scanmountgoat / arcexplorer Goto Github PK

This can probably be accomplished by using the Github API to check the latest commit to the archive hashes repo that modified the hashes file and comparing dates with the local file. The hashes can be downloaded directly rather than using Github's API to avoid any file size restrictions or rate limits.

The application can prompt the user to update hashes and display the commit message.

• Check if hashes can be updated
• Download hashes file from github
• Prompt user to update and display commit details
• Make the process asynchronous to avoid locking up the UI
• Use the file modified date for the commit or find a way to ensure dates are consistent
• Log latest commit and if update succeeded or not

This affects the search list and shared file list. There's currently no way to go the selected file or folder.

• search textbox
• add magnifying glass icon similar to vscode or file explorer instead of a "search" label
• swap icons when application theme changes
• fix overlap with textbox and icon

I recently downloaded Arc Explorer, but I noticed that it's lacking some features that are available in CrossArc.

1. It's not possible to update the hashes. That's going to be important when new software updates get released.
2. There's no option to change the language. When extracting language files from Arc Explorer, you're restricted to only the North American English files.

This should be a separate resource dictionary.

This would aim to take away the hassle of typing out the IP every time you want to connect to the arc over arc-network.

1 byte for the length and 4 bytes for the hash itself.

Crashes when trying to open file dialogue to select the data.arc, or extraction default folder. log.txt is empty.

fish: './ArcExplorer' terminated by signal SIGSEGV (Address boundary error)

dotnet --list-runtimes shows Microsoft.NETCore.App 3.1.8 [/usr/share/dotnet/shared/Microsoft.NETCore.App] as an available runtime.

Some of the shortcuts for Windows Explorer are used in Blender's file browser as well.
https://shortcutworld.com/Windows/win/Windows-10-File-Explorer-Shortcuts

• enter to go down a level
• alt + up arrow to go up a level
• left/right arrow to go up or down a level

Arc > Extract All. Finishing #16 first will make the progress indicator more meaningful, since this operation will likely take a very long time on most systems.

This should check the version number to see if it is compatible. The current behavior is to just return a boolean, which isn't very descriptive.

This should avoid any issues with API rate limits. The ApplicationSettings.json could have something like the following.

current_commit: { "sha": ".....", "last_checked": "1/1/2021 ..." }

Steps to reproduce:

1. Open data.arc
2. enter fighter/mario/model/body/c00/ to navigate to the folder
3. click the file list to finish displaying remaining items

This doesn't happen when double clicking the c00 folder. This seems to only happen when the listbox virtualization is set to "Simple". Setting the virtualization to "None" works but causes unacceptable performance when viewing large folders.

The behavior in CrossArc is to update the offset when changing region. This may or may not be a good idea to incorporate into ArcExplorer directly.

In theory, this should improve performance and memory usage on startup.

Directories with missing hash labels aren't reported by smash-arc/SmashArcNet

It may be more convenient to toggle certain settings on/off in preferences to avoid cluttering the details view.

Hashes:

• Path Hash
• Ext Hash
• Parent Hash
• File Name Hash

Metadata:

• Is Stream
• Is Shared
• Is Redirect
• Is Regional
• Is Localized
• Is Compressed

The current implementation is working but has some stability issues. This is likely some sort of marshaling bug in SmashArcNet.

Currently, ArcInstaller only can inject files with the same size or less than the original file. While ARCropolis is a great solution on console, it (at the current moment of writing this!) is not supported on either Ryujinx or Yuzu. Full on modding has been one of the goals from the outset of the Yuzu community and this could provide a solution in that field.

On faster systems this won't be noticeable, but large files may take a long time on slower drives. It should be an easy change to set the progress bar to indeterminate on single file extractions.

Improving the compatibility for windows and linux binaries will require building through github actions. This avoids issues with native dependencies using newer Rust runtime components like glibc that may not work with some operating systems.

• SmashArcNet submodule
• Properly copy dll/so files to build directory for ArcExplorer
• Build windows with github actions
• Build linux with github actions
• Build macos with github actions

The divider between the file details and file list is quite small and hard to select with a mouse.

render/shader and render/shader/ (trailing slash) are considered different directories. It doesn't make sense to handle this in smash-arc or SmashArcNet since the entries are distinct. It should be doable to handle this automatically from ArcExplorer itself.

• When typing a path like "a/b/c", load the nodes for both "a/b/c" and "a/b/c/" (trailing slash)
• When loading child nodes, strip trailing slashes from directories and remove duplicates by path
• When extracting the folder "a/b/c", extract the nodes from "a/b/c" and "a/b/c/" (trailing slash)
• Add an option in preferences to keep them separate (should be disabled by default)

The missing hash is causing none of the files to appear in the treeview. This is probably related to #8

There is a number of redundant or incomplete methods for creating nodes that should be cleaned up. The overhead of having a separate FileGridItem class for the UI should be avoided if possible.

Use a list of the nodes in the current folder rather than managing a treeview. This matches the look and feel of traditional file browsers. The grid layout makes it easier to associate additional info next to each node. This also allows for virtualization support, which will fix #9.

• Change UI to use a list instead of tree view
• Exit from the current folder
• Double click to enter a folder
• #30
• Correctly find the parent folder for manually entered paths
• #32

I've been having alot of issues with opening the data.arc file in ArcExplorer which gives me errors upon attempting to open it. I'm very confused on what's going on.

Everytime I try to merge the base game with the update file it always gives me the filesize as seen above no matter what I try. If anyone can give me advice on what to do then I would be VERY grateful. Thanks.

• update Rust lib (jam1garner/smash-arc#3)
• add regional indicator icon to regional files
• update file details view (#2)
• specify a region when extracting files
• add default region to user preferences

This should contain absolute path and most of the values from the details panel. This also allows sorting.

Errors message, it won't even view data.arc

The number of bytes can be broken up into kilobytes, megabytes, etc. The exact size should still be shown in bytes.

smash-arc already has functions for getting all file entries, so it's unnecessary to use the slower and more error prone path manipulation.

This will replace the current binding with ReactiveUI. This will require investigating how to make the search threadsafe.

Performance is consistently poor when trying to expand sound\bank\fighter_voice.

• #42
• Update Rust Lib
• Search by path
• add a way to cancel the search or clear search results
• #43

This will likely be easier with a data grid control. The UI should be similar to the "details" view in file explorer or other vertical list layouts for file browsers. The additional information could be editable from preferences or on a top bar on the main window.

• add details to file grid
• fix navigation when sorting

I've figured out that when I copy my Hashes.txt from CrossArc and paste it into ArcExplorer's directory, it works fine. But next time I open it, it'll say that it cannot read Hashes.txt. This issue can be fixed by doing the same process by copying from CrossArc but doing this every time is very impractical.

Text or image previews can be displayed in the details panel. The information can go under the details panel or replace it if more space is required.

• C# wrapper for Rust library (create project in this repo)
• display text (should be copyable but not editable)

The generated export path is invalid because of the ':' character.

Info that isn't part of the ARC and will have to be calculated.

• Compressed Size (all files and subfolders)
• Decompressed Size (all files and subfolders)
• File Count (files and subfolders)

This might simplify some of the application code and reduce the number of casts on Data context.

A setting under "Extract Location:" that holds a path to a given data.arc which automatically opens when the program would be a nice addition to this program.
Maybe add a warning, and possible reset of that path, upon opening the program if the file is moved or no longer available would be good if someone accidentally moves it.

Thank you for making this program.

• #13
• convert unsupported directory separators to a single /

Extracting files with uncracked hashes will place them on the root with the directory added to the hash in the file name instead of in their proper folder.