Week 0002

Made a couple fixes to the uefi-rs PR that re-adds the exit code check, it's merged now.

Realized today that I need some vacation, put in for next week off.

Quick doc cleanup PR for the uefi-rs template readme.

Taking a look at an issue about string conversion in uefi-rs. Some of the suggested changes have been made, but there's probably room for more conversion utilities and code cleanup. There's also quite a bit of discussion about improvements to Writeing UCS-2 strings, but I'm not clear on how necessary that is.

I realized CStr16::as_string isn't needed since to_string is already automatically implemented, put up a PR to remove it.

Now looking at removing some of the implicit conversions in the file system code. I've half convinced myself there's a bug in some unsafe code in NamedFileProtocolInfo, will have to check that more carefully. I might start by adding some more tests for the file code, since it looks like a lot of it isn't yet covered by uefi-test-runner.

Took a little time to send messages to my senators about the EARN IT Act, wherein they are once again trying to let law enforcement spy on any ol' thing without even a pretense of proper oversight. More info at the EFF.

Spent a while trying to figure out how to convert from HTML to PDF. I didn't end up finding a good option. There are plenty of tools that can do the basics of conversion: pandoc (with various "engines"), Prince, wkhtmltopdf, chrome (with --headless --print-to-pdf-no-header --print-to-pdf), weasyprint, but all of these produce somewhat ugly output that doesn't match what the HTML looks like in the browser. Firefox's print-to-pdf actually produces very nice output, but as far as I can tell there's no way to turn off the headers and footers, and there's no way to generate a pdf from the command line [bug].

So, for now maybe the best way to generate a PDF is still LaTex, which I really did not expect to be the case. Fortunately it's a lot easier to work with LaTex than it used to be thanks to the amazing tectonic project, but I really wish I could use HTML for this.

I guess the real problem is that the format we really want doesn't exist: we want the locked formatting and single-file features of PDF, but without the incredible complexity of the PDF spec.