Identify revdep issues

[hadley]

There are 229 new failures, listed below. This is too many, and suggests that we've accidentally broken something important (hopefully we've only broken 1 or 2 things that are used heavily not a bunch of small things). The first step is to find the common themes in the failures.

@tidyverse/ggplot2 can you please let me know if you have time this week to read some fraction of the failure reports (in <revdep/README.md>) and report back? I'll split the packages up over everyone who's willing to help.

Googlesheet of failures at https://docs.google.com/spreadsheets/d/1UZIFwt3j3XjnRGOMNJZvpal_CAj7Hz8aNXJt41m3JjM/edit#gid=0 — when you start exploring a failure, first put your name in the "who" column to make sure multiple people don't try understanding the same package at the same time.

[batpigandme]

I can read. At the very least, I can do some winnowing down before more expert eyes are needed. 🙂

[thomasp85]

I can put in some time

[batpigandme]

~97 of them have

Warning: replacing previous import ‘dplyr::vars’ by ‘ggplot2::vars’ 

[yutannihilation]

FWIW, my package (gghighlight) fails because it assumes an aes mapping consists of symbols, not quosures. (But, I'm OK because I plan to overhaul the implementation after the release of this ggplot2 2.3.0.)

[hadley]

@lionel- can you think of some hack to avoid the dplyr::vars() + ggplot2::vars() namespace clash? (In an ideal world people wouldn't @import dplyr and @import ggplot2 but clearly there are ~100 packages that do this, and getting them all to change is going to be too much work)

[hadley]

What's the best way to collaboratively organise these issues? Would a google doc make sense?

[karawoo]

I can read some as well, and a google doc works for me.

[thomasp85]

Maybe a google sheet where people can group issues into possible duplicates and others can begin tackling them...

[hadley]

Duh, yes, googlesheet is perfect

[hadley]

Googlesheet at https://docs.google.com/spreadsheets/d/1UZIFwt3j3XjnRGOMNJZvpal_CAj7Hz8aNXJt41m3JjM/edit#gid=0 — when you start exploring a failure, first put your name in the "who" column to make sure multiple people don't try understanding the same package at the same time.

@batpigandme can you please work on marking the packages with the "dplyr-ggplot2 vars clash"?

[hadley]

@lionel- @batpigandme Now tracking potential fix ideas for vars clash at #2585

[hadley]

Another common problem: looks like we accidental broke geom look up in qplot(). So any issue with could not find function "geom_text" (or similarly) is likely to be #2586

[clauswilke]

I see several warnings about an openMP clash, e.g.:

    > # Use image as plot background
    > p <- ggplot(iris, aes(x = Sepal.Length, fill = Species)) + geom_density(alpha = 0.7)
    > ggdraw() +
    +   draw_image("http://jeroen.github.io/images/tiger.svg") +
    +   draw_plot(p + theme(legend.box.background = element_rect(color = "white")))
    OMP: Error #15: Initializing libomp.dylib, but found libomp.dylib already initialized.
    OMP: Hint: This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see http://www.intel.com/software/products/support/.

This seems to happen with packages that work with images, and it may be a problem with the testing environment rather than indicating a real problem?