AUFIT Program

[Maartau]

New version 0.9.5515.26942

I'm @ version 0.9.5515.26892... Is there any update ? (In that case, auto-update don't run for me too )...

[DrCoolZic]

Yes this is the last release.
I am currently busy on other stuff ... so sorry nothing new

[DrCoolZic]

Here is a new version of Aufit
- Aufit 1.0.5607.30063
- BatchAufit 1.0.5609.30951

The main reason for this new version is to be able to run Aufit in batch mode (automatic mode) in order to produce disk_view, protections, and Pasti stx images without user intervention. This allow for example to analyze a complete directory tree that contains all your games images in KF or SCP format.

This is done by invoking Aufit from a command line with some specific line arguments... As using btach command line does not seems very exciting this new version comes with a new program witha nice GUI called BatchAufit that will call Aufit with the right parameters for you. You just need to specify the root directory and what type of output you want and the tree will be analyzed for you ... can take lot of time even on multi-processor machine ...
For more information read the documentation and try it ...

Aufit 1.0 has few improvements as well
- Added detection of Jupiter Masterdrive & Dragonflight protections
- Added display of out of band transitions in track data graph
- disk graphic view now contains image name & Aufit version
- Miscellaneous bug fixes
- Added auto run feature to allow batch usage of Aufit
line arg: -disk_view, -protections, -save_image, -scale x.x, suffix zz
- New project BatchAufit to run Aufit in batch mode

For now the new Aufit will not be published so you need to download from here and place everything in one directory.
To use BatchAufitexe make sure you have Aufit.exe and Ookii.Dialogs.Wpf.dll all in the same directory!

BatchAufit.rar

Aufit-doc.rar

Please report any problem

[Mug UK]

Nice work DrCoolZic

Sent by my Note III

[Brume]

Very efficient feature, thank you very much!
You saved my day (and my week)

I know you are very busy, but may I ask an ultimate feature?
Once a SCP or a RAW file is loaded, would it possible to reload it from one click? A button that opens directly last file open. It may be helpful for people like me (I'm lazy, I now) : when a floppy is damaged, I rewrite a dump on another one, until the disk is read fine.
Just a suggestion, as always

[DrCoolZic]

Based on requests new version of Aufit and Batch Aufit.

Aufit 1.0.5610.32530
- last analyzed image file is saved upon exiting Aufit. This allow to re-analyzed the last image by clicking "reload" button.
- added reload button to reanalyze the last analyzed image (either in the current run or last run of Aufit)
- in auto/bacth mode Aufit does not create anymore an xxx.done file.

BatchAufit 1.0.5610.31593
- detection of already analyzed image is based on presence of output for disk_view, or protections, or pasti file(s)
- suffix is set by default to Aufit.1.0

BatchAufit.rar

[Brume]

That was fast, thank you very much for the new features
The reload button is perfect!

[DrCoolZic]

I have problem in Batch mode: I draw disk information at same time I am still processing data....
On a machine with 8 thread this seems to cause problem on some images. Works in normal mode and crash on batch mode?
I will investigate tomorow

[Maartau]

[DrCoolZic]

Here is a new release of Aufit and Batch Aufit.
Release Aufit 1.0.5612.16387 & BatchAufit 1.0.5612.26000 - Aufit batch mode workaround + Improved batch program

I have found a hard to fix problem when using Aufit in Batch mode. As you know Aufit is a multi-threaded program capable of using up to 168 threads in some cases! But asynchronous multi-threaded programs can become extremely tricky sometimes. In the case of Aufit in batch mode you may have noticed that the program starts to draw the disk view while at same times it finishes to process the data from the read image. On a multi-core hyper threaded system you may encounter some “racing” condition where Aufit try to draw some disk data that have not yet fully processed. And believe me this is not as easy to fix as it sounds. I have implemented a work around that seems to work in most (all?) cases. I have run Aufit in batch mode on a mixture of more than 400 SCP/Kryoflux images and Aufit in batch mode failed only 1 time on my system with CPU 4 cores 8 hyper threads. I will continue to investigate for a better solution but you may have to live with this workaround for some times. I will not be able to work on Aufit for the coming weeks (travelling ).

While modifications of Aufit are focused only on this “racing” problem, BatchAufit received a major uplift.
Here is what have changed:

• BatchAufit has now a reload button working as Aufit. The program remember the last used Root directory and the reload button uses that saved last directory. This allow to start analyzing a large tree, to stop the analysis and exit program, to restart the next day BatchAufit to reload and continue where stopped.
• BachAufit has a “save log” button to save a log file of what has been done (content of the middle result window).
  The result/log window display more information: When the image analysis has started, if the analysis has succeed or failed, the elapsed time, and the CPU time used to analyzed this image. Note that on a multi-core multi thread cpu this time is bigger than the elapsed time. Typically on a system with 4 core the cpu time is roughly 4 times larger than the elapsed time, but it can be even more than that if the cpu has hyper threading capabilities (Microsoft definition: On a multiple processor system, this value would include time spent on each processor, if the thread used more than one processor).
• At the end of analysis of all images (either all done or stop button clicked) the program display the number of files processed, the number of files processed successfully, the number of files processed incorrectly, the total elapsed time and the total cpu time.
• A new textbox called MaxProcessTime has been added. By default the value is -1 and it means Aufit can run as much time as it wants, batchAufit will wait infinitely for completion. If a positive value is entered it indicates that BatchAufit will only wait for the entered number of seconds and if aufit is still running it will be killed. This value is obviously system dependent for example on my machine a 5 rotations image is usually processed in 20-30 seconds. But some complex image takes up to 90-110 sec therefore I set the MaxProcessTime to 180 and this ensure that analysis of the tree will continue even in case of Aufit failure. To help you, when BatchAufit finishes the analysis of a bunch of images it will give you the maximum and average elapsed time. You can use the max value with a margin as value for MaxProcessTime. You can leave this value to -1 to be sure all images will be analyzed but if Aufit fails it will block the analysis. Note that this value is automatically saved between runs when you exit. So you only need to set it once.

unpack everything in one dir and run.
Please report problem

BatchAufit.rar

[DrCoolZic]

New release V1.1.5671

Aufit V1.1.5671 - Major internal Change to get ready for IPF output. Visible to user is GAP data buffer in sector Data View

- Added buttons for Update: Allow to check for new version and display user’s manual. Now the update is not done automatically but on request. Click update to check if new version available and answer yes to update program
- Added button for Help: Click on the help button should open the Aufit user's manual distributed as part of the install. To work this requires to have a pdf reader installed. Please report if problems.
- Huge internal change: In the Data view the sector button now displays GAP buffers. The Gap fields contains the bytes located after the ID and the DATA fields. Aufit continue to read the bytes directly following the end of the ID/DATA fields without re-synchronization.
- Splice lines are now limited to flux above 4% in GAP only => red for inner and blue for post gap + info tip that specify location and value of out of band flux transition (read doc)
- Fix bug of wrong sync sequence detection in random flux.
- Invalid sync sequence protection is now detected.
- Major reformatting of the protection messages + added ISS

 BatchAufit V1.1.5671
- project is now ClickOnce application (means update capability from Web) and Aufit.exe is now automatically installed with BatchAufit.

The main purpose is lots of internal changes to get ready for IPF writing ...
So please feedback if problem founds

Aufit installation http://info-coach.fr/atari/software/pc- ... /setup.exe
BatchAufit installation http://info-coach.fr/atari/software/pc- ... /setup.exe

The current auto update of Aufit may work, but to be on the safe side use the above link to install new version.
programs can be started from links in the directory C:\Users\YOUR_LOGIN_NAME\AppData\Roaming\Microsoft\Windows\Start Menu\Programs\DrCoolZic\Aufit

[Maartau]

[Avanze]

Is there a separate portable version?

[Maartau]

Is there a separate portable version?

Also interested .

[DrCoolZic]

Unzip all three files in one directory
should work let me know if problem

[gothmog]

Aufit gets better and better, very good program.
A small bug report in the latest portable version above: when I click the ? button on the top right corner, Aufit crashes.

[DrCoolZic]

Thanks for the feedback - I guess it is because the program cant find Aufit.pdf
Next time I will try to do a proper portable version

copy the Aufit.pdf file (attached) to the same directory and try again.
If it still crashes create a Resources directory and copy the Aufit.pdf file in it.

[gothmog]

It fails too with the pdf in the same folder as the exe, but works fine with Aufit.pdf in the Resources subfolder.
Thanks!

[DrCoolZic]

Thanks to Gothmog http://dmweb.free.fr/ a bug has been found on STX images generated for Dungeon Master and Chaos Strike Back.

DM & CSB uses a very special way of generating fuzzy bytes: The fuzzy bytes uses flux transitions located in what I call an ambiguous area (please refer to my protection doc for more info http://info-coach.fr/atari/documents/_m ... ection.pdf ) that results in bit 7 of read bytes to change (the bytes are therefore read as $68 or $E8) more info on protection here http://dmweb.free.fr/?q=node/210

In order to correctly decode bits from flux transitions, the WD1772 need to adjust the sampling frequency (frequency adjustment) and the center of the sampling window (phase adjustment). This allow to compensate slow flux speed variation (mainly motor speed variation) as well as fast flux speed variation see http://info-coach.fr/atari/hardware/_fd ... 8002-1.pdf (http://info-coach.fr/atari/hardware/FD-Hard.php#links ). These adjustments are done by the DPLL. The phase adjustment of the DPLL used in Aufit was a bit too “aggressive” and with the triangle flux pattern used by DM & CSB this resulted in few bytes (usually 3 or 4) with bit 6 of read bytes changing. I have fine-tuned the phase compensation in Aufit and it seems that now the bytes are read correctly. More tests are underway with Gothmog (contact me if you ready to beta test).

As the technique used by DM & CSB is protected by the pattern 4,849,836 http://info-coach.fr/atari/documents/ge ... 849836.pdf this bug should only affect stx images for these two games. If you own DM or CSB can you check the Aufit’s problem (may or may not happen depending on your diskette). Load image, select the “Data view” tab, click the “Sectors” button, and scroll to sector 7. You should see after the sector content the “fuzzy bytes mask” (Sector 7 Contains Fuzzy bytes). In bold are the masks for bytes changed if you only see 80 then everything is fine (no bug) but if you see C0 or 40 please report.

With your help this bug will hopefully be fixed in the next release.

PS I have tested reading DM with HxC floppy emulator and the content of sector 7 seems to be correctly decoded. However the .stx file generated is incorrect (seems like HxC does not handle fuzzy bytes in STX?).

[DrCoolZic]

Ok it seems that the bug is fixed

It has been tested on images of 11 different floppy disks of DM & CSB (some of them not in so good condition) and now Aufit seems to be able to consistently detect correctly this kind of fuzzy bytes. In order to successfully detect you definitively need to have a pretty good DPLL to succeed in all situation . FYI the hard problem is not the frequency correction (clock s stable for DM/CSB) but the phase correction.

If you own an image of DM and/or CSB I would be interested to get it for test.
Please PM me or make it available.
Thanks

PS: I don't know when this version will be released because I am at the same time in middle of implementation of IPF format write. Meanwhile do not convert DM/CSB games

PS-PS: According to Gothmog the fuzzy bytes that you read in sector 7 must only have bit 7 changed between the different revolutions (this is consistent with the patent description). The protection is delayed and therefore it may take a long time before the game complains.
Any program interpreting at stream flux level this kind of images should be aware of this difficulty...

Thanks again to Gothmog for finding this problem. I have been playing with images converted with Aufit and never found the problem

[Jeff_HxC2001]

PS I have tested reading DM with HxC floppy emulator and the content of sector 7 seems to be correctly decoded. However the .stx file generated is incorrect (seems like HxC does not handle fuzzy bytes in STX?).

Y a qu'a demander

Here is the latest version with the STX fuzzy bytes export support :

http://hxc2001.com/download/floppy_driv ... t_beta.zip

The STX created from the Dungeon Master SCP dump work perfectly with Steem.

[JimDrew]

Nice job.

[Maartau]

PS I have tested reading DM with HxC floppy emulator and the content of sector 7 seems to be correctly decoded. However the .stx file generated is incorrect (seems like HxC does not handle fuzzy bytes in STX?).

Y a qu'a demander

Here is the latest version with the STX fuzzy bytes export support :

http://hxc2001.com/download/floppy_driv ... t_beta.zip

The STX created from the Dungeon Master SCP dump work perfectly with Steem.

Merci monsieur .

[DrCoolZic]

Thanks
I just got back home - so I will test tomorow

[DrCoolZic]

New release of Aufit v1.2 August 12, 2015

First many thanks to Gothmog and Brume for testing beta versions and very valuable feedback.

IPF version is not yet ready, but along the way Aufit internals have changed quite a bit and this was time to fix/improve the GUI as well as some subtle bugs. Here is a list of what has changed in Aufit 1.2
V1.2 August 12, 2015

• Definitively Fixed problem of racing conditions in Aufit batch mode
• Fixed bug for fuzzy bytes on Dungeon Master & Chaos Strike back (DPLL)
• Fixed bug in writing incorrectly fuzzy bytes in STX for truncated sectors
• Added Magnifier capability in Disk View
• Added Tooltips on most objects and charts of Aufit
• Fixed scale factor for disk snapshot
• Checking if passed file name exists otherwise exit application
• Fixed bug in file names
• All transparent windows in Aufit are now white
• Added detection of Dungeon Master type of fuzzy bytes
• Added Data / Time of build in the title bar
• Fuzzy bytes are now a combination of fuzzy bytes detected for all revolutions
• In Data View => Sectors detection & display of DM & CSB fuzzy bytes
• Fixed graphical bug on the tabs of the Aufit main window
• Information window renames to Event log
• Fixed bug of Rev number incorrectly reported in Event log
• Fixed bug of inner and post data gap bytes possibly filtered
• Changed some command line parameters (batch mode)
• Added save image type in batch mode (currently only stx)
• Out of band in track data content now checking for transitions over 8% in the 3 data bands
• Corrected bug in Post Gap data (incorrect length for SWS)
• Aufit.pdf is now located in the install directory.
• Aufit does not crash but report an error message if help file not found.
• Bad data in GAP are displayed in DarkOrange (Data View -> Sector command). This can be used to detect sector/track write splice
• Post GAP information displayed in tool-tip when hovering sector in data chart
• Fixed bug of incorrect message when canceling image write
• Fixed bug of image format detection/selection when loading new image

As you can see quite a lot of changes so it is probably time to release even if IPF not yet ready.
Please let me know what you think about the "magnifier" (in disk view) and the tooltips (everywhere)

If you are using the clickOnce (installed) version of Aufit and BatchAufit just click the "update button" and the program should proposed to install the new version. In all other cases (including portable version) please go to this page
http://info-coach.fr/atari/software/pc- ... /Aufit.php

Important: Please make sure you update Aufit and BatchAufit at the same time as the parameters passed have changed.