Updating generic-table-dev branch with p4lang:p4runtime:main #468
Conversation
Change CI workflow to skip publishing if PR spawned by dependabot
* Define P4Runtime API support for tables with initial entries * Add TODO asking whether the format for the contents of entries files should be specified in the P4Runtime spec. * Fix a couple of things found by linter and compiling protobuf * Update autogenerated files * Document that TableEntry const field must be false in write requests * Add an appendix describing the contents of entries files generated by p4c * Clarify some wording. * Fix Madoko lint check * Replace TODO with cross reference to new appendix on entries files and clean up Madoko formatting in that appendix. * Replace TODO with an optimistic footnote. * Propose that TableEntry has new field const true for const entries and also for const default_action * Update auto-generated files * Define has_initial_entries to be true for tables with `const entries` Also fix a couple of spelling typos. * Update auto-generated files * Address several review comments * Address some more review comments. * Update auto-generated files again * Slight change in definition of has_initial_entries flag Propose that it is true if and only if the table property `entries` is present, _and_ the list of entries is not empty. * Update auto-generated files * Add "added in 1.4.0" notes to the two new fields * Clarify the description of the content of an entries file * Fix a typo, and add is_const field to list of TableEntry fields * Address review comment in new appendix
* Fix p4lang#434: Remove obsolete TODO section in README Update the link to the auto-generated versions of the P4Runtime specification on the P4.org web site. Update the section "P4 Language Version Applicability" to version 1.2.4 of the P4_16 language specification, but list 3 known exceptions of features that have not been explicitly addressed yet. * Add P4_16 v1.2.4 language spec features that may need addressing in a future version of the P4Runtime API specification.
* Update discussion of entry priorities in constant tables * Correct description of entry priority for constant tables
Bumps [golang.org/x/net](https://github.com/golang/net) from 0.9.0 to 0.17.0. - [Commits](golang/net@v0.9.0...v0.17.0) --- updated-dependencies: - dependency-name: golang.org/x/net dependency-type: indirect ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
) * Remove 4 P4 language spec compatibility issues from the list During 2023-Sep-08 P4.org API work group meeting, it was agreed that there are no changes required to the P4Runtime API specification to be compatible with these updates in the language spec. * Add clarifying behavior of table with no `key` property back in since there are potentially open issues around p4c implementation and how it generates size field of tables in P4Info files that should be considered before considering that issue resolved.
Same role as the metadata field for table entry
Bumps [google.golang.org/grpc](https://github.com/grpc/grpc-go) from 1.56.1 to 1.56.3. - [Release notes](https://github.com/grpc/grpc-go/releases) - [Commits](grpc/grpc-go@v1.56.1...v1.56.3) --- updated-dependencies: - dependency-name: google.golang.org/grpc dependency-type: direct:production ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
|
Hi @saynb I am a bit puzzled why this PR isn't just "sync main to generic-table-dev." It would make the lineage clearer. |
|
There were some merge conflicts on that #469. So decided to do it this way instead |
|
Understood. |
|
Merging. @saynb please confirm it resolves to be same as main, otherwise we can fall back to a rebase and force-merge. |
|
@saynb got CI build error, even after retry: https://github.com/p4lang/p4runtime/actions/runs/7671762843/job/20910840903#step:5:42 |
|
@chrispsommers I can confirm that the generic-table-dev branch has no diff with main even if changes show up in a PR if you try and create one from main->generic-table-dev. |
|
Hi @antoninbas do you have any ideas about the CI failures cited above? |
|
It's a bazel build issue that I haven't had time to look into. |
|
Interesting! Looks like a regression on the main branch. |
|
Here is what changed: vs Looks like we're not yet ready for Bazel 7.x. This is easy to fix by adding a |
|
@saynb Could you do another PR to sync to main and gain the benefit of the Bazel fix? Thx. |
No description provided.