fix(#321): refactor add_to_path + sep. OS logic#322
Conversation
8c232d2 to
1f5d85e
Compare
|
PR overall looks good, but i am not sure automatically modifying in non interactive shells, should we? |
1f5d85e to
d614ef2
Compare
|
@MordechaiHadad I'm from the standpoint that if a User is wanting to install something, their expectation would be that it gets added to path and becomes usable straight away. (See - Principle of least astonishment ) With how you've decided to do it too, via a self-contained shim for each major shell ( I also would make the (reasonably) safe assumption that most people doing fully non-interactive setup's are likely in the realm of devops, fresh-setup and are familiar with the tool already or at bare min. are SSH'ing which; atleast for 2 of them- require a reasonable level of comfort or competency with shell commands/tooling. |
b44099a to
01a540b
Compare
|
Changes in those recent commits are:
|
dd2a644 to
e328788
Compare
|
The tracing update/security vulnerability patch should also unblock #320 as well |
Refactor add_to_path for handling non-interactive shells and separate OS logic to compile time guarded functions.
Deps update addresses security vulnerabilities. Relevant link: [RUSTSEC-2025-0055](https://rustsec.org/advisories/RUSTSEC-2025-0055)
Use smaller functions for readability and Err recovery.
e328788 to
9b00158
Compare
|
This has been rebased over master |
Summary
Closes #321
Refactor add_to_path for handling non-interactive shells and separate OS logic to compile time
guarded functions.
Changes are mostly the addition of a guard check if we're interactive on either stdout/stderr, wrapping prompt with tokio::time::timeout, and moving the temp config to use a RefCell over cloning.
Testing
Currently requires end-user testing, will have to further refactor for mod_tests and proper testing between unix/windows OS's.