![](https://secure.gravatar.com/avatar/565476157eb3895a9c928ffce26d988d.jpg?s=120&d=mm&r=g)
Aditya Mahajan wrote:
Hi,
I am attaching myway on using matharrow. Are there any suggestions before I release it to public?
I found some surprises while making this my way.
1. Somewhere along the line of development, \xrightarrow{text} changed to placing text at the bottom of the arrow (rather than the top as in amstex). This is not a bug, just something that works differently. Should this be changed? These arrow definitions are pretty new, so I do not think that a change at this stage will effect too many users. It may not be possible to change later for compatibility reasons.
In the interest of porting stuff from amstex, changing the behaviour is probably better.
2. Something goes wrong with palatino (used by mag-01). Have a look at the \xmapsto and \xrightleftharpoons and \xleftrightharpoons in the pdf. \xmapsto can be corrected by changing its defintion to
\definematharrow [xmapsto] [3599] [\mapstofill]
Basically, providing a bit more cusion on the left, so that the middle \relbar not overlap the \mapstochar.
Should this definition go to math-ext.tex or to something specific to palatino. I have not tested with other fonts, but what is the general strategy when definitions differ for different fonts?
something should be possible with mathcollections, I guess? (Hans knows this better then I do)
I could not correct leftrightharpoons. Why does this not work?
\definematharrow [xleftrightharpoons] [3395,3359] [\leftharpoonupfill,\rightharpoondownfill]
Because the current definition of dodoxmtharrow does not like it when the second arg starts with two identical digits. Changing it to this fixes the problem: \def\dodoxmtharrow[#1,#2,#3][#4,#5,#6][#7]#8#9% {\edef\tempa{#2}\ifx\tempa\empty %TH changed line \mathrel{\domthxarrsingle{#7}{#1}{#4}{#8}{#9}}% \else \mathrel{\domthxarrdouble{#7}{#1}{#2}{#4}{#5}{#8}{#9}}% \fi}