What happened to posting a comment with a link to a pdf and html?

[olgabot]

After updating to the most recent rootstock (https://github.com/czbiohub/kmer-homology-paper/pull/75), new PRs no longer have a comment with a link to the manuscript PDF :( I found this EXTREMELY convenient to not have to navigate through the weird GitHub actions API and remember that the PDF is an "artifact" (how is a newly published manuscript an "artifact" anyway? it's not from Ancient Egypt or anything...)

It is still available after 4 clicks (click the Actions checkmark, click Actions, click Artifacts, then click the PDF): https://github.com/czbiohub/kmer-homology-paper/runs/1454813308 -- but before it was available as a single click. Would it be possible to bring back the comment with a link to pdf and html?

[agitter]

Those comments are posted by the AppVeyor CI service. They are still available with the latest rootstock updates. I don't see any errors in your .appveyor.yml file, so I'm guessing there could be a problem with the AppVeyor settings. The last AppVeyor build was for pull request 72, before the rootstock update: https://ci.appveyor.com/project/olgabot/de-novo-orthology-paper/history

Was the repository renamed recently? I noticed the current name is kmer-homology-paper but the AppVeyor history is for de-novo-orthology-paper.

[dhimmel]

Was the repository renamed recently? I noticed the current name is kmer-homology-paper but the AppVeyor history is for de-novo-orthology-paper.

I am guessing this might be the cause. I wonder if its possible to setup AppVeyor again for the new repository name, such that you'd have AppVeyor active at https://ci.appveyor.com/project/olgabot/kmer-homology-paper.

The Manubot setup docs we have for AppVeyor are here (must expand the "Previewing pull request builds with AppVeyor"). Those docs probably aren't super relevant, since they assume a first-time setup. Fingers crossed re-enabling AppVeyor will fix the missing AppVeyor builds.

[dhimmel]

I found this EXTREMELY convenient to not have to navigate through the weird GitHub actions API and remember that the PDF is an "artifact"

Agree that access to the built manuscript on GitHub via PR artifacts is far from ideal. Looking over #322, it seems like there's a couple limitations on GitHub's end that prevent us from commenting with the built manuscript.

One idea that was floated somewhere (I forget) is that we could deploy PR manuscripts to the GitHub Pages site under a versioned link, to make each PR have a stable URL for viewing it. But this would take some work to implement.

[agitter]

we could deploy PR manuscripts to the GitHub Pages site under a versioned link, to make each PR have a stable URL for viewing it

That was a good idea. #351 has the discussion.

[olgabot]

I am guessing this might be the cause. I wonder if its possible to setup AppVeyor again for the new repository name, such that you'd have AppVeyor active at https://ci.appveyor.com/project/olgabot/kmer-homology-paper.

Yes this was it!! My bad. This paper has gone through many name changes so I have done this change before, but forgot to do it again this time. Thanks!@

[agitter]

Great, I'm glad it's working again. I'll close this issue since the comments are showing up in your pull requests.