Closed JosiahParry closed 5 months ago
What does 'intermitenly fails to install during CI' try to say? Are you building everything from source on each run?
Per your log you are using r-lib
and caches and what not. Which ... I don't use so I am not sure I can help you with that.
I tend to minimize building what I don't have to build and fetch what I can as binaries from r2u ("works great"), r-universe, CRAN. The code in the package has not changed in some time and I am not aware of any fragility. That said, we know bugs lurc and I spent most of yesterday chasing a heisenbug once covr
was activated...
For what it is worth: Six point nince seconds.
Maybe you are running out of ram or other resources? Anyway, I don't see anything with with the package.
Also, why are you wasting all these cycles? oldrel-4? Do you know when the last time was I ran R 3.6.* ?
It's a C++17 package, and it was quite demanding on the toolchain when we wrote it. I have no idea what mix you have there for a 2 year old distro release (as we get Ubuntu 24.04 next months) coupled with a five years old R binaries, It seems to work just fine on ubuntu devel and release.
Thanks for your thoughts! I use the r-lib/actions r-cmd-check-full
when working with Rust because I've been bitten by quite the number of issues with different OS in the CRAN submission process when Rust is involved. However, the point about 3.6.3 is a good one—I didn't realize that oldrel-4 was 3.6! I'm not even sure if CRAN is checking oldrel that far back (I'm not sure where I would see that)?
Removing the oldrel-4 check gets rid of the problem :) I suppose theres some compiler toolchain issue there thats beyond my head. Thank you for your time!!!
CRAN only knows three states: present (r-release), future (r-devel) and immediate past (r-oldrel). See e.g.for your arcgis package:
And if you click on the links therein you get these details:
For the next few weeks this is a little more refined as we have r-release at 4.3.3. the release cycle now in motion with r-devel at 4.4.0 and hence r-devel at 4.5.0 now. But the key point is that we still have the triplet <old-release, release, devel>. Nothing more.
Having usethis
add five releases is (IMHO) utter nonsense. Net harmful but I suppose crypto mining and LLM training are more wasteful. But in difference to the other two, you have control over this last one. So I suggest you don't.
Net harmful but I suppose crypto mining and LLM training are more wasteful.
I come here for the rants. :-)
We aim to not disappoint.
Net harmful but I suppose crypto mining and LLM training are more wasteful.
I come here for the rants. :-)
looool!
@eddelbuettel thank you for the details! To confirm my understanding the triplet is always the latest minor version difference, is that right? So we have <4.2.3, 4.3.3, 4.4.0> as the testing targets.
"Yes but" and as I wrote, because we are currently in the clearly outline beta/alpha/rc cycle it may now be 4.2., 4.3. and 4.5.0 because for a few weeks at least 4.4.0 and 4.3.3 kinda coexist.
But qualitatively this correct: old-rel is '-1', release is current and devel is what release with be next. But eg I was curious yesterday too and updated my local R-devel (which I otherwise do once every week, or every other week) and R-devel now calls itself 4.5.0:
edd@rob:~/svn/r-devel$ RD --version
R Under development (unstable) (2024-03-28 r86221) -- "Unsuffered Consequences"
Copyright (C) 2024 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu
R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under the terms of the
GNU General Public License versions 2 or 3.
For more information about these matters see
https://www.gnu.org/licenses/.
edd@rob:~/svn/r-devel$ RD -q -e 'print(getRversion())'
> print(getRversion())
[1] ‘4.5.0’
>
>
edd@rob:~/svn/r-devel$
Hey! Love the package, and I use it thoroughly in the
r-arcgis
R packages I'm developing.One thing I've been encountering is that RcppSimdJson intermittently fails to install during CI. I was curious if you might have some insight as to why it might be happening. For example this CI run https://github.com/R-ArcGIS/arcgisgeocode/actions/runs/8474106631/job/23220318803