Users/hc/gnntest

[nicelhc13]

CI/CD tests were broken due to this PR.

[nicelhc13]

@patrickkenney9801 I put a log about a file stat in CuSP (please search "File size: XX" on the log). On my end, tester.csgr was 136. But, this CI/CD test printed 128, which was different from my end. And it failed on OfflineGraph's initialization list, which sets ifstream and ofstream. So my current hypothesis is due to some unknown reason, input files on the server running the CI/CD test were broken. I am not 100% for sure but the log looks like it. If you can easily access the machine, could you please just do ls -alh on the input graph path?

[patrickkenney9801]

@patrickkenney9801 I put a log about a file stat in CuSP (please search "File size: XX" on the log). On my end, tester.csgr was 136. But, this CI/CD test printed 128, which was different from my end. And it failed on OfflineGraph's initialization list, which sets ifstream and ofstream. So my current hypothesis is due to some unknown reason, input files on the server running the CI/CD test were broken. I am not 100% for sure but the log looks like it. If you can easily access the machine, could you please just do ls -alh on the input graph path?

I can only check the file paths when the ci job is running, I'll check later today

[nicelhc13]

I confirmed that the path exists FYI if this is what you meant.

[patrickkenney9801]

@patrickkenney9801 I put a log about a file stat in CuSP (please search "File size: XX" on the log). On my end, tester.csgr was 136. But, this CI/CD test printed 128, which was different from my end. And it failed on OfflineGraph's initialization list, which sets ifstream and ofstream. So my current hypothesis is due to some unknown reason, input files on the server running the CI/CD test were broken. I am not 100% for sure but the log looks like it. If you can easily access the machine, could you please just do ls -alh on the input graph path?

I can only check the file paths when the ci job is running, I'll check later today

To follow up on this, ci has been extremely bloated lately, maybe this weekend I will be able to attempt this check

[nicelhc13]

Lemme know when you work on it. One thing that I can try is to add new input graph files not through lfs and check if they work. Now I am 90% sure that this issue is related to files.

[patrickkenney9801]

Lemme know when you work on it. One thing that I can try is to add new input graph files not through lfs and check if they work. Now I am 90% sure that this issue is related to files.

I kicked it off and it worked just fine, the second time it had a hang in the tests: https://github.com/utcs-scea/Galois/actions/runs/8532556268/job/23566845847?pr=20

Kicking it off a third

[patrickkenney9801]

Lemme know when you work on it. One thing that I can try is to add new input graph files not through lfs and check if they work. Now I am 90% sure that this issue is related to files.

I kicked it off and it worked just fine, the second time it had a hang in the tests: https://github.com/utcs-scea/Galois/actions/runs/8532556268/job/23566845847?pr=20

Kicking it off a third

Honestly I think that galois just hangs sometimes and that's a fact of life

[tewaro]

@patrickkenney9801 @nicelhc13 Where are we with these tests?

[patrickkenney9801]

@patrickkenney9801 @nicelhc13 Where are we with these tests?

I say we merge it and accept some flakes