sysreqsdb icon indicating copy to clipboard operation
sysreqsdb copied to clipboard

Coverage of the database

Open maelle opened this issue 7 years ago • 4 comments

Maybe something you've already done? If not, could this be useful?

Idea: parsing all SystemRequirements fields of CRAN packages and seeing how many of them map to an entry of the database. Not sure where the result could be published, but it could be used as a reminder to maintainers (interested in this) to add an entry to the database when needed?

maelle avatar Sep 18 '18 14:09 maelle

http://crandb.r-pkg.org/-/sysreqs?keys=[%22curl%22,%22xml2%22]

{
"curl": "libcurl: libcurl-devel (rpm) or\nlibcurl4-openssl-dev (deb).",
"xml2": "libxml2: libxml2-dev (deb), libxml2-devel (rpm)"
}

Missing ones: https://sysreqs.r-hub.io/missing

gaborcsardi avatar Sep 18 '18 19:09 gaborcsardi

Maybe the missing ones could have the maintainer as well? And we could publish it daily? Recalculating it takes some time.... I am not sure if I want to email people about these, probably not.

gaborcsardi avatar Sep 18 '18 19:09 gaborcsardi

Wow so cool that there's a direct URL for the missing ones!

You could email webdriver & shinytest authors 😸

I'm wondering whether there could be a link to that page in the README, with a note "is your package in this list? contribute to the database!". 🤔 And/or a blog post (an R-hub blog?) about system dependencies in general, with this database in particular. In rOpenSci development guide, we refer authors to this README.

maelle avatar Sep 19 '18 07:09 maelle

You could email webdriver & shinytest authors 😸

LOL.

Good idea about the link, but this list is assembled on the fly, and this is costly and heavy on the sysreqs machine, so I don't want to link to it just yet. I would need to cache it first. Also, if we link to it, then it should be HTML, not a JSON dump, so it will be a different API end point.

gaborcsardi avatar Sep 19 '18 08:09 gaborcsardi