Code coverage broken

[jacobwilliams]

Since the recent splitting of json_module into multiple files, the code coverage generation is currently broken. I temporarily disabled it in the unit tests just so they would pass, but it should be fixed.

[zbeekman]

To further complicate things, codecov.io just migrated to a config file based settings... the mapping/documentation between the old and new system is not 100% obvious, so I'll have to poke around a bit.

[jacobwilliams]

I started tinkering with it a bit...see codecov branch. Not finished. Whenever you get a free moment, you may have to clean up whatever mess I end up creating. :)

[zbeekman]

Haha fair. Yeah it looks like I won't be able to get to this this week. Hopefully next.

[jacobwilliams]

I think I've got it! I'll keep the ticket open and double check everything this week sometime.

[zbeekman]

Ok great. I'll try to get eyes on it when I can.

[jacobwilliams]

Not sure what codecov/changes — 2 files have unexpected coverage cha... error is all about in PR #195. I think perhaps because the changes are in the unit test files, which we aren't logging coverage for?

[zbeekman]

The codecov upgrade has been a bit of a dog's breakfast. This one message
was created because the #1 support request used to be hunting down changes
"that should not have appeared". In my opinion this "feature" is likely
going to cause 10x more support requests for them from people who know what
they're doing and are confused/annoyed/offended by this asinine error. Yes
my coverage changed, but, no, it is not unexpected and it most certainly is
not an error.
On Mon, May 2, 2016 at 10:21 PM Jacob Williams [email protected]
wrote:

Not sure what codecov/changes — 2 files have unexpected coverage cha...
error is all about in PR #195
#195. I think perhaps
because the changes are in the unit test files, which we are logging
coverage for?

—
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#174 (comment)

[zbeekman]

(Btw your suspicion is right, it's generated when coverage changes without
touching parts of the code where this happened, I.e. Doing anything with
unit tests)
On Mon, May 2, 2016 at 11:55 PM Zaak Beekman [email protected] wrote:

The codecov upgrade has been a bit of a dog's breakfast. This one message
was created because the #1 support request used to be hunting down changes
"that should not have appeared". In my opinion this "feature" is likely
going to cause 10x more support requests for them from people who know what
they're doing and are confused/annoyed/offended by this asinine error. Yes
my coverage changed, but, no, it is not unexpected and it most certainly is
not an error.
On Mon, May 2, 2016 at 10:21 PM Jacob Williams [email protected]
wrote:

Not sure what codecov/changes — 2 files have unexpected coverage cha...
error is all about in PR #195
#195. I think
perhaps because the changes are in the unit test files, which we are
logging coverage for?

—
You are receiving this because you commented.
Reply to this email directly or view it on GitHub
#174 (comment)

[jacobwilliams]

Ug. We need to disable the ability of Codecov to cause a test failure. I haven't had a chance to look into it further, but it looks like every commit now fails because of this. I don't like seeing all those red X's!

[jacobwilliams]

Maybe this is magically fixed now? My last commit to #206 has coverage tests passing...

[zbeekman]

yes, the pass/fail is trying to programmatically enforce coverage metrics i.e. certain minimum global coverage and coverage changes allowed by the commits, i.e. coverage can't decrease etc. Just a matter of finding the right settings and/or poking around the github project settings to turn off codecov pass/fail

[jacobwilliams]

Reduced target coverage to 70% for now. Closing issue, but we can fiddle with the settings some more if necessary.