While we're on the subject of execution in the edit buffer, it is not
entirely clear what such execution would mean. Mirko's desire seems to
be to execute the buffer directly without org-babel as an
intermediary. However, the most consistent approach would be a command
to pop back to the source buffer from the edit buffer and use
org-babel-execute-src-block to run the code. If org were to provide
edit buffer evaluation it would likely have to be that. The reason for
this is that direct execution of code in the major mode of the edit
buffer is not well defined. For some languages (e.g. racket) there is
no easy way to evaluate a buffer that does not correspond directly to
a file on the file system, which is part of why org babel exists in
the first place (to define those semantics). In theory babel languages
could provide an alternate implementation for how an edit buffer can
be run, but absent such an implementation it would be up to the user
to figure out the semantics for each of the languages they wanted to
evaluate directly. Best,
Tom
