@josi If anyone out there wants to get paid to work on an open source project and cares about an R-based IDE stack I’d welcome contributions to {debugadapter} and would be happy to help co-author an ISC proposal to work on it.
What does #rlang need to help folks easily migrate to #codeberg ?
Some ideas:
codeberg equivalent to r-lib/actionsmigration guideusethis::use_codeberg() - create new repo or provide instructions for migrating if already on GitHub, similar to use_githubusethis::use_codeberg_ci() - equivalent to use_github_action
anything else that would help people make the jump?
`R:Pak` is pretty good! Really impressed with the installation messaging, especially having not used R for a while!
#Seagl2024 presents: Culling Seagull Records: Using R to Curate Community Science Checklists
By: Richard Littauer
Saturday, 11/09 13:30PST
Category: Everything Else
Link: https://pretalx.seagl.org/2024/talk/7QPK9S/
#FOSS #FLOSS #RLang #science
I'm so desperate I actually posted a question on stack overflow about this 
Migrated my first #rlang package repo from GitHub to #Codeberg
https://codeberg.org/dgkf/options
In the process:
set up mirroring back to GH added a notice of migration with rationale adapted my most-used r-lib/actions (R CMD check, test, coverage, pkgdown) set up pages publishing
Next up, publishing a release from the new project home.
@embiggenData pretty much every package I write these days has little internal typed flavors of vapply. The convenience of lapply, the prescriptiveness of vapply, but without the clunkiness of having to always provide a prototype value.
vcapply (character)
vlapply (logical)
viapply (integer)
A tricked picked up when mucking around in many r-lib packages.
How many R package libraries is too many? As an experiment, I installed all my packages into separate directories - one package per library.
My .libPaths was ~400 paths long.
and… everything seems to be okay?
The purpose of this little experiment is for composing libraries dynamically, without relying on OS-specific features like symlinking.
Am I setting myself up for major performance issues or odd bugs down the road?
CRAN seems to be down… and although it’s a bit inconvenient, I’m also just amazed that this is the first time I’ve been impacted by CRAN downtime.
Maybe I’ve just been lucky so far, but upon reflection I’m pretty impressed it has a such a good uptime track record.
Want to use {pak} as backend for {renv} actions? Set the variable RENV_CONFIG_PAK_ENABLED = TRUE https://github.com/rstudio/renv/issues/1210 #renv #pak #reproducibility #RLang
@defuneste @johnmackintosh @stevegis_ssg @trevin_flick i tried #rlang for a while, but it never got traction, and yeah now there’s the pkg name confusion
@lionel recently introduced me to his {codegrip} package, a wonderful editor helper for reshaping function calls for RStudio & emacs.
Today I added it as a code action to {languageserver}! (currently just in a dev branch)
What are your favorite R non-standard evaluation syntax “grammars”?
What do I mean a syntax grammar? I’m thinking of things like {rlang} tidy evaluation, {glue} format strings, {ggplot2} layer building, {patchwork} operators for arranging plots, {data.tables} index operator.
Some of these are more exotic syntax extensions than others, but they all provide interfaces as little grammar ecosystems.
What works well? What doesn’t? What makes a grammar have the right “feel”?
How in the hell are we still living in a universe where there is no *complete* replacement for #Mplus?
By now there should be some #rlang code, or even #Python, that takes Mplus scripts as input and just does them. And yet, no.
Everybody I know who uses Mplus says -- with complete seriousness (!!) -- that there are R packages that implement some of its functions, but they still need Mplus to do everything.
What am I missing? Why?
@rstats #wtf #science #stats #sem #structuralequationmodeling
@djnavarro Sorry to be one of those people, but this is why I like #Nix for creating dev environments, all the tools can be defined alongside the project and independent of the OS and recreated exactly at any time, there's even an #RLang package that integrates R and Nix :)
@yabellini @code4thought Enhorabuena, @yabellini . Con esas cifras que mencionas sobre #rlang , quizá el índice #tiobe debiera mover R hacia 
(al *codiciado top 20* )
https://www.tiobe.com/tiobe-index/
So, everything was fine until #Rlang decided to re-install itself but into my user libarary on the cluster
#UCSB is hiring a #SysAdmin for #datascience environments. Boosts appreciated.
