search

daattali / ggExtra

📊 Add marginal histograms to ggplot2, and more ggplot2 enhancements
http://daattali.com/shiny/ggExtra-ggMarginal-demo/
Other
383 stars 45 forks source link

Test under ggplot2 v3.1.0 #135

Closed crew102 closed 5 years ago

crew102 commented 5 years ago

This PR adds ggplot2 version 3.1.0 to the list of ggplot2 versions that ggExtra is tested under (partly in preparation for a PR for #90 which is on the way).

daattali commented 5 years ago

Travis is finally happy! Do you think this is ready to look at now?

crew102 commented 5 years ago

so weird - nothing changed between 8d0c88d and dd303a1 that would have resulted in it going from fail -> pass, but I'll take it. Should be ready to look at.

daattali commented 5 years ago

Thanks for doing this, and trying to deal with the constant build fails!

The vector of package versions c("2.2.0", "2.2.1", "3.0.0", "3.1.0", "latest") seems to be reused in a few places and will be forgotten to be updated at some point in the future - can it be defined in one place?

crew102 commented 5 years ago

Good call

crew102 commented 5 years ago

No idea why this isn't passing. It passes for me locally (in docker container, that is). It appears to have something to do with the install of the dev version of ggplot2, though I don't see how that could be the case and have only 2 of the tests fail. Let's let this PR sit for a while and try later. I suspect this will resolve itself.

daattali commented 5 years ago

As the old saying goes, if at first you don't succeed... sit around and do nothing.

I wish there was a way to specify "pass if the two images are 99.99% identical" , like a fuzzy match.

crew102 commented 5 years ago

I think when it comes to Travis that is indeed how the saying goes

crew102 commented 5 years ago

K, it just passed (again, no idea why) so we should be all set.

...Github's really shaming me with all these force-push notifications isn't it?

daattali commented 5 years ago

Failing again..... is it always the ggplot2 dev version that fails?

crew102 commented 5 years ago

I'm not sure...It seems totally random at this point when it will fail...It's gone from pass -> fail and back again several times with the only change in the code happening in the after_script phase of the Travi build (which shouldn't affect the build status). The last commit I made wasn't even a change per se - I just split a single commit into two separate commits and it went from pass to fail.

It seems to alternate between pass/fail for no apparent reason. I'm having trouble getting the detailed results of the test output as well, which is making this a bit more complicated.

crew102 commented 5 years ago

I bet this next one passes:

crew102 commented 5 years ago

Now it's failing but all tests are passing. asdfadf

crew102 commented 5 years ago

K that last one failed b/c of a non-portable file name.

crew102 commented 5 years ago

Looks like Travis hasn't been using the latest commit for the builds. For example, the last build used https://github.com/daattali/ggExtra/commit/2d174328c0cdd44e1880dbdd8f0462f67a16371e instead of https://github.com/daattali/ggExtra/pull/135/commits/57408cde29ae63a0c42eabcc6c12a50e86948f22.