Thanks both for your quick answers,
I've implemented a solution to avoid the keyring issue by handling
properly the error, and catch it properly in test-all.R (as a better
control to deactivate CI integration tests if Zenodo token is invalid -
which is the case on CRAN).
https://github.com/eblondel/zen4R/commit/ff3c34a20b486e547bbc684a696cfe15eb8c37bd
<https://github.com/eblondel/zen4R/commit/ff3c34a20b486e547bbc684a696cfe15eb8c37bd>
zen4R has been resubmitted and Fedora CRAN builds are fine now.
Emmanuel
Le 03/09/2020 à 00:39, Gábor Csárdi a écrit :
On Wed, Sep 2, 2020 at 11:11 PM Dirk Eddelbuettel <e...@debian.org> wrote:
On 2 September 2020 at 23:59, Emmanuel Blondel (GMAIL) wrote:
[...]
| Your advice in order to solve that would be much appreciated,
Life, as they say, "is too short" so you could just comment-out the test.
Indeed, it is probably best not to run this test on CRAN's machines.
It probably fails because CRAN's keyring package is built without the
OS's keyring library, at least on Fedora.
But really, ideally you would not access the OS keyring in test cases,
except when running on a CI, or on the developers' machines.
Gabor
[...]
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
