Date
1 - 3 of 3
removing HEX_FILES_TO_MERGE?
Bolivar, Marti
Hi,
Is anybody using the HEX_FILES_TO_MERGE build system property? It was originally added by Nordic, but has since been removed from our downstream Zephyr distribution, and I personally think it would make sense to remove it from upstream as well for Zephyr 2.0 unless it's proven useful to others. There seems to be rough TSC consensus that upstream Zephyr should avoid trying to half-solve the problem of "stitching" together blobs into a final executable, in favor of letting users pick their own solutions. This seems to fall into that category. Any objections to removing this feature for Zephyr 2.0 (from actual users)? I thought I'd ping the list to check, especially since it's an officially (if briefly) documented feature in the LTS release: https://docs.zephyrproject.org/1.14.0/application/index.html Thanks, Marti |
|
lairdjm
Hi Marti,
I haven't personally used this feature at time of writing but I can imagine it would be useful for merging in a bootloader, e.g. we have a design that has 3 components: 1. Nordic MBR 2. User application 3. Bootloader So may give it a try out in the future. I imagine users of other devices have similar requirements e.g .with blobs for the ESP32. So I don't see why a working, cross platform feature like this should be removed so that 'users can come up with their own solution', if they want to do that then they can disable the feature and do as they wish already, leaving it in doesn't (as far as I know) cause any issues so why remove it? Thanks, Jamie |
|
Benjamin Lindqvist
If the feature isn't horribly misimplemented, does what is expected of it and is expected to have very little maintenance costs, why drop it? I haven't used the feature, but I can definitely picture a situation where it might be useful. On Thu, Jun 27, 2019 at 8:35 AM lairdjm <jamie.mccrae@...> wrote: Hi Marti, |
|