Put 1.0-16 on CRAN? - Githubissues

pdeffebach commented 8 months ago

Is there a timeline for putting 1.0-16 on CRAN?

We use renv to manage our R packages and it seems to default to CRAN. We want to avoid adding another repo to our .lock file.

Thank you,

Peter

edzer commented 8 months ago

I was planning to do that in the next 1-2 months.

pdeffebach commented 8 months ago

Thanks.

As a side note: Do you have any insight as to why R downloads 1.0-16 when I don't have renv enabled and 1.0-15 when I do?

As far as I can tell, both sources point to CRAN, as in I don't have r-spatial-dev in my repos list. Can you tell me all the places where the 1.0-16 might be downloaded from?

edzer commented 8 months ago

No idea. 1.0-16 is not released, so not on CRAN; you must have done something special.

kadyb commented 8 months ago

@pdeffebach, maybe you use R-universe as repository? The 1.0.16 version is released there.

pdeffebach commented 8 months ago

Thanks!

@kadyb To be clear, is the binary up on R-universe? Or just the source? Because I think I had successfully avoided making my RA install xcode and developer tools, and yet she had version 16 installed.

On my personal mac, I tried downloading sf from the R-universe repo and it ran into errors about xcode developer tools. I am on an intel mac, not M1 though. Maybe it has a source for intel but a binary for M1?

I have also not figured out why sf is downloading from R-universe. That's still a mystery.

rsbivand commented 8 months ago

https://r-spatial.r-universe.dev/sf# currently says:

pdeffebach commented 8 months ago

Sorry, I'm a bit confused. I just assumed that a .tar.gz could contain source files as well. Any time I see .tar.gz that means it's the binary?

rsbivand commented 8 months ago

No. Source is always *.tar.gz, on Windows binaries are *.zip, on macOS *.tgz.

pdeffebach commented 8 months ago

Sorry to keep bothering you with this. But maybe the troubles my RA is having downloading sf are related to packaging issues you might be interested in.

I added sf@1.0-16 to my renv.lock with the repo source being r-universe.

"sf": {
  "Package": "sf",
  "Version": "1.0-16",
  "Source": "Repository",
  "Repository": "https://r-spatial.r-universe.dev",
  "Requirements": ...

My RA tried to instantiate the .lock file with renv::init and received the error (sorry for the screenshot)

She ran renv::deactivate() and re-started R. Then ran

install.packages('sf', repos = c('https://r-spatial.r-universe.dev/', 'https://cloud.r-project.org/'))

and received the same error.

Yet, somehow, she has sf-1.0-16 installed!

I have two questions:

Do you have any idea what might be happening? I've
Is there a way to check how my RA was able to download sf@1.0-16? I looked at installed.packages() and poked around in the folder things are installed in. I was not able to find any information about what repo was used.

edzer commented 8 months ago

I had a student who had massive problems with renv. My advice was not to use it.

r-spatial / sf

Put 1.0-16 on CRAN? #2347