At 2023-04-25T10:05:14-0400, Douglas McIlroy wrote: > > $ ./build/test-groff -Tutf8 > .> nr a 3c > > .nr b 3cm > > .tm a=\na, b=\nb > > a=283, b=283 > > > This suggests that one could get away with "3in" as well. Yeesh. > > Not sure how I feel about that. I think I'd prefer to have Yet > > Another Warning Diagnostic for non-pristine input syntax. > > Beware of the infamous Postel principle, which has enabled many hacker > exploits and was embraced by the original definition of HTML: Be > conservative in what you write and liberal in what you accept. With > every browser having its own liberal idea, authors wrote HTML in local > cocoons, making HTMLdocuments hideously fragile. Ironically, Postel's > bad advice has been called the "robustness principle".
Yes. I am particularly motified by the two examples above because they are seductively misleading of the user. groff "understands" the conventional unit abbreviations "in" and "cm", they will think. And be frustrated when "pt", or others they might imagine, don't work. I'm in the choir you're preaching to. But you probably knew that--you've seen me prattle on about how much better Ada was than C. Apart from the whole "lack of industry-reconfiguring success" thing. > In the interest of portability, I would make this a hard error, though > not necessarily one that stops groff in its tracks. We might parlay > this to a wider discussion later. I might slip this in at about the same time as I deal with https://savannah.gnu.org/bugs/?60955 Regards, Branden
signature.asc
Description: PGP signature
