Generated PDF (GitHub Action): missing many images

[CameronGray1210]

Generating the PDF on local machine will properly include all images, however, the automated PDF generation on GitHub fails to render many images.

[humphd]

Which one(s)? cc @TueNguyen2911

[CameronGray1210]

Far too many to list (issue persists throughout all documents). It appears random and I don't see a pattern based on image type and syntax used in markdown - very perplexing. Here are a few to references to get things started. The generated PDF on GitHub misses the following images:

computer.md

• "Hardware": /img/programmer.png
• "Modern Computer": /img/hardware.png
• "Software": /img/software.png

information.md

• "Bytes": /img/bytenibbit.jpg
• "Memory Model": /img/ram.png
• "Addresses": /img/image21.png

Examples where images DO render correctly:

computers.md

• "Central Processing Unit": /img/cpu.png

compilers.md

• "Linux": /img/gcc.png
• "Windows": /img/cl.png

The interesting thing is when the PDF process is generated locally (on my machine anyway) it works fine.

[humphd]

This sounds like https://github.com/kohheepeace/mr-pdf/issues/37

[CameronGray1210]

Ahh.. yes that might very well be the cause. I was able to replicate the missing images even on a local build. When this happened, I then loaded all pages in the browser to build a cache, then generated the PDF and it worked. So there seems to be a link to how it uses the cached images (or the slow loading of them) to the rendering in the PDF.

[humphd]

The fix for this is to have the render wait on all images to load fully. I think it's likely to get fixed upstream.