MTR 6 ANNOUNCEMENTS - READ FIRST!

[geezerbuttz]

Good afternoon.

I completed the lockingRegister array this afternoon.
Took a bit as I rewrote it a few times to get the optimisation down.

The data is now saved with the IMF file..

I will commence on the final row of tests again. phew.
If that all checks out , I can get into the main feature code , and get this incorporated..

I will , of course , get that updated video ( with chatty audio) done as well tomorrow.

[geezerbuttz]

Fresh from my desk....

https://www.dropbox.com/s/ir4d5vglq3kpjdh/Screen%20Recording%202023-06-07%20at%2012.49.16.mkv?raw1

[geezerbuttz]

well , that was quick , in my testing a ran into a new issue.
no next vibe , and a Vobu with zero length...

That's patched... working with this now...
I guessing though it a black area patched by MTR.
Seems it blanked an area that should be played... ( older version )

[geezerbuttz]

Still finding a few odd disks that behave badly in the probe.
Updated to correct for unmarked infinite still frames.

Some HouseOfMouse disks have a huge amount of void sectors.
They were taking far to long to generate and populate the list at the end.

rewrote that code to be almost instant. less than 1 second verses 4seconds...

going through some cray ISO disks now...

Looking good.

[geezerbuttz]

spent a long time last night and this morning , running a very expendable disk.
It was reporting blocks containing no correct navigation data, and in that process , putting probe into a loop.

On closer inspection , it seems that the early versions of MTR ISO engine , wrote blank packs with no SRI table set up.
This stops anything that relies on the SRI pointer table.

I fixed this , meaning we can now scan older ISO with a high chance to get a correct output.
There is nothing better than an original disk though.

Onwards and upwards...

[geezerbuttz]

All ISO images and disks I have have passed...

Now would be the time for any disks that are new on the block , to be imaged ( any method will do ) and for them to be delivered daily for testing .

Just to confirm its a go on the probe...

I will move into MF tomorrow

[TheSystemHasFailed]

Fresh from my desk....

https://www.dropbox.com/s/ir4d5vglq3kpjdh/Screen%20Recording%202023-06-07%20at%2012.49.16.mkv?raw1

oh man.. deleted before I could see it..

[geezerbuttz]

ok , who deleted it to make some space.....

[geezerbuttz]

A very busy productive day:

Started to get the hooks for the IMF in Batch.

Removed the restrictions in Batch so titles are not missed.
Centralised the existing IMF loader to function the same between both Extraction engines.
IMF Panel is now Modal...

In process of renovating existing IS menu handler , to be run-times based , same as batch.
This will remove the buttons from the UI , instead using contextual menus for IMF handling.

Tomorrow I have to code the Batch decode ( IMF ) block to pass to the IMF loader...
Fun Fun Fun...

Nice to be back in MTR.

[geezerbuttz]

A slew of changes addition and enhancements..

1) I got the IMF loader , and decoder in batch.
currently we are only stripping the unused titles , rebuilding the table etc....
cell selection tomorrow...

2) IMF loader got a further update , on successful loading the panel dismisses itself now. ( still have manual cancel if you change your mind )

3) moved all IMF/Reset controls to contextual menu to ease UI clutter.

4) added a reset to Batch to restore the original title layout if you need to.

that's all .. cells tomorrow.
Have a good evening / day!

[geezerbuttz]

That's the end of today.
Cells did not get done as I needed to generate a complex structure that is needed for the DVDSectorProbe within MTR's batch engine.

ISO already does this , but as Sector probe was only designed to be launched from Iso , I needed to add this to get it 100% integrated.

Spent a few hours upgrading some UI elements ..
I could not believe how stupid some of the code was... I must of been pissed at the time.....

That's all done.

Added a path addition in Batch to indicate the currently selected Disk , its selectable and will open the disk for you on the desktop when selected.
Its a shortcut.

Ok , that's me.

Have a nice day / night.

[geezerbuttz]

I have to redesign the DVDSector Launch code:

At the moment its available in every extraction engine.
This is a flawed logic as , if the engines fail to load , you cannot make a map!


I need to move the whole launcher into its own class , and call it from the disk menu.
A single location to remember...


More soon when done.

[geezerbuttz]

right this all done:

Been polishing for the day. Found some classes that were redundant, I removed them.
Found a couple that I had not modernised to ARC.. Done.

Optimised the Decryption class ( its very old lol , its drawing it pension) moved a lot of internal call into blocks...

Removed the Google Kqueue implementation , and replaced it with a modern GCD implementation. ( its far smaller )

Tomorrow I want to modernise a small piece of UI code that basically fades a few panels in and out.
Once again , its very old...

Then Cells , and testing......

Close...

[geezerbuttz]

Updated the About Window handler.
Removed all the rubbish in there for a slimline approach.

Started work on the new Icons and Logo..


Anybody have any ideas, drop me a line.

[geezerbuttz]

Right , that's the new Icon and Logo done.
took longer than normal . Im rusty in the graphics department.

That leave this next week to double check the interleave cells code.
Im a little concerned that the output of Probe does not transfer well to MTR's cell definition.

I may need to do some sector matching to make sure.

Interleave cells are a nuisance.

[geezerbuttz]

A hard day.

An old bug that a few have noticed and raised , revealed itself to me.
On the title/Batch , sometime titles were missing.
I though I had cured this early last month.

No.

I was loading cell data from the probe , and I could not get a proper match.
Well , it seems the title I needed was never assessed as good in the main app.


6 hours to find the mistake , .. but i found it.

I now have to make a UI addition in batch to assist you in your choice of main feature.

In batch , even when a probe map is loaded , you may see more than one 'main title'.
The times will be similar and they will confuse you.

The only way is to look to see how many and what cells have been accessed.
To do this you need to expose the main panel of each title.

That is cumbersome . Tomorrow I will code a UI addition in Batch that exposed the current count of selected cells,
so you can see at a glance how many cells are actually active in a title..


That's me for a very long day.
More tomorrow.
G

[geezerbuttz]

Well , I had to make a compromise.

Because MF design allows for the same title to be in the extraction queue multiple times , it made no sense to place the 'active cell count' in the title panel.

This would never represent the correct value , unless you only have one entry for that title.

So the count is on each panel , visible for a quick check up..
It just takes a few minutes.

I may code a routine ( menu based ) that automatically places the title with the longest time and max cell count in the table for you.


More soon.

[geezerbuttz]

Hello All..

Well, bar the user guides , that I will do this weekend, it is done.

The MF code reads the IMF file and remaps its cells correctly.
Angle data , and interleave cells are handled correctly.

I have to run all the disks I have here through once again in Probe and MTR to generate a full process flow.
Best to make sure the end product is viable . ( triple check )

Sure there will be disks that probe will have an issue on , but as long as I can get an Iso , I can get an update out very quickly.
Probe will be hosted locally , and downloadable(?) from the forum.

Currently running the nasty japanese mermaid movie , and its goes very well.
It can take along time to probe these titles , but once again , only one person needs to probe the disk and upload it for everybody to have the solution. ( by a long time , I refer to about one hour )

Once I get this out the door , I will have a small break while I try to get some funding to begin the RnR replacement.
A little break will do me some good.

I will link the user guides directly in the forum , aswell as in the app.

G

[geezerbuttz]

Been running through all disks , no issues.

I have although , realised that it is possible for a naughty person to 'switch' IMF filenames on the documents.
This will cause a mismatch of the structure to be loaded , possibly even crashing the app.

The solution is too embed a copy of the filename within the IMF file before it is encrypted , and compare the embedded name with the current name before we accept it.


Will implement that tomorrow..

Needs to be done an a few places..

G

[geezerbuttz]

ok , that's done , a quick change , although I have to regenerate all the IMF files I have done....

This makes sure that if the file is manipulated , there is no chance it can be loaded into memory.


Back to testing / user guide tomorrow.

[geezerbuttz]

I found that the internal update mechanism was not checking for updates..
It would not hold its value.

Reworked the whole of the preferences to simplify them , and to make the more efficient.

That all works now.

Added a contextual menu to the preference paths so you can reveal the folders.

Back to scripting the User guides.

G

[geezerbuttz]

picked up a troublesome pair yesterday.

Failed in MTR 5.

MTR 6 is fine , found a couple of issues in MF with correct audio sync.
Fixed that.

Probe was missing some titles , so I optimised that not to do it ..
Titles that shared sectors already probed were being ignored.
Fixed.

Back to testing now , and building maps..

Sorry this last part is taking time...
G

[geezerbuttz]

Being me I had to complete the title preview data that I failed to do last week.
That's done.

We get a dynamic menu for each title that gives you an instant overview of the title simply by left clicking.

Been running tests , and found no subs are muxed into the MKV output????
Testing and investigating now.

G

[geezerbuttz]

Busy day.

created a new helper app for the lowly mods out there...

The IMF map system requires all map titles to be entered to a plist that is held online.
As you can imaging this could become tiring very quickly.
Not only that , it can be prone to human error.



The OS only allows you to mount ftp(s) as read only on the desktop , you cannot write to it.
So that ruled out using the OS functionality.

So....
I created a new helper app that can log into the server (sftp) , audit the IMF folder , and generate the plist for us.
All password and user credentials are stored in the keychain.

so all that needs to be done now is to run the app after you dump more IMF maps into the correct folder.

MTR:

I removed the auto save runtime of the Iso IMF files , as they contain less data than a probe IMF.
Implemented the auto load of remote IMF maps into batch mode.

im running out of things to do...
Beginning to get satisfied.
G

[geezerbuttz]

A quick video for the MODS.
Its really hurried , so I make my excuses early

https://www.dropbox.com/s/kgh28siw27j6k ... t.mp4?dl=0

[geezerbuttz]

Well , the last jobs are always the longest:

I took a peek inside the compiled app , disassembled it.
It exposes too much info, so Im obfuscating right now...

Its a complicated task , never thought to much about it in the past.
ObjC exposes so much information , no wonder stuff gets hacked all the time.

Found a few good methods that only get applied to release code , so ..
Im on it.

Stripping symbols is just not enough.

G

[geezerbuttz]

Completed the security work , rewrote the class methods as functions so the compiler strips the function names them for me.
Obfuscated the class names at build ( release ) time.
Encrypted all important strings within the app. ( release build only )

fixing a few issues in in probe , just a small speed increase.

Running final tests ( again , have to prove the updates )

Then I get the long awaited User guides done.

No more posts here , as nobody seems to read them any more.


G

[Footloose]

Completed the security work , rewrote the class methods as functions so the compiler strips the function names them for me.
Obfuscated the class names at build ( release ) time.
Encrypted all important strings within the app. ( release build only )

fixing a few issues in in probe , just a small speed increase.

Running final tests ( again , have to prove the updates )

Then I get the long awaited User guides done.

No more posts here , as nobody seems to read them any more.


G

I read them

[TheSystemHasFailed]

Completed the security work , rewrote the class methods as functions so the compiler strips the function names them for me.
Obfuscated the class names at build ( release ) time.
Encrypted all important strings within the app. ( release build only )

fixing a few issues in in probe , just a small speed increase.

Running final tests ( again , have to prove the updates )

Then I get the long awaited User guides done.

No more posts here , as nobody seems to read them any more.


G

Wait! I totally read them! I was just on Holiday in Arizona!!

[geezerbuttz]

Small update.

Found that probe does not read the ISO correctly regarding Bups.
They were not being found when they are padded to an even sector length with void packs.

This is because MTR populates all void packs with blank frame information.
Probe does not expect that , and regards a MPEG header as an indication that the BUP/IFO is false.

Altered the code to nibble to find the start of the bup.
It works , and is more likely to find the start of the bup with less chance of miss identification.

Changed some code that was making a few assumptions .
Side effect is its now far stricter about how it reads pack data and how it interprets the navigation.
Side effect , its slower.

Need to go into MTR to match the ISO tables look and feel to match the probe table look and feel...
Just a few colour cosmetics in some text items....
Nothing major..

It just about polish now....

Accuracy is more important here.