docs: add platforms reference and how-to #2080
Conversation
mr-cal
force-pushed
the
work/CRAFT-3720-multi-base-docs
branch
from
a156c79 to
3827095
Compare
Signed-off-by: Callahan Kovacs <[email protected]>
mr-cal
force-pushed
the
work/CRAFT-3720-multi-base-docs
branch
from
3827095 to
651c5be
Compare
|
@jnsgruk - do you mind reviewing this documentation? I'm looking for feedback from a charm dev who hasn't already tried building multi-base charms. For some reason, I can't add you as a PR reviewer. |
| .. code-block:: yaml | ||
|
|
||
| platforms: | ||
| [email protected]:amd64: |
There was a problem hiding this comment.
I’m kinda surprised by the difference between ‘ubuntu-22.04-amd64’ and ‘[email protected]:amd64’ - the two seem quite different and not in a very ‘obvious’ way?
There was a problem hiding this comment.
Yeah, I think I was too strict with the validation and am circumventing my own rules.
I made the validation strict to prevent ambiguous definitions, like this one that is a hybrid of shorthand and longhand notation with mismatched bases:
platforms:
[email protected]:riscv64:
build-on: [[email protected]:amd64, [email protected]:riscv64]
build-for: [[email protected]:riscv64]But I can circumvent my own validation by not using <base>:<arch> in the platform name:
platforms:
ubuntu-24.04-riscv64:
build-on: [[email protected]:amd64, [email protected]:riscv64]
build-for: [[email protected]:riscv64]I could use simple platform names like noble-riscv64 or I could drop the validation, such that the platform name is completely arbitrary when build-on and build-for are defined. @lengau, any thoughts?
There was a problem hiding this comment.
(and my examples were wrong, to add to the confusion)
There was a problem hiding this comment.
My preference would be that the only time we have limitations on the name of a platform is when it's used to infer the build-on and build-for. Otherwise Järvfjället is a perfectly good platform name.
There was a problem hiding this comment.
I'm fine with that. Once you, Claudio, and I agree on it, I can update the spec, relax the validation (which is trivial), and update the docs.
I don't think we need to block this PR, it can be a future improvement.
Signed-off-by: Callahan Kovacs <[email protected]>
There was a problem hiding this comment.
Since this is a critical topic, I've made some fairly in-depth suggestions.
I'm probably late to the party, but some of my points relate to the platform-architecture question, and affect the page and section titles. For clearer language that mimics natural speech, I'm leaning toward the notion of "build architecture" and "target architecture" (Docker uses this pair). I hope these proposals won't lead to blockages. If they do, feel free to ignore them.
I didn't give every line the same amount of attention, but I'm sure JJ and I will revisit these pages more than once later.
|
@medubelko requesting you again as I've got some follow-up questions. I haven't yet handled all the changes. |
| With both snippets, building on amd64 will produce two charms, one for | ||
| amd64 systems running Ubuntu 22.04 and one for amd64 systems running | ||
| Ubuntu 24.04. |
There was a problem hiding this comment.
| With both snippets, building on amd64 will produce two charms, one for | |
| amd64 systems running Ubuntu 22.04 and one for amd64 systems running | |
| Ubuntu 24.04. | |
| With both snippets, building on AMD64 will produce two charms, one for | |
| AMD64 hosts running Ubuntu 22.04 LTS and one for AMD64 hosts running | |
| Ubuntu 24.04 LTS. |
Since you changed it to "hosts" earlier in the page.
There was a problem hiding this comment.
I'm distinguishing here the build host from the runtime system, which is why I kept these as systems.
There was a problem hiding this comment.
Non-blocking, but what makes runtime different? Don't charms run on every kind of machine?
There was a problem hiding this comment.
@lengau I like the changes you've made. I've closed the ones we can't handle for now. The rest of these are mechanical improvements which should be easy to do, with one suggestion to resolve a semantic question. If you're short on time I can commit them today.
lengau
force-pushed
the
work/CRAFT-3720-multi-base-docs
branch
from
5cc482b to
fc7df45
Compare
Adds reference and how-to documentation for platforms, bases, and architectures in Charmcraft, including information on the new multi-base support.
Fixes #2009
(CRAFT-3720)