You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Feb 18, 2025. It is now read-only.
Repository navigation
This repository was archived by the owner on Feb 18, 2025. It is now read-only.
Add better descriptions as to what causes None to be returned. #44
Currently the documentation is vague about what causes None returns in functions. It would be very helpful to be able to not have to comb through the code to implement the api calls and know what is happening if None is returned.
Yeah I think that is a better solution that the current Some/None strategy. Having some way to understand what failed to happen is better than it failing with no context. I am not entirely sure what the original context was of this issue, let me review the code I used that faced this issue and get back to you.
The main concern when I thought about this pre-1.0 was that the errors that may show up in the failure case are hard to predict/test/standardize across platforms.
Activity
opeolluwa commentedon Apr 1, 2023
Regards this, could you provide more information on the nature of what is to be returned
opeolluwa commentedon Apr 1, 2023
If you indulge me, I could return a struct, which will contain the PathBuf if any, and an error message if none
soc commentedon Apr 1, 2023
Like a
Result?opeolluwa commentedon Apr 1, 2023
A result that returns the Ok(Pathbuf) or an Err(Strong)
Peter-Lande commentedon Apr 1, 2023
Yeah I think that is a better solution that the current Some/None strategy. Having some way to understand what failed to happen is better than it failing with no context. I am not entirely sure what the original context was of this issue, let me review the code I used that faced this issue and get back to you.
opeolluwa commentedon Apr 1, 2023
I looked it up, I think it's coming from the dir-sys crate
soc commentedon Apr 1, 2023
The main concern when I thought about this pre-1.0 was that the errors that may show up in the failure case are hard to predict/test/standardize across platforms.
opeolluwa commentedon Apr 1, 2023
Hmmm 🤔
opeolluwa commentedon Apr 1, 2023
Now, the challenge is getting the error cause, which would be better than a generic message I suggested earlier