Notepad-Tabstate-Buffer

Important

This project is being deprecated and rolled into Notepad State Library. No updates will be made to this project. All new notes, documentation, and code will be at the new repository.

These are my attempts to reverse engineer the Tabstate files for Notepad in Microsoft Windows 11.. These files are located at: %localappdata%\Packages\Microsoft.WindowsNotepad_8wekyb3d8bbwe\LocalState\TabState

There are different types of .bin files that appear to save the state of the various tabs. These tabs could be:

unsaved with text stored only in buffer
saved file with unsaved changes stored only in buffer
others?

For now, I will be focusing on getting a better understanding of the underlying structure for a new unsaved tab with text. I have not attacked the 0.bin and 1.bin files.

Please see my other repository for the Windowstate files.

Acknowledgements

NordGaren for the inspiration to take a look at this when I saw his tabstate-util
jlogsdon for lots of help
JustArion for lots of help

Overall Behavior

Opening Notepad with no existing tab(s) will create an empty "Untitled" tab with its associated bin.
Opening a file will create an associated bin.
Closing a specific tab within Notepad will delete the associated bin file(s).
Closing Notepad
- New tab with no changes will create an associated bin, 0.bin, and 1.bin.
- New tab with unsaved changes will create an associated bin.
  - On subsequent open/close of Notepad it will create an associated bin, 0.bin, and 1.bin. (Existing Tab Behavior)
- Opened file tab with no changes will create an associated bin
  - On subsequent open/close of Notepad it will create an associated bin, 0.bin, and 1.bin. (Existing Tab Behavior)
- Opened file tab with unsaved changes will create an associated bin.
  - On subsequent open/close of Notepad it will create an associated bin, 0.bin, and 1.bin. (Existing Tab Behavior)
- Existing tab fi viewed will create an associated bin, 0.bin, and 1.bin.
No bin files indicate having never opened Notepad or closed all tab(s) manually.

WORK IN PROGRESS

When/Why do 0.bin and 1.bin get created?
When/Why do 0.bin and 1.bin get deleted?
What causes a flush of the Unsaved Buffer?
The internal content of the bin file doesn't load until the tab is visible.
- We can see this behavior if you drag/drop multiple text files into notepad to open them.
Forensic correlation between sequence number of tabstate and windowstate
- If sequence number for windowstate is lower than the tabstate. Did someone delete something?

File Format

There appear to be two slightly different file formats for File Tabs and Unsaved Tabs.

File Tab

First 2 bytes are "NP"
3rd byte is unknown
- Possibly a NULL as a delimiter
- Maybe Sequence number (Stored as a uLEB128)(Always 00)
4th byte appears to be flag for saved file
Length of Filepath (Stored as a uLEB128)
Filepath as little-endian UTF-16
Length of original content (Stored as a uLEB128)
Encoding
- 0x01 ANSI
- 0x02 UTF16LE
- 0x03 UTF16BE
- 0x04 UTF8BOM
- 0x05 UTF8
Carriage Return Type
- 0x01 Windows(CRLF)
- 0x02 Macintosh (CR)
- 0x03 Unix(LF)
Timestamp stored as a uLEB128
SHA256 of the original saved file
Delimiter of 00 01?
Selection Start Index on save (Stored as a uLEB128)
- I don't think this will extend to the Unsaved tab as this seems to only show up on Save
Selection End Index on save (Stored as a uLEB128)
- I don't think this will extend to the Unsaved tab as this seems to only show up on Save
Delimiter? 01 00 00 00
- WordWrap
- Right to Left
- Show Unicode
- Unknown
Length of original content (Stored as a uLEB128)
Content
Unknown 1 byte
- Possibly a NULL as a delimiter
CRC 32 of all the previous bytes starting from the 3rd byte
Unsaved Buffer Chunks

Unsaved Tab

First 2 bytes are "NP"
3rd byte is unknown
- Possibly a NULL as a delimiter
- Maybe Sequence number (Stored as a uLEB128)(Always 00)
4th byte appears to be flag for unsaved tab
Always 01? Is this also length of Filepath like above?
Length of original content (Stored as a uLEB128)
Length of original content AGAIN (Stored as a uLEB128)
Delimiter? 01 00 00 00 (This doesn't change. Changes do appear in the 0.bin and 1.bin)
- WordWrap
- Right to Left
- Show Unicode
- Unknown
Length of original content AGAIN (Stored as a uLEB128)
Content
Unknown 1 byte
- Possibly a NULL as a delimiter
CRC 32 of all the previous bytes starting from the 3rd byte
Unsaved Buffer Chunks

0.bin / 1.bin

First 2 bytes are "NP"
Sequence number (Stored as a uLEB128)
4th number of bytes until the CRC at the end (What is this stored as? uLEB128? int? )
Delimiter? 0x00
Size of the associated BIN File in bytes (Stored as a uLEB128)
Selection Start Index on close (Stored as a uLEB128)
Selection End Index on close (Stored as a uLEB128)
Delimiter? 01 00 00 00
- WordWrap
- Right to Left
- Show Unicode
- Unknown
CRC 32 of all the previous bytes starting from the 3rd byte

Chunk Format for Unsaved Buffer

[Cursor Position][Deletion][Addition][CRC32]

Cursor position (Stored as a uLEB128)
Deletion Action (Stored as a uLEB128 indicating how many characters to delete)
Addition Action (Stored as a uLEB128 indicating how many characters to add)
- Added characters are stored as little-endian UTF-16
CRC 32 of the previous bytes

Addition Chunk

Below is an example of a chunk of bytes that represent the addition of the character 'a' at position 0.

00 - unsigned LEB128 denoting position of 0
00 - unsigned LEB128 denoting number of characters deleted
01 - unsigned LEB128 denoting number of characters added (In this case 1)
61 00 - character 'a'
BB 06 C7 CC - CRC 32 of previous bytes

Below is an example of a chunk of bytes that represent the addition of the character 'a' at position 17018.

FA 84 01 - unsigned LEB128 denoting position of 17018
00 - unsigned LEB128 denoting number of characters deleted
01 - unsigned LEB128 denoting number of characters added (In this case 1)
61 00 - character 'a'
98 07 F5 46 - CRC 32 of previous bytes

Deletion Chunk

Below is an example of a chunk of bytes that represent deletion at a position 1.

01 - unsigned LEB128 denoting position of 1
01 - unsigned LEB128 denoting number of characters deleted (In this case 1)
00 - unsigned LEB128 denoting number of characters added
E7 98 82 64 - CRC 32 of previous bytes

Insertion Chunk

Insertion chunk is a combination of the addition and deletion. This would occur if someone was to select text and paste new text into place. Below is an example of a chunk of bytes that represent pasting 3 character 'b's over 3 character 'a's starting at position 1.

01 - unsigned LEB128 denoting position 1
03 - unsigned LEB128 denoting the number of characters deleted (In this case 3)
03 - unsigned LEB128 denoting the number of characters added
62 00 - character 'b'
62 00 - character 'b'
62 00 - character 'b'
CD CD 85 6F - CRC 32 of previous bytes