Diagram synthesis is broken for many tests and examples

[cmnrd]

I noticed that many of our examples do not produce valid diagrams in Epoch anymore. Instead, only an error message regarding a NullPointerException is shown. I suspect that this is caused by the recent refactorings in #759, but it could also be any other of the recent large PRs.

Examples and tests that do not work include:

• example/C/src/Delay.lf
• example/C/src/Patterns/Chain_02_Pipeline.lf
• example/C/src/Patterns/Loop_01_Single.lf
• example/C/src/Patterns/Loop_02_SingleDelay.lf
• test/C/src/ActionWithNoReaction.lf
• test/C/src/After.lf
• test/C/src/Alignment.lf
• test/C/src/CountTest.lf
• test/C/src/DanglingOutput.lf
• test/C/src/Deadline.lf
• test/C/src/DelayedAction.lf
• test/C/src/ToReactionNested.lf
• test/C/src/TriggerDownstreamOnlyIfPresent2.lf
• test/C/src/UnconnectedInput.lf

Note that this list is not exhaustive. There are many more files that do not produce valid diagrams. I was not able to identify a clear pattern of which language elements produce valid or invalid diagrams.

I think this shows that we need to be more thorough with testing. The correct operation of our diagram synthesis is currently not validated at all. I will open a separate issue for this (see #802).

[lhstrh]

It looks like the build started failing <https://github.com/lf-lang/lingua-franca/actions/runs/1581658566> after yesterday's merge of PR #800 <#800>.

…

-- Marten Lohstroh | Postdoctoral Researcher University of California | 545Q Cory Hall Berkeley, CA 94720 | +1 510 282 9135

On Wed, Dec 15, 2021 at 7:53 AM Edward A. Lee ***@***.***> wrote: For reference, the examples cited above are in example/C not test/C. Also, in master, the diagram package is generating the wrong diagram (it shows the Eclipse object model, not the LF program). I don't know how to change this. In the error logs, I get the following: org.osgi.framework.BundleException: Could not resolve module: org.lflang.diagram [511] Unresolved requirement: Require-Bundle: de.cau.cs.kieler.klighd.lsp; bundle-version="2.0.0" at org.eclipse.osgi.container.Module.start(Module.java:463) at org.eclipse.osgi.container.ModuleContainer$ContainerStartLevel.lambda$1(ModuleContainer.java:1834) at org.eclipse.osgi.container.ModuleContainer$ContainerStartLevel.incStartLevel(ModuleContainer.java:1829) at org.eclipse.osgi.container.ModuleContainer$ContainerStartLevel.incStartLevel(ModuleContainer.java:1775) at org.eclipse.osgi.container.ModuleContainer$ContainerStartLevel.doContainerStartLevel(ModuleContainer.java:1739) at org.eclipse.osgi.container.ModuleContainer$ContainerStartLevel.dispatchEvent(ModuleContainer.java:1661) at org.eclipse.osgi.container.ModuleContainer$ContainerStartLevel.dispatchEvent(ModuleContainer.java:1) at org.eclipse.osgi.framework.eventmgr.EventManager.dispatchEvent(EventManager.java:228) at org.eclipse.osgi.framework.eventmgr.EventManager$EventThread.run(EventManager.java:339) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#801 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYD47DMVPSBI5C75BYNVM3URC2WXANCNFSM5KC7PF4A> .

[edwardalee]

OK, these errors seem to be related to recent changes to support SI time units (?). Specifically, JavaASTUtils.java has the following:

    public static TimeValue getTimeValue(Value v) {
        if (v.getParameter() != null) {
            return getDefaultAsTimeValue(v.getParameter());
        } else if (v.getTime() != null) {
            return toTimeValue(v.getTime());
        } else {
            return null;
        }
    }

This is returning null for these examples. In particular, it returns null when the time value is specified as the literal "0". TimerInstance uses the above method to set its offset and period fields, and the diagram synthesis is assuming that those fields are never null.

I think we should rewrite this code to in fact ensure that these fields are never null. Volunteer to do that?

[cmnrd]

Thanks for the analysis. I already created a fix, but I found a few other problems on the way. I will push something soon.