Specify UTF-8 for all our JSON #146
Conversation
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. To get a place to say this for the runtime.json docs, I used config.md as a template for the top-level header and blurb. The "Host-specific container configuration" phrasing comes from bundle.md, which has: The `runtime.json` file contains settings that are host specific... While I was touching the config.md lead-in, I fixed 'container' -> 'bundle'. [1]: https://en.wikipedia.org/wiki/UTF-8 Signed-off-by: W. Trevor King <wking@tremily.us>
| # Host-specific container configuration | ||
|
|
||
| The bundle's top-level directory MUST contain a configuration file called `runtime.json` with [UTF-8][] [JSON][]. | ||
| For now the canonical schema is defined in [`runtime_config.go`](runtime_config.go) and [runtime_config_linux.go](runtime_config.go), but this will be moved to a formal JSON schema over time. |
There was a problem hiding this comment.
I would have the OS specific reference be more generic than pointing out linux individually.
There was a problem hiding this comment.
and "moved to a formal JSON schema over time" where and why this comment?
There was a problem hiding this comment.
On Fri, Sep 04, 2015 at 11:44:48AM -0700, Vincent Batts wrote:
and "moved to a formal JSON schema over time" where and why this comment?
Both are just echoing the current host-independent phrasing 1. I
tried to make that clear in the commit message [2] and original PR
post with the:
“To get a place to say this for the runtime.json docs, I used
config.md as a template for the top-level header and blurb.”
text, but I'm happy to use different commit-message phrasing if that
would help make the source more obvious.
There was a problem hiding this comment.
I see that phasing. This feels like something to have roadmapped and discussed rather than alluding to.
There was a problem hiding this comment.
On Tue, Sep 08, 2015 at 03:59:13PM -0700, Vincent Batts wrote:
This feels like something to have roadmapped and discussed rather
than alluding to.
Sounds reasonable to me. Shall I open the mailing-list thread, or
would you like to?
And is there an alternative phrasing you'd like to see here in the
meantime, so we can land the UTF-8 notes in this PR without waiting on
the desired-schema-language discussion? Copying the existing phrasing
from config.md seemed like the least opinionated approach, but maybe
just dropping the “For now the canonical schema…” line with it's
Go-file links would be better?
|
peanut gallery, why not use the JSON RFC here? https://tools.ietf.org/html/rfc7159 |
|
On Tue, Sep 08, 2015 at 04:39:10PM -0700, Jonathan Boulle wrote:
How does that help? RFC 7159 has 1: JSON text SHALL be encoded in UTF-8, UTF-16, or UTF-32. The default Implementations MUST NOT add a byte order mark to the beginning of a To test the Go handling in particular, ‘json.NewDecoder(os.Stdin)’ $ cat sample.json | go run test.go but crashes with UTF-16LE: $ cat sample.json | iconv -f UTF-8 -t UTF-16LE - | go run test.go I'd rather stay away from BOMs. I'd be ok with allowing runtimes to And I see nothing about UTF-8 on json.org or the ECMA-404 it |
|
we should just put something in a glossary saying that when we say json in the spec we mean uft8 encoded json. |
|
On Wed, Sep 09, 2015 at 09:46:41AM -0700, Brandon Philips wrote:
I'd rather keep the glossary informative, and put normative stuff in |
|
+1 for what @philips said and we are currently working on a protobuf idea and discussing this during the meetings. |
|
On Fri, Sep 25, 2015 at 01:43:33PM -0700, Michael Crosby wrote:
Glossary entry it is. I'll bump #107.
This is about what authors provide in their bundles. I don't see |
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But [1] seems too glib, and I can't find a more targetted link than just dropping folks into a Unicode chapter (which is what [1] does): The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011) With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95. I'd rather put this normative requirement in the configuration-spec files, but maintainer consensus was to put it in the glossary [2,3]. [1]: https://en.wikipedia.org/wiki/UTF-8 [2]: opencontainers#146 (comment) [3]: opencontainers#146 (comment) Signed-off-by: W. Trevor King <wking@tremily.us>
I wish there was a cleaner reference for what UTF-8 was. But linking
Wikipedia seems too glib, and I can't find a more targetted link
than just dropping folks into a Unicode chapter (which is what
Wikipedia does):
With the current v8.0 (2015-06-17), it's still §3.9 D92 and §3.10 D95.
To get a place to say this for the runtime.json docs, I used config.md
as a template for the top-level header and blurb. The “Host-specific
container configuration” phrasing comes from bundle.md, which has:
While I was touching the config.md lead-in, I fixed “container” →
“bundle”.