Hello everyone!
Here is my proposal draft for GSoC
As groundwork for both these projects have been laid already, I would like to propose the unification of these two projects (File system and OS processes library)
Many thanks to @rajkumardongre for his GSoC '23 proposal to the community, I liked it quite a lot in the way it was structured and this proposal is much similar to that of his.
I would like to apologize in advance if I have missed something very obvious and would like to ask for your feedback (of any kind) on the proposal.
(I was not sure of what to include in the Timeline part of the doc, since I don’t think a very concrete plan could be provided for this project)
Let’s get your matmul PR finished and merged into stdlib, it would be a strong contribution. Do let me know if you need help working on the last gemm wrapping part.
Very happy and excited to let everyone know that my proposal about File system and Process Management library (stdlib) has been accepted by the org for GSoC 2025
The official project page for this can be found directly at the link below
My mentors would be @hkvzjal and @fxm. Thank you @Fortran-lang for accepting my application and mentors for choosing to mentor me!
Very excited to work with the mentors and the many people that I will come across during the course of the project!
pretty sure it will be a great summer for me!
Thank you everyone!
I read your proposal (at least the version I downloaded at 9 may) and have a few remarks:
The title includes “process management” (also the project objectives) but your proposal does not include anything of that kind. That is okay, but then that should be made clear in the text.
It is not entirely clear to me what you mean by “expand_user” or “expand_var” - is that tilde expansion?
What purpose has “getfd”?
Some of the function/routine names you mention are a trifle too generic, or better, are not easily identified as dealing with file or the file system: link, exists, rename.
The title includes “process management” (also the project objectives) but your proposal does not include anything of that kind. That is okay, but then that should be made clear in the text.
Oh yes, I did notice that, what I had in mind was the changes the process management could undergo due to the changes in the filesystem stuff, like making pipes generalised through derived types. I should have mentioned that yes.
It is not entirely clear to me what you mean by “expand_user” or “expand_var” - is that tilde expansion?
by expand_user I meant the $USER on Linux/MacOS and %USERNAME% on Windows, and expand_var would expand any environment variable given as an argument to it. I honestly don’t remember why I kept the expand_var when there is the get_environment_variable in the standard, I am sorry for that.
What purpose has “getfd”?
I meant the file descriptor associated with a file on unix systems.
Some of the function/routine names you mention are a trifle too generic, or better, are not easily identified as dealing with file or the file system: link, exists, rename.
I see, yeah I think they could be appropriately prefixed to avoid such confusions, using fs_ maybe.
Overall in self reflection, I think It would have been a better proposal If I would’ve briefly explained all the procedures in some detail too , I could’ve done a better job in some other places as well (some pointed out by you).
, I could’ve done a better job in some other places as well (some pointed out by you).