The owner of the repo can turn off notifications. Also you can set the pass or fail threshold for the percentage decrease so it won't give the x just from a small decrease, and you can set a general overall percentage that causes fail.
As far as coverage changing alot within a single PR (with no code changes to justify), this issue could come from the parralel builds and travis auto cancel feature. Travis will cancel a current build, if another build with the same commit history, but adding one additional commit is started. Travis is assuming you dont care about the previous build because you made recent changes. This can happen while only some of the parralel jobs sent a report to coveralls, then the other ones get cancelled. Looking at the coveralls report should clear things up. It has a sort by only files changed which should connect a report percentage to some code changes. I think we should turn off the notifications, they are a little unnecessary. The little pass or fail icon will still be at the bottom of the PR. On Tue, Sep 11, 2018, 9:48 AM Mats Wichmann <m...@wichmann.us> wrote: > > Every single pull request I get a message from Coveralls automated > checks that test coverage has decreased: > > > Coverage decreased (-1.08%) to 78.797% when pulling [Blah] into [Blah] > > > occasionally, I get a separate message which also tells me coverage has > increased, on the same PR.... > > For the one I sent just now, it is clear that coverage is completely > neutral... an extra version string was added to two lines so there > cannot be either an increase or decrease in code coverage since the > number of lines remains unchanged. > > Note that the actual report shows, as it should, no change: > > https://coveralls.io/builds/18935772 > > should those emails just be ignored, then? > > _______________________________________________ > Scons-dev mailing list > Scons-dev@scons.org > https://pairlist2.pair.net/mailman/listinfo/scons-dev >
_______________________________________________ Scons-dev mailing list Scons-dev@scons.org https://pairlist2.pair.net/mailman/listinfo/scons-dev