On Tue, Sep 23, 2025 at 06:17:27PM +0100, Rebecca N. Palmer wrote: > The upstream commit log suggests that this was a change on the server that > this test connects to, and that this might happen quite often: 1.45.8, > 1.45.10 and 1.45.11 are all plausibly relevant. > https://github.com/Bioconductor/GenomeInfoDb/commits/devel/
Is this only a test breakage, or is the package itself also broken? Is/Should r-bioc-genomeinfodbdata be used for tests? I'm asking these stupid questions, since a frequently broken test should be dropped while a test indicating actual breakage relevant for the package would be useful (but if it breaks frequently the package might not be suitable for stable releases). According to debian/changelog there once was a needs-internet autopkgtest restriction, which seems to have gotten lost at some point. This needs fixing if the test stays and needs internet access. cu Adrian
