WIRED has the vapors over third-party auth services like Google, Apple, Patreon, X, Line, etc, being used to… authenticate to web sites; specifically, the sort that paste heads onto naked bodies. I’m sure they tried to blame Elon Musk, but reluctantly had to admit that X’s auth wasn’t the only one being used.
So, nothing to see here. Literally; it’s like the folks at WIRED and Ars have never heard the phrase “pics or it didn’t happen”. 😁
I’m sure there’s a ToS issue in there somewhere, but they seem to be particularly upset at the idea of teenage boys using the sites to generate imaginary nudes of their female classmates, demonstrating that they’ve never met a teenage boy.
…asks Slashdot, quoting The Verge’s transportation editor.
Answer: electric cars.
See also “winter”, “spontaneous combustion”.
Building a test suite for a PDF generator requires some way of
validating what ends up in the binary output file. For PDF::Cairo, I
supplied a reference PDF file and used
Poppler’s pdftocairo
to render to
PNG for comparison. Unfortunately, what this really ended up testing
was the underlying libraries rather than my code, which is why the
CPAN automated tests keep breaking.
For this script, my testing can be limited to determining that known
text ends up in the correct region of the page, at the correct size,
and it doesn’t have to be precise. It turns out that Poppler’s
pdftotext
extraction utility has a -tsv
option that reports the
bounding box of each word on every page, which will suffice.
Once that’s in place, I think I’ve got everything compatible as far back as Python 3.9.x and Reportlab 3.6.x, and for the regular test suite I can just dump the internal state object to compare to a reference version stored in Configparser format.
But I still think I’ll work on page-styling first.
Markdown formatting and simple HTML accepted.
Sometimes you have to double-click to enter text in the form (interaction between Isso and Bootstrap?). Tab is more reliable.