|
From: Hans-Bernhard B. <HBB...@t-...> - 2012-05-17 23:14:02
|
[Forgot to CC this to the list...] On 16.05.2012 23:36, pl...@pi... wrote: > Now the obvious technical solution would be to parse the whole plot line > to work out what is needed and do it once. No, because that simply wouldn't work. You're overlooking the case that the argument to functions like column() can, itself, depend on the data: plot '-' using 1:column($2) I.e. we could inspect the plot command as much as we want, and still not know which input columns to buffer. |