-
Notifications
You must be signed in to change notification settings - Fork 60
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Error with vroom 1.6.4 on MacOS but not Ubuntu or Windows #519
Comments
vroom's GitHub Actions checks are all green atm and that includes macOS 12.7: https://github.com/tidyverse/vroom/actions/runs/6381873068/job/17319275234 So that makes me wonder if there's something specific to your usage or GHA or ??? Can you link to the package and the GHA logs? |
Thanks for your reply! The package is: https://github.com/OHDSI/DataQualityDashboard and GHA logs are here: https://github.com/OHDSI/DataQualityDashboard/actions/runs/6410263439/job/17404120664. However, I've looked further into this and have some more information:
btw, I am on the latest version of readr, 2.1.4 Thoughts? |
Also having a similar issue
Cannot be consistently reproduced, always broken inside checks while compiling vignette, always always works outside of checking environment. |
I also have the same "bad value" issue as above, I had to downgrade from 1.6.4 to 1.6.3 to get things working again.
|
Same here:
Where the file can be anything. I cannot consistently reproduce the error easily, it happens after doing a bit of work in the environment, so it must depend on other factors. A package I am preparing for Bioconductor consistently fails building on MacOS, but passes on Linux. Just like @pitkant, it only happens with 1.6.4. Downgrading to 1.6.3 alleviates the issue.
|
OK, I think I have managed to create a reproducible example:
In this case initialising I can also confirm the error appears only with
|
Refining @MarekGierlinski's reprex: # RSQLite does use cpp11
con <- RSQLite::datasetsDb()
x <- vroom::vroom(
I("a\tb\n1.0\t2.0"),
delim = "\t"
) |
One quick fix seems to be to build RSQLite from source with So I'm thinking that somehow CRAN RSQLite (built with "old" cpp11, not sure which exact version, but pre-0.4.6) is somehow incompatible with CRAN vroom (built with "new" cpp11 0.4.6) |
lldb backtrace. Seeing vroom compilation units should be getting their own preserve list due to r-lib/cpp11#331 so I'm not entirely sure how an "old" version of cpp11 that came with RSQLite could be affecting this Exact failure line Ends up failing in the
|
Needed until vroom issue [#519](tidyverse/vroom#519) is resolved
I was also getting the same bad value vroom error as mentioned above. I was able to work around it by adding the command readr::local_edition(1) before the lines that were causing the error as mentioned under fenr Bioconductor/Contributions#3017. |
In my R package I have several GitHub Actions testthat tests on MacOS (12.7) failing right now with an error like this one:
The same tests pass on Ubuntu, which is also using vroom 1.6.4. For some reason my Windows tests are staying on vroom 1.6.3 in GitHub Actions but when I run locally on Windows with 1.6.4 they work fine. Same tests were passing on MacOS a couple days ago on vroom 1.6.3. I tried specifying vroom <= 1.6.3 in my DESCRIPTION file but got a dependency resolution error on Mac and Ubuntu (not super familiar with how to approach such issues, though, so it's possible I did something wrong there).
I hope this is enough info for you all. Please let me know if not and I'd be happy to provide more detail. Thanks!
The text was updated successfully, but these errors were encountered: