| Commit message (Collapse) | Author | Age |
| |
|
|
|
|
|
|
|
|
| |
Only marking one line as uncovered gave 100% miss, but when grouped with
other files would be missleading, (0% + 0% + 100% \simeq 99%).
This is also incorrect, since non-code lines are marked non-covered. But
it draws the attention to the right places.
|
| |
|
| |
|
|
|
|
|
| |
Guile's coverage system often missed those. Explicitly adding coverage
information make percentage go up!
|
| |
|
|
|
|
|
|
|
|
| |
The recent rewrite where coverage reports where limited to explictly
mentioned files lead to non-covered files being completely left out.
This re-introduces them with 0% coverage, noted on the second line (to
hopefully indicate that something is amiss).
|
|
|
|
| |
Limit testrunner to only discover *.scm files.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
Unhandled exceptions thrown during tests are now caught, and properly
repported. This also allows the tests to run to completion, and print
their final summary (which I previously believed was broken when running
in multiple threads).
Also add --help flag to testrunner...
|
|
|
|
| |
Move many of the atomic procedures into proper libraries.
|
|
|