Closed matttproud closed 12 years ago
Hmm. According to http://about.travis-ci.org/docs/user/getting-started/, that key must be set and to a valid configuration. Since there is no pure C or Go implementation listed, I presume the Erlang one will have minimal overhead. Is this OK? I've had this successfully working with github.com/matttproud/golang_instrumentation for some time now.
the lint checker for travis complains that the otp_release
key is not specified
The lint checker at least doesn't complain when I change language
to go
lint checker at http://lint.travis-ci.org/
So, if you want to take a look at this, I'd appreciate it:
1.) I am willfully ignoring the linter, since using erlang without a defined otp_release
causes no problems as evidenced on http://travis-ci.org/#!/matttproud/golang_instrumentation
.
2.) Using go
or ruby
, while it validates in the linter, actually produces anomalies when it comes time for building and testing. Merely compare builds 28
and 29
from the previous link, as Travis CI throws an internal stack trace.
Is this OK? It's what I'm using now for my own other Golang projects.
Merged in. Not a fan, but I understand that the systems are inconsistent.
Hey Matt, I've pushed it out to Travis-CI but it's failing to build because it can't find any tests http://travis-ci.org/#!/pomack/thrift4go/builds/1837313
OK. I'll file another pull request to get this fixed. Again, I am terribly sorry about this. Iteratively testing (working) Travis CI is a pain in the rear.
This commit adds support for Travis CI (continuous integration) in thrift4go such that commits against the master branch will trigger an automatic build and test run against the "release" branch du jour of Go. All that needs to happen after this gets incorporated is that thrift4go register itself with http://travis-ci.org/.
This would be immensely helpful in providing automatic verification of future changes. :-)