fix: symlink issue bundling for linux #5781

[passivedragon]

What kind of change does this PR introduce?

• Bugfix
• Feature
• Docs
• New Binding issue #___
• Code style update
• Refactor
• Build-related changes
• Other, please describe:

Does this PR introduce a breaking change?

• Yes, and the changes were approved in issue #___
• No

Checklist

• When resolving issues, they are referenced in the PR's title (e.g fix: remove a typo, closes #___, #___)
• A change file is added if any packages will require a version bump due to this PR per the instructions in the readme.
• I have added a convincing reason for adding this feature, if necessary

Other information

Frankly, this fix is more from @ca0v who suggested it in the issue thread #5781, rather than me just putting it in.

I have test run this on a single project with success, but I haven't tested beyond that.

About the issue:
The bundling script fails if it has run once before, as the links it creates already exist, and it does not like that. The added -f overwrites the existing links with a new one, which should be the desired behavior.

[FabianLars]

a bit weird that this seems to be required since the AppDir is supposed to be cleaned out before that 🤔

[passivedragon]

Hmm, that definitely doesn't happen.

line 19 is the first time the directory is mentioned, and all it does is create it if it doesn't exist already
mkdir -p "{{app_name}}.AppDir"

it might need rm -r "{{app_name}}.AppDir/*" in the line directly after that and before the copy instruction, but from what I've seen from testing, that doesn't seem to be necessary.

[FabianLars]

I was talking about the rust code that runs before the bash script is invoked: https://github.com/tauri-apps/tauri/blob/dev/tooling/bundler/src/bundle/linux/appimage.rs#L38

[passivedragon]

That's pretty weird, if that failed, it should error and fail then and there to my understanding.
Wouldn't it need to error on bad input? From a quick, shallow read over rust docs seems to me like it might be happier with output_path.clone() instead of &output_path. I'm a rust noob though.

With even just this change for me though, it manages to build all files for me, all the way to the .sig file.

[FabianLars]

clone vs & won't make a difference but you're right that it should fail if it can't remove the folder, at least the way i understood the docs.

To be clear, i'm fine with merging this PR but i'm irritated that it's needed in the first place (something we probably should look into as i'd imagine that the not deleted folder could cause issues in other places too)

[FabianLars]

Oh btw, we'd need a change file similar to https://github.com/tauri-apps/tauri/blob/dev/.changes/bundler-msi-mini-version.md

[passivedragon]

That work?

[FabianLars]

yep, works for me thanks

[lucasfernog]

I think the fix doesn't make much sense and the problem might be somewhere else, but I'll take it if it works.