Mail says PREPERROR but build log says SUCCESS

[RafaelEduardoDiaz]

Hello, I would like to request help with this problem, similar to how it happened to another user. The logs say successful status, but the email says error, I can't understand what happens. Any help I appreciate.

Full log here: https://builder.r-hub.io/status/Bayeshmmcts_0.0.1.tar.gz-0e1e179e390a451e8465d9bd96ff79c7

[dmenne]

I have exactly the same problem, my package with rstan dmenne/gastempt even passes CRAN incoming, but fails on winbuilder. I assume it is because we both have a second NOTE because of the use of GNU extensions.

This is a long standing issue, its always a waste of time to wade through pages of output just to note the "success" at the end. Please fix it, r-hubbers! I never have managed to get a green output despite having 4 packages on CRAN.

[gaborcsardi]

You need to show some examples.

[dmenne]

See the example of the original poster, or

https://artifacts.r-hub.io/gastempt_0.5.4.tar.gz-6e76bbc708114f5984e5fcf1db6c9bbd/gastempt.Rcheck/tests_i386/

[gaborcsardi]

The original poster referred to a different build for PREPERROR than the linked one. The linked one is a NOTE, so green.

[gaborcsardi]

This is probably because we get a truncated log from Jenkins, so we don't see the actual R CMD check run, so we assume there wasn't one. We would need to make multiple calls to Jenkins for very long build logs. Here: https://github.com/silas/node-jenkins#jenkinsbuildlogoptions-callback

Will be fixed soon.

[dcnorris]

In case additional examples are useful ... https://builder.r-hub.io/status/original/precautionary_0.2-2.tar.gz-45f46cdb9ab546d899c536ed1b3d16b7

[samFarrellDay]

This is happening on both packages I maintain - here is an example from each: https://builder.r-hub.io/status/ParBayesianOptimization_1.2.5.tar.gz-bc6d47d807204fd1afe730114bc8d51b https://builder.r-hub.io/status/miceRanger_1.4.0.tar.gz-af30439a9ff145aba7973788f0b2a783

R CMD says OK, status is Success, email says PREPERROR like the picture above.

[njtierney]

I also have the same issue:

https://builder.r-hub.io/status/brolgar_0.1.1.tar.gz-51b120fcc8b5464cbfd955e4c1111857

[laresbernardo]

Same issue here: linux-x86_64-centos-epel: https://builder.r-hub.io/status/original/lares_5.0.1.tar.gz-53c9d9d388ca459e832e0068e0634e0b

[ycphs]

Same here

windows

[bbolker]

As far as I can tell this hasn't been resolved yet ... ? e.g.

[gaborcsardi]

Indeed has not.

[DRJP]

Hi.

I also appear to be hitting up against the same bug - always when rhub checks on Ubuntu (I've no problems when rhub checks on Fedora or Windows). The build log can be found here and the artefacts can be found here.

What are other package developers hitting this problem adding in their cran-comments.md files, because cran-summary() is returning "Error: Build failures on platforms: ubuntu-gcc-release" ?

Best regards David

[jonas-hag]

I have the same issue that I get an e-mail with PREPERROR but the status shows success (and RMD check run). See the logs: HTML, text and artifacts. On Windows, I don't have any problems.

I've found that in my logs there are some errors mentioned:

#> Failed with error: ‘there is no package called ‘shiny’’

this seems to be due to a strange DESCRIPTION file by htmlwidgets but I assume it doesn't matter as on windows it's ok and

16781 docker rm longmixr_1.0.0.tar.gz-0c8a446b4024426d8eca7058a01c2c51-3 16782#> Error: No such container: longmixr_1.0.0.tar.gz-0c8a446b4024426d8eca7058a01c2c51-3

maybe this is the reason?

Thanks for this great service! Please let me know if you need more info or if I can otherwise help to solve this issue.

[bbolker]

Still having this problem with glmmTMB (33 minute build) ...

[dmenne]

I gave up using r-hub...

[ntamas]

I'd like to add igraph to the list of packages where the mail says PREPERROR but the logs indicate success; see here.

Most likely caused by these lines at the end:

#> docker rm igraph_1.2.11.9169.tar.gz-5ebab37fc75e4ddb8b7a44f39b44f050-1

#> Error: No such container: igraph_1.2.11.9169.tar.gz-5ebab37fc75e4ddb8b7a44f39b44f050-1

true

docker rm igraph_1.2.11.9169.tar.gz-5ebab37fc75e4ddb8b7a44f39b44f050-2

#> igraph_1.2.11.9169.tar.gz-5ebab37fc75e4ddb8b7a44f39b44f050-2

docker rm igraph_1.2.11.9169.tar.gz-5ebab37fc75e4ddb8b7a44f39b44f050-3

#> Error: No such container: igraph_1.2.11.9169.tar.gz-5ebab37fc75e4ddb8b7a44f39b44f050-3

Two out of the three docker rm calls return Error: No such container:. Happens only with linux-x86_64-rocker-gcc-san.

[gaborcsardi]

No, the reason is https://github.com/r-hub/rhub/issues/448#issuecomment-793776145, i.e. the log is too long and our internal scheduler truncates it, so our app thinks that it stopped there.

No need to add more "me too" comments, thank you! :) Also no need to ask if this is still a problem, if this issue is open then it is a problem. :) Here is a somewhat longer explanation why: https://github.com/golang/go/wiki/NoPlusOne/

[gaborcsardi]

This issue is about the previous R-hub system, it does not apply to the new system, so I am closing it now. Please see https://r-hub.github.io/rhub/ for the new system, R-hub v2.