On 1-6-2011 3:52, Mojca Miklavec wrote:
On Wed, Jun 1, 2011 at 15:36, Wolfgang Schuster wrote:
Am 01.06.2011 um 15:26 schrieb Mojca Miklavec:
Dear magitians,
I discovered this in LaTeX, but ConTeXt behaves the same (plain TeX behaves differently). Can somebody please explain me why this code fails to work? (A workaround is to move \newif on top which I'm willing to do, but I'm still curious.)
\starttext
\ifx\hbox\undefined \message{invisible to tex} \newif\ifabc \abcfalse \ifabc \message{abc true breaks} \else \message{abc false breaks} \fi \fi
\stoptext
systems : begin file iftest at line 1 abc false breaks ! Extra \fi. l.7 \fi
The \ifx ends with the first \fi
I figured that out, but I still find it very weird.
from \ifabc which isn’t defined and ignored by TeX
This part somehow makes sense (TeX should not keep defining stuff and writing out stuff) ...
the last \fi is left which generates the error message
The code works when using \ifx\hboxx\undefined ... But despite the explanation, this is very very very weird. Or to quote Taco: "The trick of TeX being bug-free is that all the bugs and limitations are documented." I hope that this behaviour is documented in TeXbook.
Actually the code does what it is supposed to do .. tex's scanner does a fast scan for \else\fi and nested \if's so when you define one in a to be skipped branch you're on your own. The reason why plain tex behaves a bit different is that there \newif is defined as outer. Actually, you might like to stare for a while at the definition of \newif, especially if you love catcodes. All language have their specialities, like local l = { ["in"] = "in", out = "out", } in Lua. ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------