top post - sorry: I overlooked this thread somehow. Eric Schulte writes: > Achim Gratz writes: > >> Eric Schulte writes: >>> If you mean that there should be new syntax for setting header arguments >>> on a file or sub-tree basis w/o using file local variables, I'd be happy >>> to apply a patch. >> >> I'm thinking that something like >> >> #+PROPERTY: header-args:R :session "*R*" :exports none >> >> should work. > > I hate to change syntax, but the syntax you mention above does look both > appealing and natural. Even with working file local variables [1]. I agree - it looks very intuitive. > >> I've checked that the property interface returns the data as expected, >> but I haven't implemented anything yet. It does not seem to be an >> overly difficult endeavour, however. >> > > That is very good news. With that portion working I would agree that > this should be a fairly straightforward task. > >> >>>> But importantly, there should be no way to set a default session name >>>> without also specifying the language, regardless of which way one >>>> tries to set this up. >>> >>> If you can think of a clean way to implement this then we should go for >>> it. I doubt many existing configurations rely on this behavior. >> >> General settings for all languages should be effected by >> >> #+PROPERTY: header-args :results value :exports none >> >> and there'd be a list of header arguments (or specific values) that are >> either ignored or warned about when not associated with a particular >> language. >> >> BTW, I think the current property syntax for header arguments should be >> deprecated since it is the only place where the leading ":" is missing >> for those. >> >> Comments, thoughts? >> > > I think these are great ideas. Personally I'd love to see them > implemented. Unfortunately I don't have time to work on an > implementation currently. I'm surprised that none of the users who > motivated this discussion have chimed in. Their opinions may be more Shame on my head - here are now a few comments (and praises) following in next emails. Cheers, Rainer > valuable than my own in this regard (as I'm not a heavy #+Property > user). > > Thanks for the clean and initially tested implementation idea! > >> >> >> Regards, >> Achim. > > > Footnotes: > > [1] Thanks to Bastien my patch allowing variables like > `org-babel-default-header-args:R' to be set file-local has been > applied to Emacs. -- Rainer M. Krug email: RMKruggmailcom