I have found an interesting approach in the module...
# team-maintainers
i
I have found an interesting approach in the modules repo that I want to discuss with the hivemind before normal review. This new module uses git to get module files from GitHub before running, as the software is not distributed in any Conda repository. While I don't see any big red flags about this approach, this is certainly non-standard (I have not seen this before) and might cause issues for people who need to pre-download containers (since it skips Nextflow's caching). What do you think?
👀 1
m
Oh
that's a no go for me
r
this wouldn't work for offline usage I think
party this 5
m
but I like the craftyness
j
IMO it's pretty trivial to package stuff into a Bioconda package, so I'd much rather do that than depend on a tag in a github repo that may or may not change
party this 3
Our wonderful @James Fellows Yates even wrote the definitive guide! https://bioconda.github.io/tutorials/2024-adding-bioinformatic-software-to-bioconda.html We could point them at it and offer support if needed?
💯 1
i
Sure, on it
j
You can rename tags so that system makes no sense from the point of reproduicilbity IMO
party this 1
n
or at the very least it should just be put in a container 😅
100 rainbow 1
i
I have put a red mark on the PR so nobody accidentally approves it, let's see where it goes
❤️ 5
m
IGOR for president
s
yeah, bioconda seems the way to go here