Re: [yocto] Dynamically load patch series from fetched repositories.
Alexander Kanavin <[email protected]>
| Newsgroups | org.yoctoproject.lists.yocto |
|---|---|
| Message-ID | <CANNYZj_T6SeZTkh=fgY4g3MWNmoHKY0tg4m=eoB77ZcSGuq+MA@mail.gmail.com> |
On Mon, 29 Dec 2025 at 16:16, Danik via lists.yoctoproject.org <[email protected]> wrote: > Let me know if you have a better proposition for the implementation flow > I have presented, or if I have missed some already existing technology > in the Yocto Project. Either my colleagues or I would be eager to > contribute such a feature to the project. I'd say any patch-related effort should encourage and support a strict 'upstream first' culture for making patches, and this feels like a step in the opposite direction, or at best, something that can be badly misused. We've all seen the monstrous vendor kernels from hell, and other private vendor forks which are modified in ways that are poorly written, not documented and basically unsustainable. Even in the example above the patches contain basically no documentation or commit history of why any given patch was added (e.g. a link to an issue in a ticket system or even a commit message), whether it should be sent upstream or anything else that might help some poor developer (or consultant) to understand anything about it. That said, you probably want a 'pragmatic', 'practical' answer :) If so, then one approach is to list both the original component and the repo with patches one after another in SRC_URI, so they're both unpacked into ${UNPACKDIR}. Then sandwich a bitbake task between patch and configure that can apply patch files from the second component into the first. Appending to SRC_URI and implementing the task can be done in a class, and parametrized as needed with variables. Then inherit the class and set the variables from a bbappend, or just write the recipe directly that way. Alex