Hi, First, a few question regarding the coverage builder. 1. Why is the coverage builder executed under a French locale ? 2. Does anyone check its results. Are those results used in some way? 3. I see that it is in the list of supported builder. Is the coverage builder relevant should we reject a patch if coverage is under a certain limit? While reading the wiki and narrative documentation I could not find the answer to the above questions. I assume that coverage reporter should be run on all builder and then aggregate all results. Otherwise I expect that a coverage builder on Linux will have a lot of false positives for Windows or OSX specific code. ---------- I think that aggregating coverage for all tests modules (ie the test_*.py) file itself from all supported builders (os/reactors) should help reviewer/developer identify tests which are not executed Maybe the results could be sent to one of the trendy web services and have a nice graph. Does Twisted have this in place? Do you think this might be useful? Is there a ticket for that? Thanks! -- Adi Roiban