-
Notifications
You must be signed in to change notification settings - Fork 71
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
Add RFC to define mixins for the Bionic stack #40
Conversation
Signed-off-by: Stephen Levine <stephen.levine@gmail.com>
@jonjonsonjr as discussed |
Signed-off-by: Stephen Levine <stephen.levine@gmail.com>
Think @sclevine you meant @jonjohnsonjr |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Cannot wait to see this!
|
||
- Neither the spec nor this RFC suggest whether or not stack authors or app developers are allowed to make additional changes to stack images without using mixin metadata. | ||
- This RFC doesn't propose a way for the project to describe changes to `ubuntu:bionic` that are not official LTS Ubuntu 18.04 packages. | ||
- When package groups (like `build-essential`) or dependent packages are installed, must stack images always specify all the packages added? Or is it only necessary when stack image authors want buildpacks to be able to depend on the dependent packages? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The latter. In order to make a mixin as broadly compatible as possible you'll probably want to list both the package group and all the packages that it includes.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So your suggestion is that stack authors should add all packages to the mixin list, but are not required to? Would this imply that a given stack run/build image can have arbitrary contents that aren't specified by the ID or mixins, as long as those contents don't disappear when newer versions are published?
[This RFC](buildpacks/rfcs#40) is working it's way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working it's way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working it's way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working its way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working its way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working its way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working its way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working its way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
[This RFC](buildpacks/rfcs#40) is working its way through and it seems useful to be able to use Heroku buildpacks on `cloudfoundry/cnb:bionic`. This also aligns with how riff is compatible with [multiple stacks](https://github.com/projectriff/node-function-buildpack/blob/master/buildpack.toml#L22).
It would be nice to support another `bionic` based stack to prevent Heroku from forking all the language buildpacks. Note: While [this RFC](buildpacks/rfcs#40) is being discussed, it hasn't landed yet.
Here are some more thoughts after our chat @sclevine @hone @jkutner I think we all agreed that the I think we could enhance An example might be a buildpack author for A I would limit buildpack authors to a single pack identifier when defining their stack id. I wouldn't want Later, if a buildpack is failing to build on a new stack that doesn't work because the buildpack author is targeting a pack that includes something they are not actually required to use, the author could then go learn what exactly is in that pack and start specifying the mixins in a more selective way. They could stop requiring Anyway, that is what I was thinking while we were chatting earlier. I think having a few mixin packs targeted at the buildpack author's purpose could lower the barrier. Once they hit a limitation, the existing granularity of the mixinx would allow the buildpack author to take that dive into understanding stacks and mixins. |
Signed-off-by: Stephen Levine <stephen.levine@gmail.com>
@hone @jkutner @jabrown85 updated based on the feedback above. I recorded the |
1. The base set of packages is `ubuntu:bionic` with `ca-certificates`, `libssl1.1`, and `openssl` added. | ||
2. Mixins without an `=` are Ubuntu 18.04 LTS package names without version or architecture info from official package mirrors. | ||
3. Mixins that begin with `set=` are defined by the project as such: | ||
a. `build:set=shell-utils` represents the change set: `build:curl`, `build:jq`, and yj CLI at build-time. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
a. `build:set=shell-utils` represents the change set: `build:curl`, `build:jq`, and yj CLI at build-time. | |
a. `build:set=shell-utils` represents the change set: `build:curl`, `build:jq`, and `build:yj` at build-time. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yj is not available as an Ubuntu package, so I went with "yj CLI at build-time" instead of "build:yj
."
Maybe it's worth defining build:set=yj
as well?
text/0000-bionic-mixins.md
Outdated
# Proposal | ||
[proposal]: #proposal | ||
|
||
This RFC proposes that we define `io.buildpacks.stacks.bionic` such that: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does it expand the scope too much to set a framework for adding new stacks under the io.buildpacks.stacks
namespace or is that just a separate proposal. For instance, I think setting the set=
pattern is probably more valuable then just defining what all the sets are in this proposal. @sclevine
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We could define set=
for io.buildpacks.stacks.*
, but it might be safer to hold off until we have additional stacks to define it for. The set=
convention may be applicable to most base images that use OS packages today, but container base images could change in unexpected ways in the future.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Since the location of where this stuff needs to live or how changes are made is being considered out of scope for this RFC, we should add this to the list of unresolved questions.
text/0000-bionic-mixins.md
Outdated
The Cloud Native Buildpacks project informally defines stack build and run images with ID `io.buildpacks.stacks.bionic` as containing exactly the set of Ubuntu packages in the `ubuntu:bionic` image on Docker Hub. | ||
This RFC proposes that we formalize the existing definition of `io.buildpacks.stacks.bionic` and define mixins on that stack to be Ubuntu 18.04 LTS packages. | ||
|
||
In general, a stack is a contract that should be more like "Ubuntu Bionic" and less like "Platform XYZ's Re-distribution of Ubuntu Binoic." |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In general, a stack is a contract that should be more like "Ubuntu Bionic" and less like "Platform XYZ's Re-distribution of Ubuntu Binoic." | |
In general, a stack is a contract that should be more like "Ubuntu Bionic" and less like "Platform XYZ's Re-distribution of Ubuntu Bionic." |
text/0000-bionic-mixins.md
Outdated
In general, a stack is a contract that should be more like "Ubuntu Bionic" and less like "Platform XYZ's Re-distribution of Ubuntu Binoic." | ||
Stack images can still be provided by platforms/vendors, but they should use the standard CNB stack contracts where applicable. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I find this language a bit confusing. On first read, I thought it meant that the images must be derived from ubuntu:bionic
, but I think it really means provides equivalent functionality to ubuntu:bionic
. Otherwise images like gcr.io/gcp-runtimes/ubuntu_18_0_4
won't qualify. And it may be worth expanding functionality: does it mean that /var/lib/dpkg
should be retained and that new packages can be installed with apt-get
?
1. The base set of packages is the set of packages distributed in `ubuntu:bionic` with `ca-certificates`, `libssl1.1`, and `openssl` added. The base image does not have to be derived from the `ubuntu:bionic` image, but must contain packages that provide exactly the same functionality to processes running as the CNB-specified, non-root user. | ||
2. When invoked as root, `apt-get` must be configured to use at minimum the set of package repositories that `ubuntu:bionic` is configured to use. `apt-get update` may be needed to download the package repository indices before `apt-get install` is functional. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@briandealwis do these clarifications address your concerns?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm, thanks
@hone added the final location of the definition of |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good, thanks for updating. Let's move to FCP and make sure @briandealwis has a chance to chime back in based on your changes.
Given Brian's response, FCP will end and the RFC will merge on 2020-03-26 unless there is additional feedback. |
Signed-off-by: Joe Kutner <jpkutner@gmail.com>
Readable