Restructuring of mechanism for compiling federated programs #1221

…time unit.

…re no key-value pairs

Also added back book keeping related to control reactions.

@lhstrh

…ng purposes with @lhstrh

…Simplify getNetworkReactionTarget

It is not clear that this is working properly because the compilations seem to happen too quickly for the progress changes to be visible in the progress bar.

…for TypeScript target.

…eConfig", instead of JSON-like string.

…void a warning.

…when the main reactor is a federate.

Make preamble from fed-gen valid TypeScript code

This is supposed to fix the "conflicting types for ‘lf_reactor_c_main’" issue.

I have not vetted this merge for correctness or to see if it compiles. I will try to be available to help out with fixing the problems that will come out of this merge since I remember what I did during the merge.

This code was accidentally deleted during the merge.

As the whole directory is generated it makes sense to remove all of them.

…ile. Priority os given to what was given in cmd line though.

Fixing the loss of the hostname that is specified in a federated lf f…

…ewly failing federaed tests.

This test confuses the LspTests.

We should not refuse to compile just because the generated code looks bad.

Previously we just had an NPE. The LSP tests understand reports that a feature is unsupported, but they don't understand NPEs.

The output of the test framework is visually appealing, but it swallows useful information.

Hopefully this resolves the failing Windows LSP tests.

…leted?

…d was successful

The HashMap was not thread-safe. Probably this does not solve any problems that we actually observed, but it's hard to know.

This change seems to get BroadcastFeedback to pass locally.

This issue is the same as the one with TS. It seems very minor.

They were failing because of linker errors. Usually when CCpp fails with linker errors and C does not, it is because of the shenanigans that we used to do with our #includes. So I merged the latest changes from reactor-c/main. This introduced another problem, so I cherry-picked Erling's commit related to the LF_THREADED compile definition.

Warning: increase in TestBase.MAX_EXECUTION_TIME_SECONDS! This is apparently necessary because of the quantity of dependencies that must be downloaded in order to run Python Docker tests.

Another case of the FileConfig class simply not being finished yet.

It looks like this test does not exist in `master`. In order to limit the scope of this PR to refactoring rather than adding new features, it seems wise to omit this test from CI.

As far as CI is concerned I suspect that it might have been sufficient merely to correct the path (".resolve("src")"), but on my machine I also had to install google-protobuf probably because it was not already installed globally. This could have been done by modifying the produced package.json instead of just copying it, as is done in updatePackageConfig, but the method used in this commit also seems to work.

This bug appeared in LSP tests, but it apparently affects federated TypeScript. There are a few things to point out about this: * I think the code now does what it was originally supposed to do, but I'm not sure. * It's not clear that the result of this computation is even being used. * Ad hoc string processing is a major pain point in our code base.

Refactoring + partial fix of Docker support

Update the test framework in fed-gen to the changes made in master

This still generates wrong code. For example, AfterNoTypes fails with a segfault because the body of _lf_gendelay_0reaction_function_0 (everything after `cinit`, which ends around line 256) is wrong. But I am having trouble tracing through the code generator to find out why.

No idea why this issue does not show up in the TypeScript federated runtime tests. Relevant lines in the CI log: * Unable to import file:/D:/a/lingua-franca/lingua-franca/test/TypeScript/fed-gen/TopLevelArtifacts/src/......srclibCount.lf: java.io.FileNotFoundException: D:\a\lingua-franca\lingua-franca\test\TypeScript\fed-gen\TopLevelArtifacts\src\......srclibCount.lf (The system cannot find the file specified) * Test client received diagnostic at line 0: Invalid escape sequence (valid ones are \b \t \n \f \r \" \' \\ ) * Test client received diagnostic at line 0: Couldn't resolve reference to Reactor 'Count'.

@lhstrh

Thanks to @lhstrh for the advice -- he was exactly right.

This is an off-by-one error in the computation of WORKERS_NEEDED_FOR_FEDERATE. The code in this commit is copied from other code from master/CUtil which was last touched by Steven Wong and which is deleted in this branch. I should also point out that the previous commit, which updated the reactor-c version, probably would not have been necessary given this commit.

Fix `--clean` flag in federated generator

This matters in case of AST transformations. The test framework uses an AST transformation to set tests as CCpp, and it needs the context set in order to do the AST transformation. But the context would need to be computed after the AST transformation in order for it to get the right target. This is the least invasive way to sidestep this dependency cycle. I acknowledge that there may be performance consequences, but I'm not concerned about performance consequences until they are shown to be of practical significance. The fact is that if we do not cache, we do not have two inconsistent views of the targetConfig in the presence of AST transformations that touch the targetConfig.

This reverts commit e5d8558.

This reverts commit f10f5ce.

This also adds `e.printStackTrace()` in the test framework. I prefer that over swallowing all errors (including their stack traces) and printing a generic "something went wrong in the tests" error message instead.

Removal of `OldFedFileConfig`

The issue was not exactly a concurrency bug in the sense that data was not being set to an invalid state (AFAICT) by concurrent access. But it was in the sense that shared state was causing an error in the compilation of one federate to abort the compilation of another federate. This is OK -- it might even be desirable -- but it is a source of complexity. The root cause of the problem is that non-errors were being reported as errors. That is the only aspect of the issue that is being fixed here, and maybe the only aspect that really needs to be fixed.

Arduino needs additional information for the flash command that we could use (i.e. port where arduino is connected), so it's added as a PlatformOption

Clean up attributes in fed-gen

When I am in Cory Hall, it takes me 88 seconds to run a Python Docker test on my laptop. I think that if we make the timeout much less than 180 seconds (double what is definitely required), we will be cutting it too close. It is bad to have spurious test failures in CI.

This should not affect the compiler -- only the language server. When editing a file, the user should get feedback about syntax errors, regardless of whether their build setup involves Docker. This means that redundant work will happen -- for example, npm will be run even though all of that build-related work is supposed to be orchestrated via Docker alone. That's fine. The build-related work is only for syntax checking during development, anyway.

Fix docker tests in fed-gen

This reverts commit 3f4f156.

… and unreachable RTI

…and in TestBase because it is not effective

Commits on Oct 19, 2022

Fix missing import for FederateConfig.

hokeun committed Oct 19, 2022

Configuration menu

View commit details

Copy full SHA for d0102ab

Browse repository at this point

Copy the full SHA

d0102ab View commit details

Browse the repository at this point in the history

Commits on Jan 7, 2023

Fix compile errors in CppGenerator

lhstrh committed Jan 7, 2023

Configuration menu

View commit details

Copy full SHA for 8e9aa51

Browse repository at this point

Copy the full SHA

8e9aa51 View commit details

Browse the repository at this point in the history

Commits on Jan 8, 2023

Fix compile errors in cpp package

lhstrh committed Jan 8, 2023

Configuration menu

View commit details

Copy full SHA for 1e84ed1

Browse repository at this point

Copy the full SHA

1e84ed1 View commit details

Browse the repository at this point in the history

Restructuring of mechanism for compiling federated programs #1221

Restructuring of mechanism for compiling federated programs #1221

Commits on Aug 4, 2022

Commits on Aug 5, 2022

Commits on Aug 6, 2022

Commits on Aug 8, 2022

Commits on Aug 9, 2022

Commits on Aug 10, 2022

Commits on Aug 11, 2022

Commits on Aug 12, 2022

Commits on Aug 13, 2022

Commits on Aug 26, 2022

Commits on Aug 27, 2022

Commits on Sep 21, 2022

Commits on Sep 24, 2022

Commits on Sep 25, 2022

Commits on Sep 26, 2022

Commits on Oct 14, 2022

Commits on Oct 19, 2022

Commits on Oct 20, 2022

Commits on Oct 24, 2022

Commits on Oct 25, 2022

Commits on Oct 28, 2022

Commits on Oct 31, 2022

Commits on Nov 3, 2022

Commits on Nov 8, 2022

Commits on Nov 9, 2022

Commits on Nov 18, 2022

Commits on Nov 19, 2022

Commits on Nov 21, 2022

Commits on Dec 17, 2022

Commits on Dec 24, 2022

Commits on Dec 29, 2022

Commits on Jan 7, 2023

Commits on Jan 8, 2023

Commits on Jan 9, 2023

Commits on Jan 10, 2023

Commits on Jan 11, 2023

Commits on Jan 12, 2023

Commits on Jan 13, 2023

Commits on Jan 14, 2023

Commits on Jan 15, 2023

Commits on Jan 16, 2023

Commits on Jan 17, 2023

Commits on Jan 18, 2023

Commits on Jan 19, 2023

Commits on Jan 20, 2023