Hi Dave, > > Bug. CSTR 54 doesn't allow it, nor does > > info groff 'Changing Type Sizes' | less > > It doesn't seem quite that clear-cut to me. Neither of those sources > explicitly states what *roff does with additional parameters to .ps, > or with nonnumeric trailing characters to a numeric parameter.
A modern interpretation of the documentation would be that no further arguments are allowed, but I agree that historic implementations wouldn't have bothered enforcing that, probably for good reasons at the time. > And I bet that .ps is far from the only request that ignores both of > these. ‘.ce 3 1 4 1 5’ ‘works’. > Still, regardless of what the documentation says, there's a > commonsense argument to be made that it's a better design if the > parser nudges the user when it encounters something unexpected. Yes. > Care to open a bug report for this? Okay. https://savannah.gnu.org/bugs/?61450 > I feel like someone besides Bjarni and me should own a few bugs. There's my https://savannah.gnu.org/bugs/index.php?24047 from 2006 complaining that an -mm document has text in different places between versions 1.17.2 and 1.18. That's what first alerted me to the lack of regression testing in groff. Without it, and knowing that type is set consistently over time for the same input, we may as well use EPUB. ;-) -- Cheers, Ralph.
