🧐 Research a new solution for file management
@FastestMolasses 已经在做这个了。
开始于 2024年3月23日。
评估
这个 Issue 还没有评估数据。
描述
[!IMPORTANT]
This issue should only be picked up by CodeEdit maintainers
As a CE developer
I want a reliable foundation that handles and manages files
Because it is crucial to other functionality of the application
Our current solution for file managements is reaching its limits and there are some serious bugs present. There has recently been an attempt at writing tests for this part of the code, but it revealed that there is unwanted functionality and results.
Acceptance criteria
- There has been a thorough investigation on how file management for CodeEdit can be done most effectively
- The findings have been documented
Scope
In scope
- Research options on how load a directory recursively as workspace
- Document the findings of the research (Can be a markdown file)
- Optionally make PoCs/Demos in a separate project
Out of scope
- Implement anything of the research done in CodeEdit
Requirements (Might need further refinement)
- The system should be able to load in a directory recursively from disk
- It has to be done asynchronously, so no problems occur when a folder is loaded with large amounts of content
- It should be able to monitor changes in all the files on disk inside that directory, this doesn't include the contents of the file
- It should be able to add a files and folders to disk
Risks
- Because
CEWorkspaceFileManagerandCEWorkspaceFileare referenced at so many places in the codebase, truly refactoring the code might prove difficult. Especially since Xcode's code analysis is plain bad.
Implementation design
Problems with current implementation
There are likely multiple ways to tackle this problem, in terms of using different Apple/Swift APIs and how we design the system ourselves. The latter is the most important and something that we can control. The current setup of CEWorkspaceFileManager and CEWorkspaceFile is not working properly. There are multiple cases where an "IllegalState" is achievable.
One of the reasons this occurs is because currently the CEWorkspaceFile contains all kinds of information, it contains multiple variables & functions related to Git, some that have todo with UI and some that have to do with tab related stuff (Which arguably is also UI). It's not clear what the responsibility of the class is. It doesn't look like the class isn't following the single responsibility principle anymore. Everything that is somewhat related to it, gets "dumped" in there.
It also has some properties that in theory have something to do with it, but from a system design point of view, they don't. An example is fileDocument, from the type CodeFileDocument. You can argue that it should be there because it represents the contents of the file and is therefore related to it. But it actually has more to do with the editor and truly modifying the contents of file. And CEWorkspaceFile represents the file in things like the project navigator, where the contents of the files don't matter.
Another thing that can (and should) be looked at is abstraction. The current CEWorkspaceFileManager and CEWorkspaceFile is referenced everywhere, and is therefore hard coupled to almost everything is the codebase. This also makes it more difficult to make unit tests, as the FileManager can not be mocked or stubbed. It also is a disadvantage in cases like these, be it requires changing all the files/components that rely on it.
Things that are interesting to look into when redesigning
Right now there is one CEWorkspaceFile, which can be either a Folder or a File. Perhaps this is the right way to do and it is how we want it, but another possibility is to make a distinction between the two by making separate classes.
If PoCs are made, it is advised to develop them test driven. Since this is core functionality of the application, we should have tests in place from the get go. This way we know how we want it to behave, before it has been coded.
Tasks
- Do research for one or multiple solutions (Multiple people can work on this, but should be in contact with each other to prevent duplicate work)
- Document the findings
- Discuss findings and PoC's with the team
- 主要语言
- Swift
- 星标
- 23k
- 派生
- 1.2k
- PR 合并指标
- 30 天内没有已合并 PR
环境准备
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
CodeEditApp/CodeEdit 的其他 Issue
-
难度 2/5 1-3 小时 新手友好度 78/100
CodeEditApp/CodeEdit#2190 ·
-
难度 1/5 1 小时以内 新手友好度 15/100
CodeEditApp/CodeEdit#2195 · 10 个 reaction ·
-
enhancement
难度 4/5 3-5 天 新手友好度 35/100
CodeEditApp/CodeEdit#2185 ·
-
难度 5/5 一周以上 新手友好度 25/100
CodeEditApp/CodeEdit#2181 ·
-
bug
难度 4/5 3-5 天 新手友好度 35/100
CodeEditApp/CodeEdit#2180 ·
查看 CodeEditApp/CodeEdit 的全部 Issue
相似的 Issue
-
auth type: bug
难度 2/5 1-3 小时 新手友好度 90/100
googleapis/google-cloud-swift#1260 ·
维护者通常 1 天内回复
-
catalog-audit
难度 2/5 1-3 小时 新手友好度 68/100
维护者通常 1 天内回复
-
0. Needs triage bug
难度 2/5 1-3 小时 新手友好度 78/100
维护者通常 1 天内回复
-
clawsweeper:linked-pr-open clawsweeper:no-new-fix-pr clawsweeper:source-repro impact:ux-friction issue-rating: 🦞 diamond lobster P2
难度 2/5 1-3 小时 新手友好度 86/100
steipete/RepoBar#140 · 1 条评论 · 1 个 reaction ·
维护者通常 1 天内回复
-
area:dictation bug P2
难度 2/5 1-3 小时 新手友好度 88/100
uttrflow/uttrflow-swift#2551 ·
维护者通常 1 天内回复