Add XML generator for sonarqube generic coverage report

[akmll]

Add command line option '--sonarqube' with output report file name to generate sonarqube generic coverage report as described in https://docs.sonarqube.org/latest/analysis/generic-test/

[latk]

Thank you for this PR, that looks fantastic. It is fine to leave the PR open while you work on it. The Travis test failures seem to only be about code style issues (unused variables).

What I need from you before I can merge:

• Please add your name to the AUTHORS file
• Please add a mention of this format to the README file.
• The recent pull request #307 moves to the lxml backend which reduces memory usage. I think this would be a good idea here as well. If you run into any Python 2 vs 3 problems, it's fine if new features only work under 3.
• I need at least one test case for the generator. This involves:
  1. creating a new target in the test's Makefile,
  2. running it manually to create the output file
  3. manually verifying the output
  4. moving the file into the test's reference folder
  5. extend the gcovr/tests/test_gcovr.py test runner to run this additional target if it is available

Please let me know if you run into any problems.

[codecov]

Codecov Report

Merging #308 into master will decrease coverage by 2.78%.
The diff coverage is 16.41%.

@@            Coverage Diff             @@
##           master     #308      +/-   ##
==========================================
- Coverage   94.89%   92.11%   -2.79%     
==========================================
  Files          15       16       +1     
  Lines        1823     1890      +67     
  Branches      315      327      +12     
==========================================
+ Hits         1730     1741      +11     
- Misses         46      101      +55     
- Partials       47       48       +1

Impacted Files	Coverage Δ
gcovr/configuration.py	97.22% <ø> (ø)	⬆️
gcovr/sonarqube_generator.py	15.87% <15.87%> (ø)
gcovr/__main__.py	91.21% <25%> (-1.84%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c018e2b...9c93208. Read the comment docs.

[codecov]

Codecov Report

Merging #308 into master will decrease coverage by 2.78%.
The diff coverage is 16.41%.

@@            Coverage Diff             @@
##           master     #308      +/-   ##
==========================================
- Coverage   94.89%   92.11%   -2.79%     
==========================================
  Files          15       16       +1     
  Lines        1823     1890      +67     
  Branches      315      327      +12     
==========================================
+ Hits         1730     1741      +11     
- Misses         46      101      +55     
- Partials       47       48       +1

Impacted Files	Coverage Δ
gcovr/configuration.py	97.22% <ø> (ø)	⬆️
gcovr/sonarqube_generator.py	15.87% <15.87%> (ø)
gcovr/__main__.py	91.21% <25%> (-1.84%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c018e2b...9c93208. Read the comment docs.

[codecov]

Codecov Report

Merging #308 into master will decrease coverage by 2.78%.
The diff coverage is 16.41%.

@@            Coverage Diff             @@
##           master     #308      +/-   ##
==========================================
- Coverage   94.89%   92.11%   -2.79%     
==========================================
  Files          15       16       +1     
  Lines        1823     1890      +67     
  Branches      315      327      +12     
==========================================
+ Hits         1730     1741      +11     
- Misses         46      101      +55     
- Partials       47       48       +1

Impacted Files	Coverage Δ
gcovr/configuration.py	97.22% <ø> (ø)	⬆️
gcovr/sonarqube_generator.py	15.87% <15.87%> (ø)
gcovr/__main__.py	91.21% <25%> (-1.84%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c018e2b...9c93208. Read the comment docs.

[codecov]

Codecov Report

Merging #308 into master will decrease coverage by 0.18%.
The diff coverage is 88.46%.

@@            Coverage Diff             @@
##           master     #308      +/-   ##
==========================================
- Coverage   94.89%   94.71%   -0.19%     
==========================================
  Files          15       16       +1     
  Lines        1823     1874      +51     
  Branches      315      324       +9     
==========================================
+ Hits         1730     1775      +45     
- Misses         46       49       +3     
- Partials       47       50       +3

Impacted Files	Coverage Δ
gcovr/configuration.py	97.22% <ø> (ø)	⬆️
gcovr/tests/test_gcovr.py	97.75% <100%> (ø)	⬆️
gcovr/__main__.py	93.24% <100%> (+0.18%)	⬆️
gcovr/sonarqube_generator.py	87.23% <87.23%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c018e2b...d83b899. Read the comment docs.

[akmll]

Sorry, I didn't notice it was reopened. I am quite a newbie in python so this might take time, and I use msys2 on windows, it seems some of the test cases would fail.

[latk]

No worries. I'm thankful for any work that you can do! The tests assume a specific GCC version (GCC 5) and may fail under different versions. This is not a problem, as long as Travis/Appveyor are green. If extending the tests is too tricky then I can look into it myself. TBH this is the first time in forever that a new test target is added, so this would be nontrivial for me as well. The part about tests that I definitely need from you is that you confirm that the generated XML is accepted by Sonarqube, for example by using your modified version of gcovr in your build pipeline. If that seems to work, I'm confident it will be all right for others as well. Again, thank you very much for working on this, and please reach out if I can help.

…

On Mon, 6 May 2019 at 14:12, akmll ***@***.***> wrote: Sorry, I didn't notice it was reopened. I am quite a newbie in python so this might take time, and I use msys2 on windows, it seems some of the test cases would fail. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#308 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AANTJD7H7XIQJNDPMNOGIXTPUAOCNANCNFSM4HK5VHBA> .

[akmll]

Finally I have made the unit tests to success. The uncovered lines of code share across the generators, so I think it is OK.

I have tested the generator with a real project under enterprise environment and it works like a charm, the coverage information is just the same as the html report generated.

I read in the PR #307 that the memory usage is high with xml.minidom. I was thinking about generating the xml text directly since the format is rather simple. I am not very familiar with python, so I am not sure if that would help.

Another thing needs to mention is that there is no stdout print option for this sonarqube generator, I could add it easily if there is such scenario.

[latk]

This looks great! I'll probably fiddle a bit with the details while merging, but all the necessary stuff is there now. Thank you for working on this :)

[latk]

OK, I've merged this manually! I decided not to use the string-based output generation but rather ported to lxml. Your idea to pass the output file name via --sonarqube rather than -o was good, so I added that to all other formats as well.

Thanks again for working on this!