Add IS-11 support by lo-simon · Pull Request #474 · sony/nmos-cpp

lo-simon · 2025-11-14T13:39:07Z

Add IS-11 NMOS Stream Compatibility Management support.
This is superseded IS-11 support #271 which was initially created by @N-Nagorny

…move Flow Compatibility API helper functions into details namespace

…andler

…raint Set is a subset of another one

…zable

…alue with debug message

…callback to parse "transport_file" and Stream Compatibility Management API callback to perform application-specific receiver validation with its transport file.

… be initialized after" warning

jonathan-r-thorpe · 2026-01-21T16:25:11Z

+        auto output = edid_support
+            ? nmos::experimental::make_streamcompatibility_output(output_id, device_id, true, boost::variant<utility::string_t, web::uri>(edid), receiver_ids, model.settings)
+            : nmos::experimental::make_streamcompatibility_output(output_id, device_id, true, receiver_ids, model.settings);


Hmmm, what's the behaviour if EDID support is set to true in the config? Does this pass any of the EDID tests in the NMOS Testing tool?

jonathan-r-thorpe · 2026-01-21T16:33:41Z

+            if (!nmos::caps::meta::enabled(constraint_set)) continue;
+            for (const auto& sender_caps_constraint_set : sender_capabilities)
+            {
+                const auto intersection = nmos::experimental::get_constraint_set_intersection(sender_caps_constraint_set, constraint_set);


Should we worry about the Windows 2022 warning?
declaration of 'intersection' hides function parameter

Rename the variable intersection to constraint_set_intersection to prevent declaration of 'intersection' hides function parameter warning.

jonathan-r-thorpe · 2026-01-21T17:03:15Z

+        }
+
+        // Constraint B is a subconstraint of Constraint A if:
+        // 1. Constraint B has enum keyword when Constraint A has it and enumerated values of Constraint B are a subset of enumerated values of Constraint A


"Constraint B has enum keyword when Constraint also has it" perhaps?

Co-authored-by: jonathan-r-thorpe <64410119+jonathan-r-thorpe@users.noreply.github.com>

…o prevent `declaration of 'intersection' hides function parameter` warning

# Conflicts: # Development/nmos-cpp-node/node_implementation.cpp

simonbaren · 2026-02-26T07:51:17Z

Hi @lo-simon and @jonathan-r-thorpe thanks for reviving this PR. I've been using the fork for long time now and will be glad to see it in main.
@lo-simon If you need additional set of hands, feel free to contact me.

jonathan-r-thorpe · 2026-03-05T09:38:51Z

Hi @lo-simon and @jonathan-r-thorpe thanks for reviving this PR. I've been using the fork for long time now and will be glad to see it in main. @lo-simon If you need additional set of hands, feel free to contact me.

Hi @simonbaren - have you tried running this branch against the NMOS Testing tool IS-11 test suite? We're having a problem when we run the test suite with EDID switched on that we're trying to understand.

simonbaren · 2026-03-05T09:58:02Z

@jonathan-r-thorpe yes we did. But we have a custom node-implementation (to interact with our low-level stuff).
I know that we did fail some of the IS-11 EDID tests with errors COULDN'T TEST.
I can try to gather which specific tests it was.

If you want I can try to run the code as-is against nmos-testing to see what is missing.
Note - that there are tests (can't recall the test case number) that in the testing procedure you need to read the advertised EDID with an external tool to verify the test - i.e there is a manual test for it. It happens on the tests with the active constraints_sets.

simonbaren · 2026-04-15T10:41:40Z

@jonathan-r-thorpe I had another look on the IS-11 project - the EDID code (constraints_sets) is just a stub and doesn't actually modify the effective_edid. Might that be the issue you are seeing?

N-Nagorny · 2026-04-28T17:53:58Z

Hello guys! Glad to see the PR was revived.

Yes, three tests fail.

Test ID	Test name	Description	Resolution
test_01_03	Inputs with Base EDID support reject an invalid EDID	The test PUTs a nulled binary to Base EDID. EDID parsing was out of scope for the reference impementation, so `nmos-cpp` accepts any EDID (returns 204).	We can implement basic EDID support. https://github.com/N-Nagorny/libedid could come in handy here, though I can't commit to maintaining it. Alternatively, we can `--ignore` it in `run_nmos_testing.sh`.
test_02_03_05_01, test_02_03_05_02	Verify for inputs supporting EDID that the version and the effective EDID change when applying constraints	The test tries to constrain a Sender to certain `grain_rate` for video and `sample_rate` for audio respectively. `node_implementation.cpp` declares Sender Capabilities that allow only single value of `grain_rate` and `sample_rate`, so Node can't adhere to `constraint_sets` from the tests and returns 422. This valid response should result into "Could Not Test" but the test suite has a bug in it (fixed in AMWA-TV/nmos-testing#852).	Revive AMWA-TV/nmos-testing#852

simonbaren · 2026-04-28T18:05:11Z

Hi @N-Nagorny .
About test_02_03_05_01 and test_02_03_05_02 - those are very Generator reliant.
I mean - if your source Generator comply to EDID changes - test will fail.
The solution to that (was applied by IPMX I think) is to add a manual step for those tests. When test runs, the tester is prompted to go and read the EDID from the device and check that it changed.

As a semi-automatic test, it is possible to read the effective_edid and validate that it was constrained according to the request, but this is kind of wrong, because the test should check what is on the wire side and not the nmos side.

regarding test_01_03 - it pass for the implementation we run (we do use the libedid: Edid::parse_edid_binary(edid)) .

N-Nagorny · 2026-04-28T20:50:08Z

I mean - if your source Generator comply to EDID changes - test will fail.

Should it have been "doesn't comply" or I'm missing something?

was applied by IPMX I think

Could you please share links?

because the test should check what is on the wire side and not the nmos side

Effective EDID just tells what the modeled device gives to its upstream input. NMOS Tests don't check deeper than the NMOS side by design, like they don't capture UDP streams after activating a Sender.

Could you please tell more about your implementation and run AMWA-TV/nmos-testing#852 against it?

simonbaren · 2026-04-29T07:31:06Z

@N-Nagorny

Should it have been "doesn't comply" or I'm missing something?

Yes, "doesn't comply" is correct — my mistake.

Could you please share links?

Sorry but I don't have a direct link. It's an additional test configuration where the tester is prompted to read the EDID from the DUT at a certain point and verify it changed on the wire.

Effective EDID just tells what the modeled device gives to its upstream input. NMOS Tests don't check deeper than the NMOS side by design, like they don't capture UDP streams after activating a Sender.

The UDP analogy doesn't hold here. A sender that doesn't stream what it's supposed to would still fail downstream — the UDP stream itself is independently verifiable. But an EDID change that never reaches the wire has no such natural verification. A node could trivially update effective_edid on the NMOS side without ever pushing the change to the physical input — making it look compliant while not actually being so. The example node_implementation is actually a perfect illustration of this: there is no real HW behind it, so effective_edid is just a value in memory with no physical effect. That's exactly the gap the manual step is designed to close, and it's also why a node that only "looks good" on the NMOS side is not truly IS-11 compliant.

N-Nagorny added 30 commits November 17, 2022 01:57

Add Flow Compatibility Management API base route

d5e46c1

Add IS-11 resource makers and implement all GET operations

487818e

Implement PUT and DELETE for /inputs/{inputId}/edid/base

79d2024

flowcompatibility_api.cpp: refactor repeated code

634967d

Add IS-11 and BCP-004-01 JSON Schemas

6d2a212

Implement PUT and DELETE for /senders/{senderId}/constraints/active, …

ca8c804

…move Flow Compatibility API helper functions into details namespace

Implement States of Sender and Receiver

4b8f590

Fix bst::optional nullopt values

fff7420

Allow making Outputs without EDID

3de8159

Remove "senders"/"receivers" arrays from Input/Output properties

142f765

Add Sender Capabilities to flowcompatibility_active_constraints_put_h…

7ba0287

…andler

IS-11: update schemas, apply the new spec name to the sources

b11402e

flowcompatibility -> streamcompatibility

ae50dd1

Node: replace EDID examples with dummies, add minor fixes

d230c3a

Sync Sender/Receiver states with IS-11 spec

2bcaba0

Prevent IS-11 restrictions violation

16941f7

Fix is_subconstraint(), add unit tests

f2b8918

Get rid of list of parameter constraints used for checking if a Const…

6604027

…raint Set is a subset of another one

Align bst::optional usage in IS-11 with the rest of code

583969d

Replace string literals with platform-independent version

b3cf776

Convert EDID binary stored as utility::string_t to std::vector<uint8_t>

3755db2

Sync IS-11 Status with the spec

986820e

Make Sender validator of streamcompatibility_behaviour_thread customi…

c8f111b

…zable

streamcompatibility_behaviour: make namespaces full

278a802

is_subconstraint(): fix the rational case

4836364

Add match_constraint() for arbitrary JSON value

b48f40f

is_subconstraint(): fix comparing of rationals

8141286

Add constraints matcher for Senders

c5d2000

Add validate_receiver() callback, expand validate_sender()'s return v…

2bac1b6

…alue with debug message

Remove senderReceiverSubrouteType

e94eddf

lo-simon added 3 commits November 25, 2025 15:09

Remove unnecessary notification

2c503d0

This document requires update, remove it for now

49175c9

Update to include IS-11 and other missing IS info

b91c599

lo-simon marked this pull request as ready for review November 25, 2025 15:18

Merge branch 'sony:master' into IS-11-patch

53ca6f0

jonathan-r-thorpe reviewed Dec 15, 2025

View reviewed changes

Comment thread Development/nmos-cpp-node/node_implementation.cpp

lo-simon added 2 commits January 7, 2026 01:03

Use the common "validate_sdp_parameters" function for Connection API …

b8cb28b

…callback to parse "transport_file" and Stream Compatibility Management API callback to perform application-specific receiver validation with its transport file.

Reorder node_implementation data members to prevent "data member will…

5455a94

… be initialized after" warning

jonathan-r-thorpe reviewed Jan 21, 2026

View reviewed changes

Comment thread Development/nmos-cpp-node/node_implementation.cpp Outdated

jonathan-r-thorpe reviewed Jan 21, 2026

View reviewed changes

Comment thread Development/nmos-cpp-node/node_implementation.cpp Outdated

jonathan-r-thorpe reviewed Jan 21, 2026

View reviewed changes

Comment thread Development/nmos/constraints.cpp

jonathan-r-thorpe reviewed Jan 21, 2026

View reviewed changes

lo-simon and others added 8 commits January 22, 2026 09:05

Update Development/nmos-cpp-node/node_implementation.cpp

22993d2

Co-authored-by: jonathan-r-thorpe <64410119+jonathan-r-thorpe@users.noreply.github.com>

Update Development/nmos-cpp-node/node_implementation.cpp

0d72ebd

Co-authored-by: jonathan-r-thorpe <64410119+jonathan-r-thorpe@users.noreply.github.com>

Add N-Nagorny's IS-11 doc

9dc5bf2

Merge remote-tracking branch 'origin/master' into IS-11-patch

78fa408

Rename the variable intersection to constraint_set_intersection t…

5df9e87

…o prevent `declaration of 'intersection' hides function parameter` warning

Merge branch 'nmos-cpp-master' into IS-11-patch

053589a

# Conflicts: # Development/nmos-cpp-node/node_implementation.cpp

Remove blank spaces

841cdf4

Add missing ; for lambda function

764946c

Conversation

lo-simon commented Nov 14, 2025

Uh oh!

Uh oh!

jonathan-r-thorpe Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

jonathan-r-thorpe Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

lo-simon Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jonathan-r-thorpe Jan 21, 2026

Choose a reason for hiding this comment

Uh oh!

simonbaren commented Feb 26, 2026

Uh oh!

jonathan-r-thorpe commented Mar 5, 2026

Uh oh!

simonbaren commented Mar 5, 2026

Uh oh!

simonbaren commented Apr 15, 2026

Uh oh!

N-Nagorny commented Apr 28, 2026

Uh oh!

simonbaren commented Apr 28, 2026

Uh oh!

N-Nagorny commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

simonbaren commented Apr 29, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

N-Nagorny commented Apr 28, 2026 •

edited

Loading