setspace with non standard point sizes.

[davidcarlisle]

https://tex.stackexchange.com/questions/734794/spurious-text-when-using-setspace-and-scrextend-with-decimal-font-size

here \@ptsize is 0.5 we could avoid the issue by rounding up(?) to the nearest integer before doing the \ifcase but if changing this area we might also want to support other values eg extsizes which leaves this command with its "natural" value rather than ten less.

\documentclass{article}\usepackage[fontsize=10.5pt]{scrextend}% 0.5 10.5, 12.59996

%\documentclass{article}\usepackage[fontsize=14.5pt]{scrextend}% 4.5, 14.5, 17.39995

%\documentclass[14pt]{extarticle} % 14, 14.4, 17
%\documentclass[12pt]{extarticle} % 12, 12, 14.5

\makeatletter
\show\@ptsize
\show\f@size
\showthe\baselineskip
\usepackage{setspace}

\begin{document}
\onehalfspacing
text
\end{document}

[komascript]

Such definitions of \@ptsize were the origin of setspaceenhanced.sty. However, the package now offers far more options and is completely independent of \@ptsize, working instead with \f@size and \f@baselineskip. This means it also works with classes that do not define \@ptsize in the first place.

But I know you don't like the extension because, under certain circumstances, it can lead to different results in existing documents. Of course, a rollback could also be offered for such cases.

[davidcarlisle]

@komascript I wonder... It's true of course that changing anything may potentially affect some documents, so doing nothing is a possibility, if we do anything we end up with old setspace, new setspace and setspaceenhanced as three incompatible variants which isn't clearly a win.

I don't know setspaceenhanced in detail. if we made setspace be that, are there easily documented cases where existing documents would change? Or a reasonable setup where setspace.sty was setspaceenhanced followed by some compatibility adjustments?

[komascript]

Since setspaceenhanced has been completely freed from the use of \@ptsize and thus also from a fixed assignment \@ptsize → factor, and instead performs a calculation of the factor based on \f@size, there is no compatibility mode whatsoever. Of course, one could be added, but for those who need absolute compatibility, it is IMHO easier to use the LaTeX rollback mechanism for packages.

If setspace were to offer everything that setspaceenhanced offers in the future, I would have no problem withdrawing setspaceenhanced or converting it into a wrapper package. If the email discussion about setspace (and lscape) had not been so unsatisfactory and inconclusive at the time, the package would not even exist.

[komascript]

BTW: The maximum compatibility with setspace currently available is achieved by using \usepackage[byselectfont=false,keepfontsize=false]{setspaceenhanced}.