Quoting Wolfgang Schuster
2007/8/9, Aditya Mahajan
: On Thu, 9 Aug 2007, Wolfgang Schuster wrote:
2007/8/9, Hamid Kamrani
: Aditya,
In the example you give with CMID you have only two column preambles whereas the table has 3 columns and if one adds the missing preamble then one gets an error. Some how your new code breaks the calculation for number of columns.
-Hamid
He forgot only to put a \NC before the \AR.
Sorry about that. I thought I pasted from a correctly running example.
Here is another attempt, which I feel is the cleanest way to go about this whole thing. Right now the skip between the columns is not configurable, but it should be possible to use the get the value of the skip.
The attached code typesets an example that Wolfgang had sent off-list. I have only changed two lines of the table.tex.
Aditya
Hi Aditya,
what do you think about
\setuptables [\c!columndistance=2em]
This goes against the way I understand table macros. With this, you can only have equal width columns. In tables, you can specify specific widths of specific s and o keys. AFAIU, the table macros divide that the specified tabskip by 2, and place it in the data column. Since the rule column comes next, the rule column automatically gets the same tabskip, so in appearance we the specified column width. This is also the reason why cmidrules do not work. If you specify a hrule inside a data column, it spans half of the visual space between the columns. What needs to be done is that the glue be added inside a local group to the rule column, with the global tabskip set to zero. The previous solution that I posted adds a 2em tabskip glue to the rule column, and a 0pt tabskip glue to the data column. After thinking again about it, I think that this is not the right way. There is no need to add a 0pt glue to the data column. We can add the glue to the rule column inside a group. I will see if I can get this working. Aditya