Hi, so I ran lcov on master again. The results can be found here: http://people.canonical.com/~bjoern/lcov_reports/master-2015-06-10_23.49.39/ @Cloph: Can you copy that directory over to dev-builds? I dont know for how long I can host 1GB there. In general, this went well -- except that sw_unoapi_3 reliably deadlocked for me on that commit (so some writer coverage might actually be underreported). Its a far fetch to compare this to the last coverage run as configure switches[1] and lots of other things might be different. But lets do it anyways just for fun: Between 2014-11-03 and 2015-06-10 (7 months): Line coverage went from 40.8% to 42.9% (+5%). Function coverage went from 50.1% to 49.9% (~unchanged). Absolute lines hit went from 901,356 to 953,406 (+6%). Absolute functions hit went from 177,780 to 191,916 (+8%). I also had a deeper look at one code area: writer core (sw/source/core). See attached spreadsheet for the details. Line coverage went from 54.4% to 56.3% (+4%). Function coverage went from 67.6% to 68.7% (+2%). There are 2996 LOC more covered by tests (1.72% of writer core LOC). There are 208 functions more covered by tests (1.33% of writer core). There are 3721 LOC less not covered by tests (4.91% of writer core LOC). There are 139 functions less not covered by tests (2.84% of writer core). Best, Bjoern [1] For reference: My run was done with the switches shown here: http://people.canonical.com/~bjoern/lcov_reports/master-2015-06-10_23.49.39/lcov-configure.txt
Attachment:
lcov-writer.ods
Description: application/vnd.oasis.opendocument.spreadsheet