fontinst bug(s) ?

[Lars Hellström]

>I do indeed see in the log-file that t1c.etx is called, but in the output,
>(test2.ps_full, test2.pdf_full_Acro, or test2.pdf_pdftex)
>the bold small caps are basically the size of true caps: about 40% larger
>than the real (regular) small caps...
>
>Conceivably this is just a "feature"...

It's not a bug anyway. The amount by which the upper case is scaled to fake
small caps is stored in the fontinst integer smallcapsscale, and latin.mtx
uses the default value 800 for this. You could set it to some other value
before the \latinfamily, e.g.

  \setint{smallcapsscale}{600}
  \latinfamily{pad}{}

would get you smaller faked smallcaps, but the more you scale, the more the
appearent weight changes. You can't get everything right, but I agree not
having the same smallcaps x-height in all weights must look very strange.

BTW, nice to hear that someone (other than the maintainers) is reading
fisource.

Lars Hellström