XQuartz incompatible with mac OS Tahoe?

[Chris Law]

Hi geomorph community,

I have a few undergraduates trying to install geomorph onto their Macs (OS Tahoe). However, it turns out there might be a compatibility issue between XQuartz and Tahoe OS. Has anyone figured out how to get XQuartz to work on their Tahoe OS or a way to get 3D figures to display when using rgl/geomorph? Below is the error displayed when loading geomorph.

1: In rgl.init(initValue, onlyNULL) : X11 error: GLXBadContext 

2: 'rgl.init' failed, will use the null device

Thanks!

Chris J. Law

,

 

PhD | he/him

Principal Research Scientist | Department of Biology

Affiliate Curator | Burke Museum

University of Washington, Seattle, WA

cj...@uw.edu | chrisjlaw.com

[Adams, Dean [EEOB]]

Chris et al.,

 

Yes, there is a compatibility issue with MacOS Tahoe and XQuartz. This is a high-level OS issue which unfortunately has a cascading effect to all R-packages that utilize RGL, including geomorph (see RGL’s web site where they state: "Currently Tahoe, the most recent version of macOS, does not support OpenGL in XQuartz, and it appears unlikely such support will be coming.")

 

The geomorph team is aware of this new challenge, and we are attempting to come up with strategies that mitigate the issue.  None of our current options are optimal, and we are trying to navigate a path for geomorph that retains the maximal usage for the largest number of users across multiple OS platforms.

 

The good news is that when you load geomorph on Tahoe, what you received was just a warning, not an error. Thus, geomorph did load, and as such all of its analytics and data analysis functions should run as intended. However, as you noted, the 3D graphics components will not, as they depend on RGL which is affected by the XQuartz issue.

 

We will obviously keep the geomorph community updated with any solutions we come up with, or any changes that are necessitated by this change due to the OS compatibility issue.

 

We thank you for your patience while we deal with this. Sometimes the R-world throws us package developers rather unforeseen curveballs!

Best,

 

Dean

 

Dr. Dean C. Adams

Distinguished Professor

Director, Ecology and Evolutionary Biology Graduate Program

Department of Ecology, Evolution, and Organismal Biology

Iowa State University

www.public.iastate.edu/~dcadams/

phone: 515-294-3834

--
You received this message because you are subscribed to the Google Groups "geomorph R package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/geomorph-r-package/CAA202XYBpiDgHMd3o-KF_Mwg9eH8Kokqoj_yi4gTv-sN1aFHbA%40mail.gmail.com.

[mura...@gmail.com]

Apple stopped supporting OpenGL (which is RGL provides) over a decade ago in lieu of their own 3D rendering library Vulcan and finally pulled the plug on it on Tahoe. The only real solution if Duncan Murdoch (maintainer of RGL) considers using Vulcan API on Apple os. Looks like the issue is open on RGL site, and the only solid fix seems to compile RGL from the source and then use web widgets instead of built-in rendering. 

[mura...@gmail.com]

Forgot to paste the link to track https://github.com/dmurdoch/rgl/issues/488

[Chris Law]

Thanks for the confirmation, Dean. Yes, the undergrads are having fun learning to use the analytical side of geomorph. I guess they'll have to get used to seeing TPS plots for now!

To view this discussion, visit https://groups.google.com/d/msgid/geomorph-r-package/SJ0PR04MB8306C29CC825E5B4965EA766A26AA%40SJ0PR04MB8306.namprd04.prod.outlook.com.

[Mike Collyer]

Dear geomorph Community,

Out of curiosity, are there many of you who use plotly for visualization?  This seems like a potential avenue for change, because the plotly R package is a way to make the plotly.js software work with R, and because web browsers can handle JavaScript.  This could be a safer alternative for 3D visualization tools, but I do not have a sense, e.g., if one can make functions that select points from interactive graphs.  Another appeal is that embedded interactive graphs can be generated for html documents in RMarkdown or Quarto.  So having pop-up rgl windows would not be a necessity.

However, I have read that plotly has discontinued developing documentation for R, and it appears the focus is on continued development for python, where it is apparently more flexible and advanced.  Perhaps a motivation is that within Quarto, one can use either R or python scripts, so spreading themselves thin with multi-platform options limits the developers opportunities to enhance their software in fewer platforms.  I’m not sure, just a guess.  The paranoid side of me wonders if the vicissitudes imposed with CRAN updates just makes keeping up with documentation too crazy?  (We have killed some documentation in the past because R base updates rendered difficult-to-overcome errors for making help documents.  Trying to hit moving targets can be really frustrating for developers who do not get paid for this.)  So, I question whether switching to visualization tools that do not update documentation is safe for users.

I have used plotly sparingly, and liked it quite a bit, but have not taken a deep dive into it to know of limitations.  If any of you have either warnings or plaudits to offer, I would like to hear them.  Some of you work more in 3D visualization, so you might have sage perspectives.  I do not know if leaving rgl behind and using more modern 3D technology is an option to explore, and it is not something that could happen overnight, but the motivation for such an endeavor could be amplified if we knew geomorph users would appreciate it.

Many thanks!

Mike

To view this discussion, visit https://groups.google.com/d/msgid/geomorph-r-package/CAA202XZNZ8rvaSnzN-cTRrukVbqrp4xW%3D%2Bw01-XAmMvWMS0hug%40mail.gmail.com.