releasing 0.6.1

Question

releasing 0.6.1

IndrajeetPatil opened this issue 3 years ago · 21 comments

IndrajeetPatil commented 3 years ago

@const-ae Would you be open to doing a minor release 0.6.1 sometime soon?

It will be good to have the new testing infrastructure and website URL on CRAN.

Answer 1 · 2021-02-22T08:45:10.000Z

Yes I think this is a good idea.

Are there more features to add and bugs to fix or can I review the code in it's current state and send it off to CRAN?

Answer 2 · 2021-02-22T09:33:56.000Z

No, currently there are no additional features to add or bugs to fix. There hasn't been much addition to the code since the last release. It's just a lot of tests and other infrastructural components were added.

Also, I ran revdepcheck and everything looks good!
#84

I have also updated cran_comments.md for submission.

Answer 3 · 2021-02-22T13:25:23.000Z

Btw, can you please change the source for github pages to following:

We are now using github actions to build the website.

Answer 4 · 2021-02-22T14:12:53.000Z

Okay, done :)

Answer 5 · 2021-02-22T14:22:48.000Z

Okay, cool. Let me know when you have submitted to CRAN!

Answer 6 · 2021-02-22T15:12:10.000Z

Okay, so I just went through the devtools::release() checks. And I have two questions

Now that we are building the website with Github actions on the gh-pages branch, can we delete the docs folder on the master or is that still needed?
There is a weird WARNING while running the checks:

    ─  checking for empty or unneeded directories
   Removed empty directory ‘ggsignif/WIP’
   Removed empty directory ‘ggsignif/_snaps’
     NB: this package now depends on R (>= 3.5.0)
     WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/dodge_comparison_e590628d7e6d7d433e967afd28f467cd.RData’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/dodge_comparison_e590628d7e6d7d433e967afd28f467cd.rdx’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/faceted_comparison_69c8749b9264a70296877658d75692b1.RData’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/faceted_comparison_69c8749b9264a70296877658d75692b1.rdx’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/simpe_comparison_0904ec0533f7aab714e0411dded9c4f7.RData’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/simpe_comparison_0904ec0533f7aab714e0411dded9c4f7.rdx’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/unnamed-chunk-2_a171c02e5427c1cabb4140b102de7964.RData’  WARNING: Added dependency on R >= 3.5.0 because serialized objects in  serialize/load version 3 cannot be read in older versions of R.  File(s) containing such objects:  ‘ggsignif/tools/README-cache/unnamed-chunk-2_a171c02e5427c1cabb4140b102de7964.rdx’
─  building ‘ggsignif_0.6.1.tar.gz’

It seems that it is not actually registered as a warning by R CMD CHECK, because the check is still successful. Do you know where that warning comes from and if it's save to ignore?

Answer 7 · 2021-02-22T15:22:32.000Z

That's strange.

There is no ggsignif/tools directory.

Maybe try with devtools::check_win_devel() and see if you see the same thing?

Answer 8 · 2021-02-22T16:04:53.000Z

Okay, the test from Winbuilder came back as successful (https://win-builder.r-project.org/bTj4Z995iMa7/). However the issue actually persist, which you can see when you download the zip-file of the package. The following line

Depends: R (>= 3.5.0)

is added to the DESCRIPTION file. This is really weird :/

Answer 9 · 2021-02-22T16:09:21.000Z

I have just restarted Rstudio and now the warning is gone. At least locally. I have submitted a second round to Winbuilder. Let's hope, restarting actually "fixed" the problem

Answer 10 · 2021-02-22T16:12:28.000Z

Strange indeed. I have also run devtools::check_win_oldrelease(), so you should get an e-mail about that.

Let's see if this has something to do with our caches in local systems.

Answer 11 · 2021-02-22T16:37:46.000Z

Okay, good idea. I'll keep you posted :)

A separate thing: I also ran a test on Rhub, which produces a NOTE about \Sexpr in the documentation (https://builder.r-hub.io/status/ggsignif_0.6.1.tar.gz-43032db7327f484c8a37b1d108c6a27b):

 967#> Package has help file(s) containing install/render-stage \Sexpr{} expressions but no prebuilt PDF manual.

I haven't used the \Sexpr macro before, but I found a thread from lifecycle package, who now recommend to use

#' `r lifecycle::badge("deprecated")`

instead of the \Sexpr macro. Do you think this is worth fixing?

Answer 12 · 2021-02-22T16:50:11.000Z

Hmm, in that case, I would much prefer to get rid of that from the docs and not gain an additional dependency.

We can instead show in the README that the package is in stable state.

Will make a PR now.

Answer 13 · 2021-02-22T16:58:01.000Z

I am also slightly modifying the README title to match DESCRIPTION title.

Answer 14 · 2021-02-23T10:32:42.000Z

Let me know if something else comes up in r-hub or winbuilder checks.

If not, we are good to go! 😄

Answer 15 · 2021-02-23T10:34:09.000Z

Yeah, I just ran another check with Winuilder and the tools/README-cache/ thing popped up again. Now as a warning: https://win-builder.r-project.org/CP4M36O13Vgs/

Answer 16 · 2021-02-23T10:36:08.000Z

I wonder if it is related to the cache that was removed in PR 79 (https://github.com/const-ae/ggsignif/pull/80/files#diff-72778b58969c8ca8268402860b0e003e3d213a26c812bc9f9b928395c284c99fL20). But it just doesn't make sense, because the cache was removed. So I am right now trying to figure out, how R remembers that there is some cache...

Answer 17 · 2021-02-23T10:42:25.000Z

Yeah, this is totally weird and unexpected behavior.
Can you check if there are hidden files on your system that contain this cache, which is being packaged into the file submitted to winbuilder?

I just did winbuilder checks, and you should be getting emails about them. Let's see what we get there. If we also observe the same problem with my system, then we need to dig deeper into this.

Answer 18 · 2021-02-23T10:59:29.000Z

The result from winbuilder came back and look good now (https://win-builder.r-project.org/RmSGxeQBrqyl/00check.log). I think, I'll just risk an submit. If there actually is a problem, the package will be auto-rejected and then we can dig deeper. But it probably doesn't make sense right now to chase this if it could just turn out to be a weird fluke on my system.

Answer 19 · 2021-02-23T11:01:55.000Z

I'll keep you posted how the submission goes.

Answer 20 · 2021-02-23T13:42:51.000Z

The update is accepted and on its way to CRAN. Thank you so much for making this release happen 🎉 🥳 🎉

Answer 21 · 2021-02-23T13:44:15.000Z

Yaay!!! 🎉

Thanks for letting me be a part of this process 😅