> I created the ds FIND*COMMA from DATE*FULL and used .substring > to find the pairing ``, ''. I created the ds DATE*SHORT and > subjected it to the successful .substring request. > The only thing worth mentioning is that FIND*COMMA must be > redefined at the beginning of each nested conditional, having > been irrevocably altered during the previous conditional.
Would you mind posting your solution? I'm curious as to how you coded it. Anyhow, in an earlier message I said I didn't think it was possible to do all the required processing inside a single string invocation. It turns out I was wrong, it's entirely possible. Here is a sample solution, which strips off everything up to the first comma, irrespective of context: .\" ---------------------------------------------------------------- .de DATE*SHORT \c .ds AA \\*[DATE*FULL] .ds BB \\*[DATE*FULL] .YY .AA .. .de YY .ds CC \\*(BB .length BB \\*(BB .if \\n(BB>1 .substring BB 1 .substring CC 0 0 .if '\\*(CC',' .ZZ .if \\n(BB>1 .YY .. .de ZZ .if \\n(BB>1 .ds AA \\*(BB .if \\n(BB=1 .ds AA [single existing comma is last character] .nr BB 0 .. .\" ---------------------------------------------------------------- .ds DATE*FULL Friday, December 31, 2038 .sp 3c Full date is: \*[DATE*FULL]. .br Short date is: \*[DATE*SHORT]. .\" ---------------------------------------------------------------- Of course, it's inefficient, because it has to repeat all the processing every time you expand DATE*SHORT. > > Or there's a `.while' IIRC. :-) The ".while" construct isn't strictly necessary, because the roff language supports recursion. 8-)
