Should meta.strandeness exists or not? :thread:
# team-maintainers
m
Should meta.strandeness exists or not? 🧵
🔥 3
not permitted via linting, could be done via args I assume, but could be misleading if it suddenly vanish
m
I don't remember where the discussion was, but it was decided no. I asked about this when implementing the linting for meta keys
m
ok, then can you approve the linked PR as well please?
👀 1
l
I have the same issue with github.com/nf-core/modules/pull/12625 Should I also remove it? Simon wasn't for it.
m
For me, such breaking changes should be properly documented
I have no issues doing this, but we need to notify people and provide them with ways to fix it
A simple blogpost showing a config files with the ext.args to use would do the trick for me
In my opinion the linting is always right, if we have issue in a modules, due to linting, we either fix linting or fix the modules
But we should be more aware about such changes, and be more explicit about it
This conversion really don't need to happen since a decision was already made
m
I made an effort to start fixing modules once linting was implemented, but there was a long list to fix. My plan is to get to more at the hackathon.
m
I do think a blog post would help more, I totally forgot we decided that
You are not alone Mahesh, let us share the burden
m
The list keeps changing so I'm using the find command here: github.com/nf-core/modules/issues/8716
I thought there was an actual list, but I realise it's just the find command as a pixi task locally.
s
I think it was one of those decisions that was just made
I specifically said to keep it
But obviously my nemesis Dr Evil (@James Fellows Yates) said to get rid of it. https://nfcore.slack.com/archives/C043UU89KKQ/p1774078044671669
🤣 1
m
@Simon Pearce are your nemesis only Brits? Are you secretly French?
🇲🇫 1
🤣 3
a
One of the issues here is that we lack a coherent versioning system in modules. Makes it very hard to give users an easy way of seeing a module has breaking changes
💯 1
j
Are there instances where strandedness should be an input rather than arg?
One of the issues here is that we lack a coherent versioning system in modules. Makes it very hard to give users an easy way of seeing a module has breaking changes
These days I always preview a diff before updating a module because there have been so many changes to input structure for e.g. samtools modules
j
I've always previewed it, is that so bad?
party this 1
m
I always check
I don't trust other people
And me even less