-
Notifications
You must be signed in to change notification settings - Fork 157
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
UUIDs instead of id database #102
Comments
Just so I get it:
I really like the idea of a root UUID, which may even be different for each release, thus allowing a program to verify that they have loaded the correct version of a VSS spec. That said, I have two hesitations about UUIDs:
That said (again), I would be happy to update the tooling to support this, and retire the original int-based ID file. My $0.02. |
Yes, absolutely agreed. I got sometimes the complaint, that strings with varying length might not be so easy to handle in certain environments and that a fixed size unique identifier would help. I didn't want to replace the pathname as a primary identifier but just add another option. |
Sounds good to me. |
@magnusfeuer, sorry I've overseen your comment. That would be fantastic! |
Will do. Give me a week or so. |
@danielwilms Starting on this now. I will remove Signal IDs since they can be implicitly created on both pub and sub side by:
|
@danielwilms The generated CSV, fidl, and JSON files looks ok at a cursory glance, with UUID values having replaced integer IDs everywhere. I cannot speak for the cnative format since I haven't read up on that particular subsystem. I will also add a vspec2hdr script that generates a C header file with all the signal specs. I may also add a vspec2py that generates python code with the signal spec as wll. |
Added vspec2c.py that generates hdr and C file that encodes the entire signal spec and provides access functions. Demo code that exercises the auto-generated code: At this point I am ready to merge into master |
|
Merged to master. |
This is now fixed in the PR#108 "C-native tool updated with UUID support. " |
Thank you, @UlfBj . Merged. |
Considered to be a duplicate of uint8[] Has previously not been supported by tooling so removing it from documentation shall not cause any backwards compatibility issues. See COVESA#102 and COVESA#103
As the process of generating the id database is not so transparent and fail proof, I would suggest to replace them with uuids, to have them truly unique and having a fixed length. To do so and not having an arbitrary string length as input the namespace system of uuid v5 could be used:
That can be done recursively to the leafs. I think it offers some advantages over the old system as mentioned above. Any opinion on this?
The text was updated successfully, but these errors were encountered: