John,
I finally go around to merging your patch set. I'm assuming this patch
set will apply to the stable 5 branch as well. Thank you for your
contribution to KiCad.
Cheers,
Wayne
On 07/10/2018 11:55 AM, John Beard wrote:
> Hi Wayne,
>
> (Feel free to tell me to come back to this later if
John,
I didn't forget about your qa test patches. They are on my todo list.
I've just been busy with life. I will have some free time over the
weekend to test them and get them merged.
Cheers,
Wayne
On 7/24/2018 6:31 AM, John Beard wrote:
> I think the first three of these patches are still
I think the first three of these patches are still worth committing
now. That will allow "make test" to be used to collect all our
(working) tests and give us somewhere to hang new tests so that they
can used.
We can discuss whether or not we want "make test" to also rebuild the
tests separately,
Hi Wayne,
(Feel free to tell me to come back to this later if busy with v5!)
On closer inspection, qa_geometry is NOT a general test of geometry
code, but rather a few tests to ensure that some refactored code
(SHAPE_POLY_SET) matches the original CPolyLine-based code.
As for the fillet code
I'm not sure what bothers me more, the fact the test fails or the fact
that the test used different fillet code as the pass/fail criteria
rather than a hard coded known correct test. Which fillet code was/is
correct? We really do need to start doing a better job with our
geometry tests and
Note to Wayne: Nothing here concerns v5 release, I was just trying to
get a geom test working for future.
On Mon, Jul 9, 2018 at 9:12 AM, John Beard wrote:
> Hi,
>
> Let's come back to the make test behaviour after v5, I think we'll
> need to discuss that separately. However, I think this does
Hi,
Let's come back to the make test behaviour after v5, I think we'll
need to discuss that separately. However, I think this does illustrate
why we need the tests to be runnable easily, otherwise they suffer
bit-rot, and then the tests are useless.
Looking at that change, the test is now
Hi
I guess we could add it to the qa target somehow? What I don't
particularyly like with this patch is that executing "make test" does not
check for dependency changes.
Back to the status about qa_geometry... it did pass a long time ago, doing
a bit of git bisect points at this commit as the
Hi,
Are the qa_geometry test supposed to all work?
When I run `qa_geometry`, I get 1160 errors like this:
error: in "ChamferFillet/Fillet": check { chainPoints.begin(),
chainPoints.end() } == { polyPoints.begin(), polyPoints.end() } has
failed.
Mismatch at position 0: [ 40 | 14 ] != [ 40 | 12
9 matches
Mail list logo
