PM.


#17 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Sun Mar 20, 2016 9:03 PM

Ups, you're true, I was riding the wrong horse, sorry. The "bootable folder" A:/D1: isn't really usable with SpartaDOS. The folder option gets just used here to copy files to hard disk images.

Nevertheless, if SD 3.2G could be made available, it would omit the errors built into SD 3.2F (hopefully).

It seems extracting the boot files out of a spartados ATR is a nontrivial process. I took a quick look at it just now, but I don't think I'll be able to make very much progress right now without taking a more in depth look. Given what time it is, I think I'll just go to bed instead, and catch up on lost sleep in the last semester of school.



#18 GoodByteXL OFFLINE  

GoodByteXL

    Moonsweeper

  • 339 posts
  • Location:XL heaven

Posted Mon Mar 21, 2016 11:17 AM

Thanks to the power of the VirtualBox, you get five builds of RespeQt R3 here: http://www.abbuc.de/...wnload/#respeqt

 

it is built for

debian 8.2 x86

debian 8.2 x86_64

debian 7.9 x86_64 QT5static

Ubuntu 12.04 LTS x86_64 QT5static

Ubuntu 14.04 LTS x86_64 QT5static

debian 8.2 x86_64 works just fine under Linux Mint Cinnamon 17-2 64bit  :-)

 

It seems extracting the boot files out of a spartados ATR is a nontrivial process. I took a quick look at it just now, but I don't think I'll be able to make very much progress right now without taking a more in depth look. Given what time it is, I think I'll just go to bed instead, and catch up on lost sleep in the last semester of school.

Don't know anything about this process, but if I can help please let me know.



#19 morelenmir OFFLINE  

morelenmir

    Stargunner

  • 1,544 posts
  • Location:West Yorkshire, Great Britain

Posted Mon Mar 21, 2016 11:42 AM

RespeQt release 3
    * removed settings migration feature due to foreseeable
    incompatibilities between AspeQt settings and RespeQt settings
    * code cleanup by TheMontezuma to reorganize various SIO constants into
    a pair of enumerations
    * updated contact info in readme.txt (thanks to Kyle22 IIRC)
    * fix for hardware handshaking issue in Windows which could cause data
    which are not commands to be interpreted as commands. (issue #21)
    * fix for performance issue with Software Handshaking using wired
    SIO2PC (issue #21)
    * Software handshaking for bluetooth (TheMontezuma)
    * fix for serial port name in log windows (TheMontezuma)
    * format command timeout value fixed (TheMontezuma)
    * bug fix for COM ports higher than COM9 in windows (TheMontezuma)
    * Spanish language update, thanks to ascrnet
    * reorganization of messy HTML in about.html, thanks to ascrnet
    * Spanish version is included in the about.html, thanks to ascrnet
    * fixed broken language option in about dialog (#14)
    * reorganization of messy HTML in RespeQt User Manual-English.html,
    thanks to ascrnet
 
NOTE: software and 'none' handshaking have small risks of data loss when used alongside real atari devices. The risk arises from the slight possibility that some consecutive 5 bytes will form a valid format command and cause a drive in RespeQt to format. The risk is miniscule, but existent, so it's worth mentioning. Hardware handshaking does not carry this risk.
 
https://github.com/j...espeQt/releases
 
There is source, as well as a win32 build, and I suspect DrVenkman will upload an OSX build when he gets a chance.
 
I am officially declaring the win32 build as stable now, since nobody has reported any issues with the win32 releases that have not been fixed.
 
the current ever-growing todo list:
  put NONE and software handshake data loss warnings in documentation (there's always something I forget to do when a release comes around)
  need to deal with AspeCL at some point
  translations
  sioworker.cpp: have devices provide their own string for debug output? - SioWorker::deviceName
  Qt4 version?
  R: support
  folder imaging write support
  SDFS folder imaging
  better printer support - MX-80 and 1020, possibly others. direct to file, full 8 bit support, etc.
  VAPI/ATX
  conditional compilation for nativeMenuBar stuff in OSX
  PCLINK.SYS support?
  XF551 compatibility issue: http://atariage.com/...ased/?p=3427201

 

Many thanks Joey Z!!!  I am looking forward to testing this out.

 

If its a question of votes then please count me as a 'Yay' for removing that question mark from the PCLINK.SYS entry!!!



#20 TheMontezuma OFFLINE  

TheMontezuma

    Dragonstomper

  • 674 posts
  • Location:Hildesheim, D / Kraków, PL

Posted Mon Mar 21, 2016 12:49 PM

I can report some progress regarding PCLINK. The sio2bsd code (integrated locally to RespeQt) compiles, however does not work yet. I'm working on that, but it will take a few more days.



#21 morelenmir OFFLINE  

morelenmir

    Stargunner

  • 1,544 posts
  • Location:West Yorkshire, Great Britain

Posted Tue Mar 22, 2016 6:27 AM

A few weeks ago I mentioned I was experiencing a total connectivity failure with RespeQT seeming to lose or 'forget' its virtual COM port via my SIO2PC, despite being plugged in all the time.  The consensus seemed to be this was a local driver issue, which did indeed seem logical.  I may be jumping the gun a little, but I think I have solved the problem.  After some investigation I discovered there were at least three separate copies of the FTDI drivers on my machine; two had been installed from the official releases and one from Microsoft via Windows Update.  I guess it could have been these pieces of software were conflicting with one another and causing the issue.  Accordingly I used the 'FTClean' tool which used to be available directly from tthe FTDI 'Tools' website and supposedly removes every trace of their drivers.  This it seemed to do - with the qualification that while the official FTDI releases were indeed deleted, the Windows Update/Microsoft drivers were not.  Whatever the case, after rebooting several times - principally due to the totally unrelated and ongoing nVidia 'black screen' issues - and leaving the machine overnight I returned this morning to power-up both PC and A8 to find...  RespeQT was immediately recognized as an SIO device by the A8 and its emulated drives worked fully.

 

So; a fix?  It seems so for now, although I have been hopeful in the past and then found that problems lingered.  It is also possible that the new code introduced with r3 also aided the problem.  I guess time will tell.

 

For those interested I have included a copy of the 'FTClean' utility here as it seems to be no longer available direct:

 

Attached File  FTDI Cleaner.rar   369.94KB   178 downloads

 

So far as I know it is free of viruses/malware and I have used the app several times myself without seeming harm.  However, it goes without saying that all files downloaded from the internet should be treated with a degree of suspicion until you are personally satisfied with their safety.



#22 GoodByteXL OFFLINE  

GoodByteXL

    Moonsweeper

  • 339 posts
  • Location:XL heaven

Posted Tue Mar 22, 2016 12:15 PM

For those interested I have included a copy of the 'FTClean' utility here as it seems to be no longer available direct:

Should be here for Windows:
http://www.ftdichip....GUI_Readme.html

#23 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Tue Mar 22, 2016 7:20 PM

Ups, you're true, I was riding the wrong horse, sorry. The "bootable folder" A:/D1: isn't really usable with SpartaDOS. The folder option gets just used here to copy files to hard disk images.

Nevertheless, if SD 3.2G could be made available, it would omit the errors built into SD 3.2F (hopefully).

I think I have managed to extract the boot files out of the 3.2G ATR I found, can you please see if you can test this in place of the $bootspa/$boot.bin file for folder boot of sparta?

 

EDIT: old file was wrong, fixed now.

 

EDIT: file was wrong again, working on fixing now. fixed now hopefully

Attached Files


Edited by Joey Z, Tue Mar 22, 2016 8:08 PM.


#24 Joey Z OFFLINE  

Joey Z

    Dragonstomper

  • Topic Starter
  • 876 posts
  • Location:Hoffman Estates, IL

Posted Tue Mar 22, 2016 8:36 PM

I think I have managed to extract the boot files out of the 3.2G ATR I found, can you please see if you can test this in place of the $bootspa/$boot.bin file for folder boot of sparta?

 

EDIT: old file was wrong, fixed now.

 

EDIT: file was wrong again, working on fixing now. fixed now hopefully

This file has been reported as working, and has been pushed to git. It will be included in the next release.



#25 Bikerbob OFFLINE  

Bikerbob

    Dragonstomper

  • 686 posts
  • Location:Mississauga ON Canada

Posted Sat Mar 26, 2016 9:30 PM

This fork of the software works with the original Nick Kennedy serial SIO2PC does it not? I just pulled mine out and want to start using it with a windows based interface.

 

thanks

 

James





Reply to this topic



  


0 user(s) are browsing this forum

0 members, 0 guests, 0 anonymous users