Conversation
erickt
commented
Sep 24, 2020
- Switch many cases of must, optional to MUST, OPTIONAL to match RFC-2119.
- Change Freeze-Attack to use MUST instead of "should".
- Update links to https://theupdateframework.io
- Fix some typos.
- Standardize convention from "(if any)" to ", if any," since that was more common.
- Switch from "Check signatures" to "Check or an arbitrary software attack".
* Switch many cases of must, optional to MUST, OPTIONAL to match RFC-2119. * Change Freeze-Attack to use MUST instead of "should". * Update links to https://theupdateframework.io * Fix some typos. * Standardize convention from "(if any)" to ", if any," since that was more common. * Switch from "Check signatures" to "Check or an arbitrary software attack".
The rest of the TUF spec refers to each part of the document as $SECTION.$STEP. This updates the application workflow to also include the section number. This disambiguates referring to "2.2" from either referring to the "Thread Model and Analysis" from checking for a timestamp rollback attack.
There was a problem hiding this comment.
Thanks for the clean up, @erickt! The section number prefix in the client workflow steps IMHO is really helpful.
I took the liberty to rebase this PR on the recently updated master, which now includes #111, #116, #117 and a quick hotfix for the release checker script (d4c2f4b).
While rebasing I updated the version number and resolved some conflicts, and I also added a commit that fixes some references to sections whose numbers changed here (and in #116).
Here is the diff of your original branch and my updated and forced-pushed version of it.
Maybe you or someone else (@mnm678, @joshuagl) can take a quick look and (re-)green-light this.
| } | ||
|
|
||
| where PUBLIC is in PEM format and a string. All RSA keys must be at least | ||
| where PUBLIC is in PEM format and a string. All RSA keys MUST be at least |
There was a problem hiding this comment.
I understand why, but do we actually enforce this? Do we reject RSA public keys that correspond to private keys < 2048 bits?
There was a problem hiding this comment.
AFAIK we do not in the reference implementation.
Update references to adopt section and step number changes in: - theupdateframework#116, which added a "persist root metadata" step and thus pushed back the "check freeze attack" step; and in - a recent commit that added the section number as prefix to all steps of the client workflow. Co-authored-by: Trishank Karthik Kuppusamy <trishank.kuppusamy@datadoghq.com>
|
Thanks for taking another look, @erickt and @trishankatdatadog! I added your suggestion and also cherry-picked your related #123, @trishankatdatadog! |