[tex-k] Where does conversion of PostScript pts to TeX pts occur?

[Karl Berry]

    But TeX thinks that 1bp is actually the fixed-point integer 65781, as
    demonstrated by executing this TeX code:
    \newdimen\onebigpoint \onebigpoint=1bp 
    \newcount\testbpvalue \testbpvalue=\onebigpoint 
    \the\testbpvalue 

pdftex also outputs 65781 for that document, thankfully.

    one_bp := 65782; {65781.76} 
    ...
    This strikes me as a potential problem when comparing bigpoint
    dimension values.

Because pdf_px_dimen:=one_bp, so we might end up comparing 65781 from
TeX-world calculations to 65782 from pdftex-world px calculations?
It sounds plausible, but I don't know. (The couple of other usages of
one_bp in pdftex do not look problematic.)

In any case, that setting of one_bp has been there forever. Even
assuming we could construct a case where 1bp (TeX) != 1bp (px),
I don't think it would be a good idea to change it. Whatever problems
there are must have already been worked around.

    no comments in the code on the motivation for rounding it up even
    though TeX rounds down.

My wild guess would be that Thanh or Martin or whoever defined it did
not realize that TeX rounded down, just did the calculation from scratch
and did the natural rounding. --best, karl.