ropensci-org / rotemplate Goto Github PK

Not sure why yet @jeroen

https://docs.ropensci.org/grainchanger/

Hi there,

I guess this issue is related to how pkgdown websites are deployed on https://docs.ropensci.org.

It looks like the filename of the logo matters, using rcites_logo.png I still have my logo but also the RopenSci one, see the difference between https://docs.ropensci.org/rcites/ and https://ropensci.github.io/rcites/.

I guess this is just a matter of adding a sentence (or a link) in the readme mentioning what the logo filename should be!

Hope this helps and thanks for your work on this!

See: https://docs.ropensci.org/DataSpaceR/

See the docs column in https://ropensci.r-universe.dev/

Most problems are probably pre-existent, and not due to the new pkgdown. Still this may be a good moment to try and fix them, otherwise they won't get new docs. Maybe retire some things that are broken and unmaintained.

Some of the links in the readme (including for the example travis.yml) are still pointing at the ropenscilabs repo.

https://github.com/ropensci/citesdb/blob/ca579791a574b04e04c755c569a916a838db2ef8/pkgdown/_pkgdown.yml#L6

Hi,

I recently moved my package to rOpenSci. All things worked as described, but the Google Analytics tracking are not showing up in the code. Is there something I miss?

The package documentation website can be found at https://docs.ropensci.org/mctq/ .

Cheers!

See for example: https://docs.ropensci.org/tesseract/articles/intro.html

According to gabor, this was fixed. The tibble pkgdown site itself does not seem to suffer from this: https://tibble.tidyverse.org/reference/tibble.html

Back before weathercan was part of rOpenSci, I had a red theme for the package docs. This was overridden by the rOpenSci theme, so I ignored it but never removed the red theme. Now in the new theme, parts of that original red show through (https://docs.ropensci.org/weathercan/).

If this is just something I should fix on my end, I'll close the issue and deal with it 😁, but I thought I'd mention it in case it's undesirable behaviour from the template.

In the README.

rotemplate is having some issue related to image rendering. The bug is recent.

Example (today - 2021-12-03)

Source: https://docs.ropensci.org/mctq/

Print from 2021-11-10

Source: https://web.archive.org/web/*/https://docs.ropensci.org/mctq/

I was only able to create the Travis project under travis-ci.com

At the moment travis::use_travis_deploy() doesn't support .com (but should soon? cc @krlmlr)

See this pages, in the bibtex content, there is no year arguement, so the year argument in the printed citation is ???? , any way to fix it?

If there is a vertical scrollbar, we also get horizontal scroll :(

I'm sure it wasn't there before.

See the block that starts with QLD_hotdates in https://docs.ropensci.org/bomrang/articles/use_case.html

Everything seems underlined as one big hyperlink.

Hi:

When having a look to my CHANGELOG on a smartphone I can’t see the CRAN release date:

However copy-pasting I see that the element is there.

cffr 0.1.1

CRAN release: 2021-11-11

A css issue?

I.e. some sort of wrapper around usethis::use_pkgdown() that'd have the write template in it, and same for usethis::use_pkgdown_travis().

After switching the template to proxima-nova, the paragraph headers seem pretty thin. Our main site uses a bold headers.

.

If relative URLS are used to link to images in README.md or index.md then they are broken when rendered to docs.ropensci.org

example: https://docs.ropensci.org/tidyhydat/

In our staff meeting on July 31 we discussed again the question if we should build and host docs for unsupported and silly ropenscilabs packages.

We decided we want to keep providing docs for everything in the registry, but display labs package with a an adapted theme that emphasizes this is not an official on-boarded ropensci project, but rather an experimental package.

cf ropensci/codemetar#192 (comment)

cc @cboettig

It seems that pkgdown2 now hardcodes a mathjax 2 include, instead of making that part of the theme 😢

The result is that for packages that use mathjax, the site now has both mathjax 2 and 3 in the header https://docs.ropensci.org/jagstargets/articles/introduction.html

Pkgdown should really not force us to use an (outdated) mathjax library.

I've experimented with pkgdown before and one thing I noticed is that CoC and Contributing links are not automatically added. Would you think about that in an ROpenSci template? Also (cough).

cf ropensci/dev_guide#138 (comment)

Hi there!

When my visdat package is being built I get this error on GH actions.

== Building pkgdown site =======================================================
Reading from: '/Users/runner/work/visdat/visdat'
Writing to:   '/var/folders/24/8k48jl6d249_n_qfxwsl6xvm0000gn/T/Rtmps0XlXh/file816547b99c9'
-- Initialising site -----------------------------------------------------------
Copying '../../_temp/Library/pkgdown/assets/bootstrap-toc.css' to 'bootstrap-toc.css'
Copying '../../_temp/Library/pkgdown/assets/bootstrap-toc.js' to 'bootstrap-toc.js'
Copying '../../_temp/Library/pkgdown/assets/docsearch.css' to 'docsearch.css'
Copying '../../_temp/Library/pkgdown/assets/docsearch.js' to 'docsearch.js'
Copying '../../_temp/Library/pkgdown/assets/link.svg' to 'link.svg'
Copying '../../_temp/Library/pkgdown/assets/pkgdown.css' to 'pkgdown.css'
Copying '../../_temp/Library/pkgdown/assets/pkgdown.js' to 'pkgdown.js'
Error: Error: rotemplate is not installed
11 Removing worktree 11111111111111111111111111111111111111111111111111111111111
Running git worktree remove \
  /var/folders/24/8k48jl6d249_n_qfxwsl6xvm0000gn/T/Rtmps0XlXh/file816547b99c9
Execution halted
Error: Process completed with exit code 1.

https://github.com/ropensci/visdat/runs/3187831393

I'm not sure that I've changed anything in my pkgdown YAML to trigger this - but looking at the pkgdown for magick and their github workflows - perhaps I don't actually need this pkgdown YAML or GH action?

@jeroen if I do that what will I break in the docs pipeline? 😉

See comment in: #28 (comment)

Look if our template contains {{{ }}} fields that we actually do not want to be parameterized.

When I navigate to https://docs.ropensci.org/rtweet/ on my smartphone, the color of the navbar dropdown links make difficult to see the text:

The other day I realized that on rtweet the buttons to reference, vignettes and articles are missing from the header.

I think this is not a problem of building the website, but it might be related to other peculiarities of the package(?).
Thanks for providing these templates!

Cf rtweet

cc @llrs

We can add a 404.md somewhere to get a better 404.html page, similar to: https://docs.ropensci.org/doesnotexist

Currently we get something standard like: https://docs.ropensci.org/antiword/doesnotexist

Hi,

I recently moved my package to rOpenSci. All things worked as described, but the search field of the package documentation website simply doesn't work. Is there something I miss?

The package documentation website can be found at https://docs.ropensci.org/mctq/ .

Cheers!

I think the TOC for articles is still a bit large for example magick vignette. The tidyverse one seems a lot smaller but I'm not sure we can copy that: https://tidyr.tidyverse.org/dev/articles/pivot.html

When viewed on small screens, the navbar covers the content of the page, like here (Firefox on a Moto G5 plus with Full HD screen):

This is equivalent to tidyverse/tidytemplate#42

{pkgdow} now uses bootstrap-toc plugin for the ToC sidebar in articles.
This breaks all ToC of packages using {rotemplate}. Example: tic

See the PR diff for changes.

Also this commit in mlr3pkgdowntemplate might be helpful to get it fixed quickly.

Test with https://cards-dev.twitter.com/validator

See for example: https://docs.ropensci.org/tesseract/articles/intro.html

This is pretty unfortunate wrapping. I think we should either decrease the font-size a bit, or otherwise lower options('width') when rending the site such that it fits the block.

rotemplate is not rendering math notations. Possible issue with mathjax.

Example:

• Source: https://docs.ropensci.org/mctq/reference/msl.html

It would be great if resources shared between all rOpenSci pkgdown websites (e.g. hexlogo.png, ropensci.css, link.svg, pkgdown.css, pkgdown.js) could be served from the same URL. For example:

https://docs.ropensci.org/ropensci.css instead of https://docs.ropensci.org/PACKAGE/ropensci.css

(in the code <link href="../ropensci.css" rel="stylesheet"> instead of <link href="ropensci.css" rel="stylesheet">).

This would enable caching across all rOpenSci pkgdown websites, thereby increasing speed, and decreasing bandwidth usage.

I saw issue #14 but I think there should be more consideration of accessibility. I wanted to submit a new issue because the low navbar contrast actually makes it very difficult to read without hovering over each item (which turns them white- they should probably be white already).

I am not an expert on accessibility either by any means, but there are great tools that help a significantly with getting web content closer to WCAG standards

I understand if we do not have the resources or expertise to make our pages meet every single criterion. None of the web content I produce does either. But I feel like the low contrast navbar is pretty egregious currently and as a low-hanging fruit we could easily do better there.

Making the font white and even bold would make a huge difference. Technically, even that will still fail the contrast test, due to the lightness of the blue background. But still a major improvement. Hovering over items can always be indicated by changing the background color rather than the foreground text, similar to how selected items change. Again, I am not an expert either, so there is nothing special about these particular suggestions.

Anyway, it does sound like people have already met to discuss accessibility, but while there are seemingly countless accessibility guideline items, some could be prioritized to address and are very easy to do so. This one seems very pertinent. While I do have outlier-level poor vision, I at least benefit from corrective lenses (though they do make everything I see appear notably smaller in size than what most people see). There will be plenty of people who find the navbar more difficult to read than I do. I just would like to encourage revisiting this and adoption of some of these tools if they have not been used previously.

• good idea?

• in practice, not sure this can happen in the theme.

some use_blabla function would be great.

when the screen is not wide. the button should be the same color as other text in the navbar.

All all hands we discussed it would be nice to see which pkgdown sites and which pages are visited most.

In particular use the same footer.

Our wrapper needs to overwrite the template:params however I keep breaking one case when fixing another.

It seems that if we pass a list() in our overrides, the params merge (instead of replace) the ones from the _pkgdown.yml. In this scenario we get an error like here.

Error : Can't find Bootswatch theme 'readable' (template.bootswatch in '_pkgdown.yml') for Bootstrap version '5' (template.bootstrap in '_pkgdown.yml')

So instead I tried to set params=NULL in our overrides and manually set mathjax = TRUE later in the wrapper, but then I found out this had broken our mathjax docs again.

So now I reverted it again, but we really need unit tests that test at least these two scenarios.

I think the update to pkgdown 2.0.2 broke our navbar. There is too much top-margin, and not enough bottom margin now. I think it has to do with our hexlogo in the nav bar.
Probably related to r-lib/pkgdown#1989

Before (pkgdown 2.0.1)

After (pkgdown 2.0.2)

Step to reproduce:

• Visit e.g. https://docs.ropensci.org/lightr/
• Scroll down until top navbar disappears (because of headroom.js)
• Notice how the navbar shadow is still here

This sounds like a bug to me. The shadow should vanish once the element it comes from isn't here anymore.