Uh oh!
There was an error while loading. Please reload this page.
Improve Travis / tox setup + coverage - #851
Conversation
d0aad44 to
cef19ebCompareMainly inspired by using pytest directly in tox, for easy test selection etc.
cef19eb to
c3b7066CompareAdded coverage reports via codecov, which can be seen at https://codecov.io/gh/pallets/click/branch/improve-travis-tox. |
16b780d to
b61c167CompareThis hopefully works around `complex/complex/commands/__init__.py` and `termui/termui.py` showing up in `examples` (at the root).
blueyed
commented
Sep 28, 2017
@untitaker@axu2 |
blueyed
commented
Sep 28, 2017
Also please authorize/install codecov on https://codecov.io/gh/pallets/click/. |
untitaker
commented
Oct 1, 2017
via email
I wouldn't include codecov, particularly its notification settings are
extremely buggy in my experience. I feel like it's more maintenance effort than
it brings insight …On Thu, Sep 28, 2017 at 02:14:18PM -0700, Daniel Hahler wrote:
@untitaker@axu2
Can you take a look, please?
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#851 (comment) |
blueyed
commented
Oct 6, 2017
@untitaker
Do you mean that you have or want to delete the duplicated comments then? If this is the issue you mean it might be solvalble by @codecov just fixing it in the end.. ;) From my point of view codecov provides a lot of valuable insights. |
ThiefMaster
commented
Oct 6, 2017
IMO even just a single codecov comment on every PR is somewhat annoying. I like getting notifications from GitHub for new PRs and comments since that's convenient to see if there's anything to comment on, a PR to review, etc. But codecov comments? I'm tempted to overcome my laziness and configure my email filter to send the notifications about them straight to the trash folder. But that aside, IMO these comments are more noise than adding useful comment. |
blueyed
commented
Oct 6, 2017
@ThiefMaster Otherwise you could also just block the bot user (https://github.com/codecov-io), and be done with it I guess. |
blueyed
commented
Oct 6, 2017
Especially for PRs codecov's reporting is valuable, since it will tell you if coverage drops and if the diff is covered etc. |
untitaker
commented
Oct 6, 2017
via email
If codecov were about to "just fix" anything they would have done that a long time ago. …On 6 October 2017 21:38:40 GMT+02:00, Daniel Hahler ***@***.***> wrote:
@untitaker
What do you mean by "extremely buggy"?
What I am seeing sometimes is multiple comments, also only one should
be there and edited.
> maintenance effort
Do you mean that you have or want to delete the duplicated comments
then?
If this is the issue you mean it might be solvalble by @codecov just
fixing it in the end.. ;)
>From my point of view codecov provides a lot of valuable insights.
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#851 (comment) -- Sent from my Android device with K-9 Mail. Please excuse my brevity. |
blueyed
commented
Oct 6, 2017
@untitaker |
untitaker
commented
Oct 6, 2017
via email
Hence my suggestion to not have any such service. …On 7 October 2017 00:09:43 GMT+02:00, Daniel Hahler ***@***.***> wrote:
@untitaker
It is still the best service for it, isn't it? coveralls is much worse
and there is no other.
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#851 (comment) -- Sent from my Android device with K-9 Mail. Please excuse my brevity. |
blueyed
commented
Oct 6, 2017
What about the benefit to see if a PR covers its diff? |
blueyed
commented
Oct 6, 2017
And even in general, I think you can grasp a lot of useful info from https://codecov.io/gh/pallets/click/tree/improve-travis-tox/click already. |
untitaker
commented
Oct 6, 2017
via email
I never disputed that the information provided by the service would be useful. But for that it would have to work properly, and that just has not been my experience at all. Flaky reports and/or random build failures because their weird CLI tool was unable to contact S3 or whatever.
Bot comments are a non-starter, way too spammy, …On 7 October 2017 00:13:51 GMT+02:00, Daniel Hahler ***@***.***> wrote:
And even in general, I think you can grasp a lot of useful info from
https://codecov.io/gh/pallets/click/tree/improve-travis-tox/click
already.
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#851 (comment) -- Sent from my Android device with K-9 Mail. Please excuse my brevity. |
untitaker
commented
Oct 6, 2017
via email
Ah, I see you used travis-retry for that purpose, fair enough.
I guess we can merge this and also immediately block the bot org-wide. That should do it. …On 7 October 2017 00:13:51 GMT+02:00, Daniel Hahler ***@***.***> wrote:
And even in general, I think you can grasp a lot of useful info from
https://codecov.io/gh/pallets/click/tree/improve-travis-tox/click
already.
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#851 (comment) -- Sent from my Android device with K-9 Mail. Please excuse my brevity. |
blueyed
commented
Oct 6, 2017
@untitaker |
blueyed
commented
Oct 6, 2017
btw: it might be better to use pytest-cov here, since it would also track subprocess automatically - not sure if that is used in tests?! |
untitaker
commented
Oct 10, 2017
Thanks! Codecov seems to be already activated. |
Mainly inspired by using pytest directly in tox, for easy test selection
etc.