Annotate when using ctool

[ulrich.vdh@…]

gcov generate annoted c file.

Is it possible to use it to retreive annotation in the c:coverage and not only the percent of coverage ?

Thanks

[ulrich.vdh@…]

Patch to make it work

[anonymous]

I have made a patch to add visualisation of result of the gcov tool in brownser annotate.

Thanks

[hefloryd]

We have been using this patch for a while and it seems to work fine. Please apply it if possible.

[Martin Schulze <spam.martin.schulze@…>]

Now with ===== detection and skipping (with warning!) lines of unknown format

[Martin Schulze <spam.martin.schulze@…>]

I was pleased to see this patch but it did not work for me. The problem was that my gcov output had another line annotation: ===== for unexecuted lines that are in exceptional code paths.

Unfortunately the code did not cope well with this and errored out with an exception (pattern.search( line ) did not find anything so .groups was called on a None Type?).

Long story short: I revised the code a little bit to make it more robust and handle the aforementioned format as well.

There are still many improvements to make. Basically any exception in the code will throw it off track and give no meaningful error message to the server side which makes it hard to "debug".

[osimons]

Thanks!

What I'd really like is for the patch to include unit tests. Extend bitten.tests.ctools.GCovTestCase to test all supported formats and features. It makes the code a lot easier to support and maintain.

[osimons]

Replying to osimons:

Extend bitten.tests.ctools.GCovTestCase to test...

That should be bitten.build.tests.ctools.GCovTestCase of course.

[Martin Schulze <spam.martin.schulze@…>]

Okay, I have some slight improvements on the new code and two test that fail on the old patch and work on the new one:

• detection of unreached exceptional code (test_unreached_exceptional_path_is_zero)
• marking of lines that don't match the expected pattern (test_mark_unknown_lines)

I am tempted to test whether there are warnings on such lines but I don't know the framework good enough to test for this.

I have never before written any tests for bitten (or trac or python for that matter) so I would like some feedback on things I should improve upon (aka a code review).

I mostly copied together my code from the preexisting test and extended one of their helper functions (_create_file) for creating a dummy file with the possibility to give the file's contents. I hope I did not break anything there but these tests still pass.

[Martin Schulze <spam.martin.schulze@…>]

Improved code and tests

[osimons]

Superb. Thanks! I've committed the patch in [1027].