[R] arrow R package 17.0.0.1 installs without GCS enabled on MacOS

[cklunch]

Describe the bug, including details regarding any error messages, version, and platform.

Installing the current version of arrow from CRAN on Mac is successful, but with arrow_with_gcs() == FALSE. I've tried the options on https://arrow.apache.org/docs/r/articles/install.html for custom installs, but haven't been able to get GCS capability enabled. If I downgrade to arrow 15.0.1, GCS connections are working fine.

I see the same behavior on Sonoma 14.7 with Intel, R version 4.4.0, and on Sequoia 15.1.1 with Apple M2, R version 4.4.2.

Looking at PR #42210 it looks like this may have been a development choice. If so, is there a simple way to install with GCS, or to enable GCS after installation?

Component(s)

R

[amoeba]

Hi @cklunch, sorry for the trouble. This is a bit of a pain point with the package.

Can you please try this and let us know if it works? From an R console:

Sys.setenv("NOT_CRAN" = "true")
install.packages("arrow", type = "source")

If that runs successfully, you should see GCS enabled once you load arrow and run arrow_info().

[cklunch]

@amoeba Thanks for the quick response! I'm getting more informative errors this way, here's the start of the install output:

trying URL '[https://cran.rstudio.com/src/contrib/arrow_17.0.0.1.tar.gz](https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcran.rstudio.com%2Fsrc%2Fcontrib%2Farrow_17.0.0.1.tar.gz&data=05%7C02%7Cclunch%40battelleecology.org%7Cd64564578acc435bbd3c08dd0e64a618%7Cf44d2ab390994d85998610165a8619f5%7C0%7C0%7C638682548174516801%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=ewiZoiPq7PSKt5s89ZaEN6771wyqx1rJTaMZQCFsd1A%3D&reserved=0)'
Content type 'application/x-gzip' length 4492506 bytes (4.3 MB)
==================================================
downloaded 4.3 MB

* installing *source* package ‘arrow’ ...
** package ‘arrow’ successfully unpacked and MD5 sums checked
** using staged installation
*** pkg-config not found.
*** OpenSSL not found
*** Found local C++ source: 'tools/cpp'
*** Building libarrow from source
    For build options and troubleshooting, see the install guide:
    [https://arrow.apache.org/docs/r/articles/install.html](https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Farrow.apache.org%2Fdocs%2Fr%2Farticles%2Finstall.html&data=05%7C02%7Cclunch%40battelleecology.org%7Cd64564578acc435bbd3c08dd0e64a618%7Cf44d2ab390994d85998610165a8619f5%7C0%7C0%7C638682548174533581%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=Z6oY%2FA6d3fa0Cjf%2FPjaD9mxLsw0A01Zj%2BTzGy5toJc4%3D&reserved=0)
**** cmake 3.26.4
**** pkg-config not installed, setting ARROW_DEPENDENCY_SOURCE=BUNDLED
****  S3/GCS  support  requires version >= 1.0.2 of openssl-devel (rpm), libssl-dev (deb), or openssl (brew) ; building with  ARROW_S3=OFF and ARROW_GCS=OFF
**** arrow

Consistent with that message, arrow_info() after install has both GCS and S3 disabled:

Arrow package version: 17.0.0.1

Capabilities:
               
acero      TRUE
dataset    TRUE
substrait FALSE
parquet    TRUE
json       TRUE
s3        FALSE
gcs       FALSE
utf8proc   TRUE
re2        TRUE
snappy     TRUE
gzip       TRUE
brotli     TRUE
zstd       TRUE
lz4        TRUE
lz4_frame  TRUE
lzo       FALSE
bz2        TRUE
jemalloc   TRUE
mimalloc   TRUE

Memory:
                 
Allocator mimalloc
Current    0 bytes
Max        0 bytes

Runtime:
                       
SIMD Level          none
Detected SIMD Level none

Build:
                                   
C++ Library Version           17.0.0
C++ Compiler              AppleClang
C++ Compiler Version 16.0.0.16000026

OpenSSL 1.1.1 is on my machine, so it looks like the arrow install can't find it.

[amoeba]

Thanks. Can you please try install OpenSSL@3 with

brew install openssl@3

And then running a similar command to above (with one tweak):

Sys.setenv("ARROW_R_DEV"=TRUE, "NOT_CRAN" = "true")
install.packages("arrow", type = "source")