Could you please add inet_pton existence check in configure.win too like it was stated in #5019 (comment)?

@StrikerRUS Thanks for the reminder. I've added that in ff8c604. Could you please help to check?

BTW, do you have any idea why some actions of R packages with gcc and R4.1 hangs?

/gha run r-solaris

/gha run r-valgrind

/gha run r-valgrind

It looks like the valgrind build is being triggered by comments correctly

For example, https://github.com/microsoft/LightGBM/runs/6122965595?check_suite_focus=true

But is then failing prior to even building the R package.

https://github.com/microsoft/LightGBM/runs/6122967527?check_suite_focus=true

/usr/bin/git init /__w/LightGBM/LightGBM
  Initialized empty Git repository in /__w/LightGBM/LightGBM/.git/
  /usr/bin/git remote add origin https://github.com/microsoft/LightGBM
  Error: fatal: unsafe repository ('/__w/LightGBM/LightGBM' is owned by someone else)
  To add an exception for this directory, call:
  
  	git config --global --add safe.directory /__w/LightGBM/LightGBM
  Error: The process '/usr/bin/git' failed with exit code 1[28](https://github.com/microsoft/LightGBM/runs/6122967527?check_suite_focus=true#step:4:28)

I suspect that was just missed in #5152. Sorry about that! Will put up a PR shortly.

/gha run r-solaris

Workflow Solaris CRAN check has been triggered! 🚀
https://github.com/microsoft/LightGBM/actions/runs/2292821204

solaris-x86-patched: https://builder.r-hub.io/status/lightgbm_3.3.2.99.tar.gz-66562bfd23ac42ff942995a915f1b396
Reports also have been sent to LightGBM public e-mail: https://yopmail.com?lightgbm_rhub_checks
Status: failure ❌.

/gha run r-valgrind

Workflow R valgrind tests has been triggered! 🚀
https://github.com/microsoft/LightGBM/actions/runs/2292822851

Status: success ✔️.

The r-solaris check fails with the following information

/usr/bin/ld: cannot find -lgfortran
collect2: error: ld returned 1 exit status
make[1]: *** [/usr/share/R/share/make/shlib.mk:10: Matrix.so] Error 1
make[1]: Leaving directory '/tmp/Rtmph9P7SX/R.INSTALL58396355090c/Matrix/src'
ERROR: compilation failed for package ‘Matrix’
* removing ‘/usr/local/lib/R/site-library/Matrix’

/gha run r-solaris

Workflow Solaris CRAN check has been triggered! 🚀
https://github.com/microsoft/LightGBM/actions/runs/2303038422

solaris-x86-patched: https://builder.r-hub.io/status/lightgbm_3.3.2.99.tar.gz-4f49e421e97542b4876d80c696893832
Reports also have been sent to LightGBM public e-mail: https://yopmail.com?lightgbm_rhub_checks
Status: success ✔️. (Fake status set manually to pass CI checks. Actually, Solaris support was removed in #5226.)

The r-solaris check fails

Still fails.

Maybe worth reporting to https://github.com/r-hub/rhub/issues?
I believe our actions cannot affect Matrix package compilatibility.
@jameslamb WDYT?

WDYT?

I agree! I see the following error at the top of the logs from the Solaris jobs, which I suspect is the root cause.

46#> Fatal server error:

[  47](https://builder.r-hub.io/status/lightgbm_3.3.2.99.tar.gz-4f49e421e97542b4876d80c696893832#L47)#> Could not create server lock file: /tmp/.X22-lock

I found one issue with a similar error, where maintainers point to "GitHub rate limiting" as a problem: r-hub/rhub#492 (comment).

I also see the following in the logs:

Error: package or namespace load failed for ‘ps’ in dyn.load(file, DLLpath = DLLpath, ...):
 unable to load shared object '/export/home/XKCQ2Gb/R/00LOCK-ps/00new/ps/libs/ps.so':
  ld.so.1: R: fatal: relocation error: file /export/home/XKCQ2Gb/R/00LOCK-ps/00new/ps/libs/ps.so: symbol psll_memory_uss: referenced symbol not found

Anyway, I can try tonight submitting to R-hub from master. If that works, then maybe that means this was a temporary issue. If not, then I'll report it to {rhub} at that issues link.

Anyway, I can try tonight submitting to R-hub from master

I just built the package like this:

sh build-cran-package.sh

Then manually triggered Solaris builds like this:

package_tarball <- paste0("lightgbm_", readLines("VERSION.txt")[1], ".tar.gz")
rhub::check(
    path = package_tarball
    , email = "FILL_IN_EMAIL_HERE"
    , check_args = "--as-cran"
    , platform = c(
        "solaris-x86-patched"
        , "solaris-x86-patched-ods"
    )
    , env_vars = c(
        "R_COMPILE_AND_INSTALL_PACKAGES" = "always"
    )
)

With email set to the one linked in #5159 (comment).

Statuses should be sent there once the jobs complete.

Links:

• https://builder.r-hub.io/status/lightgbm_3.3.2.99.tar.gz-49175e584edf47c9aa4e04e74ea750d9
• https://builder.r-hub.io/status/lightgbm_3.3.2.99.tar.gz-de5f17b897844c598adddbd49782e707

Hasn't this meant that our token was overwritten?

I just get it! Matrix installation happens inside wch1/r-debug, not at RHub.

Job fails even before submitting the LightGBM package to RHub.

Hasn't this meant that our token was overwritten?

It would surprise me if that was true. I assumed that when I was prompted for a new token, R-hub was issuing me an additional token for use with that email, and that you could have multiple associated with the same email address.

If that's not true and we find out that the token has been overwritten, then I apologize and I'll put up a PR to fix the comment-triggered jobs.

Job fails even before submitting the LightGBM package to RHub.

ooooh I see! I was wondering where @shiyu1994 saw failures related to {Matrix} when I didn't see that in the logs.

I guess then, based on #5159 (comment), we need to install gfortran into that job's environment prior to building the package.

This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.