Coverity Fix: Close opened mmd libraries

[tylerzhao7684]

Type: Resource leak (RESOURCE_LEAK)
mmd_library is not properly closed under several exit conditions, we need to close it before exiting

[tylerzhao7684]

Seems like the mmd library is accessed globally, need another way to fix this

[tylerzhao7684]

Confirmed that it is this line that causes the unit test to fail
@zibaiwan Any idea on this?
Coverity doesn't like mmd_library (A pointer) to be lost even though the library is accessed globally

[tylerzhao7684]

I looked at #227 and used a similar method to close the library
Also looked at #264 and I think it caused a problem because in some cases dlsym need to be called on libraries that are opened globally

[tylerzhao7684]

@zibaiwan Integration test passed with this implementation
Review please!

[zibaiwan]

@tylerzhao7684 , as we are very close to merge this in. Assuming the window testing passes, could you please squash your 4 commits into 1 commit as they essentially all for fixing one issue?

The configuration in this repo is set to merge all 4 commits into the repo instead of squashing them into 1 automatically, to allow one large PR can have multiple meaningful commits. You can choose the way you like to do the squash, e.g., rebease -i or reset --soft, and then do a force push. Let me know if you have any questions!

Thanks you!

[zibaiwan]

@tylerzhao7684 , can you please take a look at my previous comment to squash the change? This PR is ready to merge once the commit is cleaned up.

[zibaiwan]

@tylerzhao7684 , can you please take a look at the workflow failure? My guess is, cov-errors.txt is now empty based on the empty output of the cat cov-errors.txt because you fixed the last Coverity issue. However, I don't understand why count=$(grep -c '^$' cov-errors.txt) && echo "$(( $count / 2 ))" would fail in this case, instead of just outputting 0.

One possible guess is: cov-errors.txt didn't exist and the grep failed, but it didn't make sense why cat cov-errors.txt didn't fail earlier. Can you please investigate a bit and this PR likely needs a supplement change in the workflow as well, otherwise the main branch workflow will show the failure status.

[tylerzhao7684]

This previous workflow didn't fail because #277 was not merged in, so even if this PR fixes one coverity issue, the other one still remain as it is not merged
Here's the previous workflow run

Now it fails because it couldn't find cov-errors.txt, which is a good sign because now we don't have coverity issues anymore!
I'll work on a small fix in this PR to make this workflow succeed when cov-errors.txt is not found

[zibaiwan]

This previous workflow didn't fail because #277 was not merged in, so even if this PR fixes one coverity issue, the other one still remain as it is not merged Here's the previous workflow run

Now it fails because it couldn't find cov-errors.txt, which is a good sign because now we don't have coverity issues anymore! I'll work on a small fix in this PR to make this workflow succeed when cov-errors.txt is not found

I am not 100% sure if it indeed failed due to cov-errors.txt is not found, because in the workflow, cat cov-errors.txt step didn't fail. I tested it locally on one of the linux machine, this would give an error like this cat: dummy.txt: No such file or directory, but maybe the workflow machine has a different linux distribution therefore no error? But if it indeed caused by no file found, I 100% agree this is a good sign!

[zibaiwan]

This previous workflow didn't fail because #277 was not merged in, so even if this PR fixes one coverity issue, the other one still remain as it is not merged Here's the previous workflow run

Now it fails because it couldn't find cov-errors.txt, which is a good sign because now we don't have coverity issues anymore! I'll work on a small fix in this PR to make this workflow succeed when cov-errors.txt is not found

A separate commit for the workflow fix is more appropriate (no need to squash) but can be merged in the same PR.

[tylerzhao7684]

This previous workflow didn't fail because #277 was not merged in, so even if this PR fixes one coverity issue, the other one still remain as it is not merged Here's the previous workflow run
Now it fails because it couldn't find cov-errors.txt, which is a good sign because now we don't have coverity issues anymore! I'll work on a small fix in this PR to make this workflow succeed when cov-errors.txt is not found

I am not 100% sure if it indeed failed due to cov-errors.txt is not found, because in the workflow, cat cov-errors.txt step didn't fail. I tested it locally on one of the linux machine, this would give an error like this cat: dummy.txt: No such file or directory, but maybe the workflow machine has a different linux distribution therefore no error? But if it indeed caused by no file found, I 100% agree this is a good sign!

You are correct, there is a cov-errors.txt.
The workflow failed because grep returns non-zero exit code when it finds no match, I added the fix now