> From: [EMAIL PROTECTED]
> Date: Mon, 2 Nov 1998 09:54:27 -0500
Hi Joel, I think you found the solution ...
> \if{#2}\empty\relax\else\leftline{\Bigtype #2}\fi}}%
> ^ ^....BRACKETS here around #2 are all that changes.
>
> This ultimately explains most of what you were seeing. When the \if tested
> #2 without brackets, it ultimately tested just X, and (for some reason I can't
> explain) concluded it was empty, [...]
... and I found the explanation and wrote it already to Don in a
private mail.
\if#2\empty\relax
If you now insert something which starts with twice the same character
e.g. XXXDesus TeX compares the first X with the 2nd X(!), eats them
because they are equal and returns the remainder "XDesus" and "\empty\relax"
as result. If the first two tokens of #2 are not equal then the "else"-part
is executed and the whole string displayed :-)
Up to now this hasn't been detected because the else-part for an
empty string is nearly the same as the \relax in the then-part :-)
-- Werner
and then concluded the macro (\relax)... It
> appears that this was tested a second time (accounting for the two X removals),
> and then (here's the fun part), it was left with inline text "XDessus" but was
> NOT executing your macro \titles anymore, so it just set the text as \tenrm
> XDessus. (TeX was just doing what TeX would do when it runs into inline text.)
>
> By putting brackets around the argument in that setting, it appears that TeX is
> forced to consider all of the argument to determine the logic outcome, and (when
> I ran it) it appeared to be more what you might expect.
>
> By the way, all this can be seen by putting \tracingall% right before the first
> instance of your macro. A patient person can then search the resultant .log
> file for a step-by-step record of what TeX is doing. This is where it can be
> seen that Tex was processing your \if test based only on the first X.
>
> I hope the result here is just what you need. It is possible that the \if test
> activity has more complications than this suggestion might handle, (for
> example)... I worry that maybe the brackets will make it always fail the test,
> then you wouldn't have what you wanted either... but as was once written
> "Sufficient for the day is the trouble thereof."
>
> Regards
> Joel Hunsberger
> [EMAIL PROTECTED]
>
> ______________________________ Reply Separator _________________________________
> Subject: TeX Problem
> Author: [EMAIL PROTECTED] (Don Simons) at SMTPpost
> Date: 11/1/98 9:16 PM
>
>
> In the following tex file, can anyone figure out why "Dessus" and
> "XXXDessus" come out in different fonts? And what happened to the
> first two X's in "XXXDessus"?
>
> --Don Simons
>
> ============================
> \input musixtex
> \font\BIGfont=\fontid bx10 scaled\magstep4
> \font\Bigfont=\fontid bx10 scaled\magstep2
> \hsize=524pt
> \vsize740pt
> \def\nbinstruments{1}
> \setstaffs11
> \setclef10
> \makeatletter%
> \def\titles#1#2{\kern-\lin@pos%
> \kern-3.2\Interligne\kern-\parindent\kern-\afterruleskip%
> \kern-\sign@skip\vbox{\vskip#1\Interligne%
> \if#2\empty\relax\else\leftline{\Bigtype #2}\fi}}%
> \makeatother%
> \startmuflex\startpiece\addspace\afterruleskip%
> \znotes\zcharnote{16}{\titles{2.0}{XXXDessus}}\en%
> \alaligne
> \znotes\zcharnote{16}{\titles{2.0}{Dessus}}\en%
> \Endpiece
> \vfill\eject\endmuflex
> \bye
>
>
