isA<processorPolyPatch>

fedvasu · January 8, 2018, 21:07

isA<processorPolyPatch>(curPatch)

it always gives me 0, I am using 2.4.0.

I have verified it using Pout and patch name, even for processor patches.

Why?

fedvasu · March 3, 2020, 18:47

Seriously what is going on here?

This doesn't work as expected AT ALL. It always gives me zero.

Sometimes even isA<empytPolyPatch> also has the exact same behavior.

This is a really dark bug or I just don't understand anything at all.

raumpolizei · March 4, 2020, 03:12

Indeed, this sounds strange. Could you try to decompose a simple domain using the simple decomposition method? For instance a cube decomposed in 8 smaller domains, with each subdomain having 3 of its 6 patches as contact patch with other subdomains. Run decomposePar and check the log carefully. Then in a test application (I would suggest $WM_PROJECT_DIR/application/test/volField as starting point). You could try something like

Code:

// create the mesh
...
// mesh is fvMesh object - your mesh
auto& boundaries = mesh.boundaryMesh();
forAll(boundaries, patchID)
{
    const fvPatch& currentPatch = boundaries[patchID];
    Pout<< "patch " << currentPatch.name() << nl
            << "polyPatch check: " << isA<processorPolyPatch> << nl
            << "any additional check here" << nl;
}

I did not run the code on my machine but the syntax should be approximatelly correct. I'm using OF 16.12 so it could be slightly different for your version. Check your code guide (if you built the doxygen doc locally) or the extended code guide for your OF version.

January 8, 2018, 21:07	isA<processorPolyPatch>	#1
fedvasu Member Join Date: Oct 2013 Posts: 92 Rep Power: 12	isA<processorPolyPatch>(curPatch) it always gives me 0, I am using 2.4.0. I have verified it using Pout and patch name, even for processor patches. Why?

March 3, 2020, 18:47	Still UNSOLVED or UNEXPLAINED	#2
fedvasu Member Join Date: Oct 2013 Posts: 92 Rep Power: 12	Seriously what is going on here? This doesn't work as expected AT ALL. It always gives me zero. Sometimes even isA<empytPolyPatch> also has the exact same behavior. This is a really dark bug or I just don't understand anything at all.

March 4, 2020, 03:12		#3
raumpolizei Member Join Date: Dec 2018 Location: Darmstadt, Germany Posts: 87 Rep Power: 7	Indeed, this sounds strange. Could you try to decompose a simple domain using the simple decomposition method? For instance a cube decomposed in 8 smaller domains, with each subdomain having 3 of its 6 patches as contact patch with other subdomains. Run decomposePar and check the log carefully. Then in a test application (I would suggest $WM_PROJECT_DIR/application/test/volField as starting point). You could try something like Code: // create the mesh ... // mesh is fvMesh object - your mesh auto& boundaries = mesh.boundaryMesh(); forAll(boundaries, patchID) { const fvPatch& currentPatch = boundaries[patchID]; Pout<< "patch " << currentPatch.name() << nl << "polyPatch check: " << isA<processorPolyPatch> << nl << "any additional check here" << nl; } I did not run the code on my machine but the syntax should be approximatelly correct. I'm using OF 16.12 so it could be slightly different for your version. Check your code guide (if you built the doxygen doc locally) or the extended code guide for your OF version.