A new coverity scan would probaby be a good idea

[cmorty]

It's probably I good Idea to run a new coverity scan (The last one was run mid last year). Maybe even set up a travis job (https://scan.coverity.com/travis_ci).
Also there are some old patches (#5627, #5637, #5616 and #5617) that might apply or can be closed. Not sure about their state.

[albert-github]

• Coverty scan would be indeed nice
  • having it with travis-ci would be nice as well, but it should not run with each (proposed) patch or merge
  • who will interpret the results
  • who will fix the results / create (proposed) patches
• [PATCH] Fix potential null pointer dereference in src/lodepng.cpp (Origin: bugzilla #736850) #5627 was a false positive and is already closed
• [PATCH] Fix missing break in switch case statement (src/namespacedef.cpp) (Origin: bugzilla #737234) #5637 I think has been solved in the mean time
• [PATCH] Fix another potential null pointer dereference in src/index.cpp (Origin: bugzilla #736393) #5616 I think has been solved in the mean time
• [PATCH] Return with error in case a null pointer is provided to QBuffer::writeBlock (Origin: bugzilla #736394) #5617 I placed a comment

[cmorty]

  * having it with travis-ci would be nice as well, but it should not run with each (proposed) patch or merge

Wouldn't work due to rate limiting. But running it every other day would be a start.

  * who will interpret the results
  * who will fix the results / create (proposed) patches

Currently you need to register to get access. If possible, the results could be public. That could get some drive-by pull-request. I personally would not worry too much. And if one is looking for a bug looking at the list of potential issues is often a good Idea.

OT: Receiving PRs isn't enough. They need to be merged, too. Having looked through the PRs (that's often the first thing I do when finding a bug) I found that there are many open PRs that are pretty old (up to 5 years). I'm not sure whether they merge at all or whether there will ever be someone to rebase them. Maybe adding something like https://probot.github.io/apps/stale/ could help sort out the PRs (and issues) that are still relevant. IM(!)O doxygen is loosing quite a bit o developer power by not merging or rejecting those PRs.

[albert-github]

A bit off topic but I did register and after login I can see the "View Defects" button but clicking on it results in a 503 error:

Service Unavailable

The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.

Apache/2.4.29 (Ubuntu) Server at scan3.coverity.com Port 443

[cmorty]

I thought that it really was a temporary thing. Maybe @doxygen or Martin Ettl can have a look at it. They are the project admins

[albert-github]

@cmorty I was under the impression that you were able to access the defects, but apparently you couldn't either.

I took some actions (maybe they will bring something):

• asked the question at stack overflow: https://stackoverflow.com/questions/56165239/unable-to-view-defects-of-coverity-scan-503-service-unavailable
• asked the same question to scan-admin@coverity.com (tried to login to the coverity / synopsys community but could not submit a support case).

[albert-github]

I just got from synopsys an email back that they have resolved the problem and I can see the defects again.

I now looked quickly at the results, would be good to have a new run.

[cmorty]

Still not working for me. :( ... It's not a 503 any more but:

It may take a few minutes before you can view your defects, when you change your email or password or sign-in with Github for the first time.

[albert-github]

I can see you in the list of members and when I understand it well you can login.
My advise write a nice email to scan-admin@coverity.com will, probably, take a little while as they are based in another time zone.

[albert-github]

We do a regular run of coverity.
I also have to assume that the access problems should be solved as well (as there are no further messages about it).

I think this problem has been solved in the mean time.

[albert-github]

An automatic github actions procedure has how been installed to run on a regular base.