Sunday, December 8, 2002 Hans Hagen wrote: HH> At 07:24 PM 12/7/2002 +0100, you wrote:
def\startalign{% \startformula \let\\\cr \grabuntil{stopalign}\doeqalign } \def\doeqalign#1{\eqalign{#1\cr}\stopformula}
HH> ^^ better make that \crcr in order to catch HH> users \\ Will fix. I just "copied" the AMS defs ...
\let\stopalign\relax
These allow me to write:
\startalign a & = b = \\ & = c \stopalign
HH> isn't \startmathalign a better choice? Maybe. I just "converted" the AMS-LaTeX "align" environment, so I kept the align name.
and everything is fine. Yet I cannot do
\placeformula[hook] \startalign a & = b = \\ & = c \stopalign
because of expansion problems. Why? How can I define the stuff?
HH> took me a couple of cd's to uncover this -) HH> (1) placeformula has to look ahead, and the next command is not seen as HH> something math; so, HH> (2) i made another lookahead, and then HH> (3) we ran into those \cr & and # things that may not always show up HH> everywhere, so HH> (4) i had to rewrite a low level macro that i didn't like rewriting, i.e. HH> (5) i had to use a variant i'd always thought that could be avoided HH> so, you spoiled my afternoon -) :P Sorry, but it's your choice to work on sundays ;) HH> btw, if you isolate/collect the ams math that's not yet in m-newmat, i can HH> see if/how i can add it there in a configurable way, so if some things HH> fails, we can look into it later Ok. I'm "filling it up" as it comes while typesetting/converting some notes (part of them were hand-written, part of the in LaTeX form). HH> you can try the version i just uploaded Will do. -- Giuseppe "Oblomov" Bilotta