Rainer M Krug wrote: > Ken Mankoff <mank...@gmail.com> writes: >> On 2014-03-11 at 08:47, zwz wrote: >>> In my setup, there is >>> (setq org-export-exclude-tags '("private" "exclude") >>> >>> and In my test.org: >>> >>> * test >>> >>> ** Not exported :exclude: >>> #+BEGIN_SRC ditaa :file test.png :cmdline -E >>> +--------+---+---+---+---+---+---+---+ >>> +---+---+---+---+---+---+---+---+ >>> x | 0 cRED | 0 | 0 | 0 | 0 | 0 | 0 | 0 | | 0 | 0 | 0 | 0 | 1 | 0 | >>> 1 | 1 | >>> +--------+---+---+---+---+---+---+---+ >>> +---+---+---+---+---+---+---+---+ >>> #+END_SRC >>> >>> ** blah blah >>> blah blah blah >>> >>> >>> When I try to export it to pdf, the test.png is still generated, >>> although it is not used for the pdf at all. >>> So I think the export procedure may be optimized for more efficiency. >> >> No, because I often have code and sections I don't want exported, but >> I want their side-effects active. For example, code with sessions >> where part is not exported, but I need that code run so code >> elsewhere, using the same session, is able to run and be exported. > > AFAIK, this depends if you use a session. When you are using a session, > *all* code blocks are evaluated, if you do not set the header argument > session, only the ones which are exported are evaluated - precisely the > reason you give.
We should know if zwz uses a session or not. Because, IIUC, what he reports is that all code blocks are evaluated, even the ones which are not exported. Best regards, Seb -- Sebastien Vauban