Talk:PAT0 (File Format)

From Custom Mario Kart
Jump to navigation Jump to search

@Wiimm, I'm not sure about 6 sections here. I know that's the only way that our system works, but from what I've seen, the item at address 0x20 into the file is a length, and at 0x24 always seems to be 0. Only the first two pointers actually seem to be sections (though the next two also sort of work sometimes).

Chadderz 15:26, 28 June 2011 (CEST)

Some PAT Statistics

Here you can find different statistics:

  • A list with all header members and section sizes. The old statistic at this discussion page counts many more PAT hits, because the BRRES files of objects counted for every track.
  • 2 lists with all base data of section 0.
  • 3 lists with all string list data of section 0.
  • A directory with wpatt decoded examples of all 41 different PAT files found in MKW. 3 examples use the complex output, the other 38 the simplified.

Different PAT0 format

Today I have analysed PAT0 again and I think, section 0 is total different from the description. I will goon with the analysis in the next days and will then write down my results. But I will ask, if anyone can confirm or disagree the current description.

-- Wiimm (talk) 00:36, 9 April 2015 (UTC)

I disagree with the current description also. Just looking at my unfinished PAT0 editor for CTools, I seem to parse a BRRES Index Group at section #0 offset. This then points to entries which match the format described in the section 0 section on the page. I then parse the rest of the file by looking at NSTR and reading that many values from section 1 and section 3. A comment in my code says "Something is wrong with these; misinformation?" so obviously I found a problem when I wrote it. I resolve section 1 values to strings, but I don't do anything further with section 3 values. I ignore section 2.
Chadderz (talk) 05:37, 9 April 2015 (UTC)
For the moment I'm ready with section #0. Now (in some days?) I will write some analyses code that lists all PAT data found in MKW to compare the values and also to verify the new data structure. Then I will also compare the old description.
Wiimm (talk) 10:39, 9 April 2015 (UTC)
Ah, so that's why Wiimms SZS Tools have been destroying my PAT0s. :p --Torran (talk) 13:21, 9 April 2015 (UTC)
Only, if the data structure differs from the standard (hope so). That's a problem of creating from scratch. I had tested extraction and composition of BRRES files without faults, but when data is moved, a wrong string offset calculation can make it invalid.
Wiimm (talk) 19:12, 9 April 2015 (UTC)
My analysis is complete for the moment. Now I'm theoretical able to create a PAT file from scratch scanning a self made text file (like the KMP compiler).
Wiimm (talk) 23:08, 9 April 2015 (UTC)