search

irods / irods_client_library_rirods

rirods R Package
https://rirods.irods4r.org
Other
6 stars 5 forks source link

Releasing patch 1.0.2 #45

Closed MartinSchobben closed 10 months ago

MartinSchobben commented 11 months ago

I intend to submit this patch to CRAN. This contains the updated documentation of Mariana. However, due to the missing dependency in the the irods demo, all GitHub workflows fail for now (irods/irods_demo#46). Any idea how to circumvent this for now?

trel commented 11 months ago

PR https://github.com/irods/irods_demo/pull/47

MartinSchobben commented 11 months ago

Submitted patch to CRAN. Let's wait if that goes through and then merge this PR with a tag for 0.1.2.

MartinSchobben commented 11 months ago

It's on CRAN now: https://cran.r-project.org/web/packages/rirods/index.html

MartinSchobben commented 11 months ago

Can you merge this and release with tag 0.1.2?

trel commented 11 months ago

lots of conflicting files at the moment - not yet mergeable.

MartinSchobben commented 10 months ago

yes, the snapshots are updated weekly. I will see that I rebase this week. I believe the snapshots are created on Monday.

trel commented 10 months ago

Oh, I see that now. And my name is on all of them. Why is anything getting pushed weekly if there are no code changes? That feels surprising and weird/wrong.

MartinSchobben commented 10 months ago

I added this to enable testing without an iRODS server/internet connection (e.g. on the CRAN servers). I have the weekly updates to capture potential changes in the HTTP response. But maybe weekly is a bit much.

I presume I can also trigger the event by a change in the submodule irods_demo.

trel commented 10 months ago

Oh, I see. Is there a way to remove/ignore the timestamps from responses files? Then you could detect any 'real' changes more easily?

MartinSchobben commented 10 months ago

ah yes, that should be possible as well. It entails changing inst/httptest2/redact.R.

MartinSchobben commented 10 months ago

I fixed the HTTP snapshots that now not include dates anymore. The tag for version 0.1.2 now belongs to commit 4144b54. This commit is the same as the "truth", from the point of view CRAN (R).

trel commented 10 months ago

so this is okay to be merged, as is? five commits?

MartinSchobben commented 10 months ago

yes, that is fine. Up to commit https://github.com/irods/irods_client_library_rirods/commit/4144b545566069a01c8a54870f1051532775024e is needed to stay in sync with CRAN (0.1.2) the last two fix the snapshots in the dev version.

~Regarding fixing the automated snapshot generation. I now have to update the snapshots again :-)~

trel commented 10 months ago

Got it.

One last note - when rebasing, the SHA of that commit will change (again).

If that matters, then we need to make sure we get things merged (with a stable SHA) before we declare a release. If not, then we should be good to go.

Merging!

MartinSchobben commented 10 months ago

That makes sense. Never thought about that. In the normal situation the R developer will be push a package to CRAN from the main branch. That is not possible in our situation, where I push to CRAN from my own development branch. This is, however, all only considered as good practice and not mandatory.

trel commented 10 months ago

You can still push to CRAN from main - just have to wait until we merge first?

MartinSchobben commented 10 months ago

Right! That would be better then.