intelligent-agent / redeem

Firmware for Replicape
http://wiki.thing-printer.com/index.php?title=Redeem
GNU General Public License v3.0
36 stars 44 forks source link

Add Travis CI #156

Closed ThatWileyGuy closed 5 years ago

ThatWileyGuy commented 5 years ago

With these changes, Travis CI can now validate commits and pull requests. There are four jobs:

The formatting scripts are (lovingly) stolen from what Richard came up with for CircleCI. I originally was trying to add test support to his config, but getting C++ tests to work on CircleCI seemed a lot harder than Travis.

Travis's results from when I pushed this commit to my own Redeem repo are at https://github.com/ThatWileyGuy/redeem/runs/47338987

Wackerbarth commented 5 years ago

I have no preference for Circle-CI over Travis. I used Travis on other projects and switched to Circle-CI only because someone had started to use it on the project.

However, I do have an issue with the differences in the output of clang-format when run on various platforms. The generated style on this Travis run differs from that on the Circle-CI run.

I also ran the test of the file in question on the BBB. It produced the same result as the Circle-CI run. (Both consider that d88945e is introducing an error.)

So, before switching, the Travis configuration will need to be modified to match the other results.

ThatWileyGuy commented 5 years ago

Yes, I suspect we need to lock the version of clang-format. I'll take a look this evening.

ThatWileyGuy commented 5 years ago

I have locked clang-format to version 6.0. I believe this is the in-box default for Ubuntu Bionic (18.04).

Wackerbarth commented 5 years ago

@ThatWileyGuy - clang-format version 6.0 is what appears on the BBB. But, as I read it, that was also the version that Travis had selected. So, I am confused as to why one works one way and the other in another way.

ThatWileyGuy commented 5 years ago

@ThatWileyGuy - clang-format version 6.0 is what appears on the BBB. But, as I read it, that was also the version that Travis had selected. So, I am confused as to why one works one way and the other in another way.

The build logs I linked to from the original PR description show this in job 2.4: clang-format version 7.0.0 (tags/RELEASE_700/final)