Building on what Alan said, with actively maintained projects, the build can fetch either release tarballs or a specific git hash. If we (NuttX) need to apply any patches, we keep those in our repo and apply them at build time.
A good example you can look at is our Python port: https://github.com/apache/nuttx-apps/tree/master/interpreters/python There are other examples in the apps tree. Just look around for anything with patches. :-) On Thu, May 14, 2026 at 9:48 AM Alan C. Assis <[email protected]> wrote: > Hi Felipe, > > Thank you very much for that great news! > > I will suggest you what was the Greg's approach for previous projects > included inside NuttX Apps: > > If the project is actively maintained, then you shouldn't include the > source code inside nuttx-apps/ > > But if the project is mostly a dead project: a project that hasn't been > actively maintained for many years (i.e. 3 or more), then it is fine to > include the source code there. > > BR, > > Alan > > On Thu, May 14, 2026 at 10:22 AM Felipe Moura Oliveira < > [email protected]> > wrote: > > > Hi all, > > > > I am working on a preliminary Dropbear server port for NuttX and would > like > > to confirm the expected integration approach before moving further. > > > > As a starting point, I followed David’s ESP-IDF-based steps and, after > some > > adjustments, I was able to get Dropbear running on an ESP32-C3. I then > > started porting it to NuttX. > > > > For the initial proof of concept, I placed the Dropbear server code under > > apps/netutils and kept the integration as simple as possible. With this > > approach, I was able to get it working. > > > > Before improving the port, I would like to confirm whether this is the > > correct location for the Dropbear source code, or if there is a more > > appropriate place in the NuttX apps tree. > > > > I also have a question about service initialization. Currently, I need to > > manually start the Dropbear application. My expectation is that the SSH > > server should be started automatically when enabled in the configuration. > > > > However, as far as I understand, there is no generic apps autostart > > mechanism that works across all boards. The alternative would be to add > > board-specific startup logic in each board bring-up code, but I would > > prefer to avoid that if possible. > > > > What would be the recommended approach for initializing this kind of > > network service in NuttX? Should this be handled by board bring-up logic, > > NSH initialization, an application-level startup mechanism, or some other > > pattern? > > > > Any guidance on the preferred architecture would be appreciated before I > > continue refining the port. > > > > > > -- > > *Felipe Moura de Oliveira* > > *Universidade Federal de Minas Gerais* > > Linkedin <https://www.linkedin.com/in/felipe-oliveira-75a651a0> > > <https://twitter.com/FelipeMOliveir?lang=pt-br> > > >
